Jim Breyer

1. Snapshot and testable edge thesis

Breyer's proposed edge is institutional interpretation across technology cycles: combine prepared thematic research with founder, university, corporate, and specialist networks; recognize a simple high-engagement wedge; sponsor decisively; then use recruiting, governance, and patient capital to help an adaptive team navigate institutional change. [researcher inference · moderate] profile edge · HBS doctrine · Circle reflection

The thesis applies most plausibly to early-stage software/infrastructure from network computing through vertical AI, especially where technical product and regulated or incumbent institutions must meet. It is not a general claim about public-market timing, commodity technology, or late-stage price selection. [researcher inference · moderate]

Predictions. A fresh Breyer-linked investment should be more likely than a random venture deal to involve a repeat founder or trusted network introduction, a narrow/high-frequency product, interdisciplinary authority, differentiated data or distribution, staged proof, recruiting intensity, and a long holding horizon. [researcher inference · moderate] fintech sourcing · current healthcare doctrine

Falsifiers. The thesis fails if a complete ledger shows no advantage outside Facebook/Circle; ordinary results are weak after reserves/fees; decisive sourcing and work consistently belong to other partners; the published method appears only after success; or network access predicts selection but not outcomes. [researcher inference · strong]

Confidence boundary. Facebook, Circle, 2001/2005/2008 direct speech, healthcare writings, and filings support the mechanism. Missing IC records, cost bases, denominator, failures, and value-add measurement prevent a stronger causal or performance claim. [researcher inference · insufficient public record] source coverage

2. Core philosophy

Prepared mind before the meeting

Research the technology category, map users and institutions, and build a view before a company appears. The meeting should update a prepared hypothesis, not create a theme from charisma. [investor-stated · strong] Stanford 2005

Simple product, important market

Prefer an intelligible entry product whose use or value is visible, then test whether it can win a meaningful share of an important market. Simplicity is an entry condition, not an excuse to ignore hidden policy or operational complexity. [investor-stated · strong] HBS 2008

Adaptive team over frozen plan

Select founders who listen, learn, recruit, and change strategy while retaining ambition. Test this prospectively through references and observed updates; otherwise “adaptability” becomes hindsight. [investor-stated · strong] Stanford listening

Capital follows risk reduction

Finance the next decision-changing milestone and preserve a buffer. Cash available, runway, reserves, and justified burn are separate. [investor-stated · strong] PBS 2001 · Stanford burn

Institutions can validate or kill

Regulation, universities, medical centers, global partners, and distribution incumbents can supply expertise or barriers. They can also create legal shutdown, false validation, concentration, and conflict. [researcher inference · strong] Prosper lab · Walmart proxy

Hold through cycles, but do not confuse duration with price immunity

Structural technology adoption can survive sentiment cycles, and difficult markets can improve entry quality. Yet Accel VIII and the AOL–Time Warner forecast show that secular correctness does not rescue fund size, price, or company integration. [researcher inference · strong] PBS contradiction · Forbes adverse record

3. Sourcing

Breyer's public sourcing system has five channels:

1. Category preparation: explicit research before Facebook and later AI/healthcare theses. [investor-stated · strong] Stanford 2005
2. Partner discovery and pursuit: Chien and Efrusy in Facebook show institutional sourcing before senior sponsorship. [retrospective witness · moderate] Forbes 2011
3. Repeat founders: Jeremy Allaire/Circle and the Harbor operator rationale reduce relationship uncertainty but not business risk. [retrospective witness · moderate] Circle note · Harbor note
4. Universities and specialists: Stanford, medical institutions, and present healthcare collaborators supply technical/domain discovery and references. [investor-stated · strong] TechCrunch note · Paige note
5. Global/portfolio/co-investor referrals: IDG-Accel, fintech founders, and portfolio networks extend geography and adjacency. [investor-stated · moderate] fintech note

A reusable sourcing audit records introducer, prior relationship, economic conflicts, independent channels searched, unavailable populations, and which thesis existed before the lead. Network density is an advantage only if it preserves disconfirming search. [researcher inference · strong]

4. Picking

Team

Observe listening, update quality, recruiting pull, integrity under bad news, complementary domain/technical authority, and ability to state the simple wedge. Do not convert pedigree into evidence. [researcher inference · moderate] Circle board/recruiting note

Product

Ask whether usage or workflow value is high-frequency and measurable. Facebook's campus engagement is the strongest public example; Subtle and Iterative show the later workflow version. [researcher inference · moderate] Stanford 2005 · Subtle note · Iterative note

Market

Define the important market through a budget owner, replacement path, or new category—not a broad technology label. A simple product needs a credible path to large share. [investor-stated · strong] HBS note

Data and defensibility

Separate access from ownership, consent, representativeness, refresh, portability, and lawful commercial use. Regulation is defensibility only after legal operation is proven. [researcher inference · strong] AI note · Prosper order

Price and ownership

The public philosophy is thin. Breyer discusses cycles, fund size, and importance, but no accessible rule gives target ownership, valuation ceiling, expected-value hurdle, dilution budget, or loss ratio. A modern implementation must add them. [researcher inference · insufficient public record]

Decision card

[researcher inference · strong]

5. Deal and portfolio mechanics

The earliest explicit mechanic is sequential risk removal. Company milestones should distinguish technical feasibility, product engagement/workflow use, legal/regulatory operation, distribution, economics, and organization. The next tranche belongs to the binding risk, not the most flattering metric. [researcher inference · strong] PBS note

Portfolio construction requires an additional layer absent from company storytelling: opportunity count, target ownership, partner bandwidth, initial checks, reserves, correlation, duration, and fund size. Accel VIII is the warning that ample LP demand can exceed the investable opportunity and governance capacity. [researcher inference · strong] Accel VIII lab

No public record establishes Breyer's loss ratio, reserve policy, recycling, pro rata rules, follow-on decision rights, ownership targets, or sell discipline. Circle's retained IPO shares and Facebook's trust/fund holdings show patience, but filings alone do not reveal why, for whom, or at what return threshold. [researcher inference · insufficient public record] Facebook filing · Circle filing

6. Vehicle and incentive fit

Accel, Breyer's trust, Breyer Capital, IDG-Accel China, public boards, and philanthropic/university roles are distinct. Each changes capital duration, authority, conflicts, information rights, compensation, and return attribution. [contemporaneous record · strong] vehicle table

Accel's LP model can create fee/fund-size incentives and reserve constraints. The 2011 profile's 2.5% fee and 30% carry apply to the reported context, not every vintage. Breyer Capital may provide longer duration and concentration freedom, but private allocation, governance, and succession rules are unknown. [retrospective witness · moderate] Forbes 2011

The practical vehicle test asks:

• Does the vehicle survive the validation/holding period?
• Who approves initial and follow-on capital?
• Which person or vehicle owns information and opportunity rights?
• What ownership and reserve burden can one company consume?
• Which board, personal, fund, and family interests require disclosure or recusal?
• Are liquidity and outcome reported at company, vehicle, LP, or person level?

[researcher inference · strong]

7. Team attribution and work with founders

Breyer's best-supported founder work is senior sponsorship, recruiting, board governance, cross-institutional introductions, and patient relationship continuity. The Circle interview says board meetings should start with adverse issues; the mechanism is earlier truth and faster resource allocation. Efficacy is not independently quantified. [retrospective witness · moderate] Circle note

Facebook requires explicit role separation:

• Chi-Hua Chien — discovery in reported accounts.
• Kevin Efrusy — sourcing/pursuit in reported accounts.
• Jim Breyer — senior sponsor/negotiator, director, and personal-trust investor.
• Accel partnership and fund entities — institutional approval/capital.
• Mark Zuckerberg and team — company building and operating outcomes.

[retrospective witness · moderate] Forbes team account · [contemporaneous record · strong] Facebook filing

The same rule applies to current healthcare work: Mustafa Cheatham, clinicians, scientists, founders, other partners, and institutions retain authorship and diligence credit. “Breyer-backed” is a capital/association statement, not a full causal account. [researcher inference · strong] current-thesis note

8. Views on the craft

Venture is preparation plus judgment under incomplete information, not spreadsheet certainty. Early-stage selection centers on people and an important/simple wedge; capital and governance should help the team learn rather than force a frozen plan. [investor-stated · strong] HBS note

The craft is also institutional. Breyer repeatedly crosses company, fund, university, global, public-board, and family-office contexts. That breadth can generate interpretation and recruiting advantages, but it also raises allocation, independence, and attribution burdens. [researcher inference · strong] Walmart proxy

Finally, venture is power-law and incentive-sensitive. A few outcomes may drive the result; fund demand and carry can distort behavior; and a correct theme can coexist with bad prices and weak funds. The proper response is explicit denominators and vehicle-level economics, not hero narratives. [researcher inference · strong] investment Pareto

9. Signature bets and decision process

Facebook

Prepared category work plus a simple/high-engagement product, adaptive founder, senior sponsorship, board work, and long holding. Critical controls: Chien/Efrusy/Accel attribution; trust versus funds; beneficial ownership versus economic interest; sale versus return. [researcher inference · strong] Facebook lab

Circle

Repeat-founder relationship plus financial-infrastructure thesis and long-duration governance. Critical controls: 2013 rationale is retrospective; affiliated holdings/sale are not personal return; retained shares remain market risk. [researcher inference · strong] Circle lab

Paige and the healthcare-AI cohort

Domain/data/team/institution thesis applied to a high-value workflow. Critical controls: commercial rights, multi-site validation, regulatory and buyer path, cohort denominator, and reserves. [researcher inference · strong] Paige lab

Prosper as negative signature

Fintech/regulatory potential met an invalid securities-offering path. Critical control: the SEC order names Prosper, not Breyer. The case belongs in the philosophy because it tests a recurring institutional-moat claim. [contemporaneous record · strong] Prosper lab

10. Mental models and maxims

1. Prepare narrowly, then decide on a simple high-use wedge. Theme work directs attention, while user/workflow evidence constrains narrative.
2. Back learning velocity and recruiting, not plan fidelity. Record how founders update and recruit before outcomes are known.
3. Capital is a sequence of risk purchases. Fund the next binding proof and preserve survival buffer.
4. Regulation is a gate before it can be a moat. No defensibility credit until the company can lawfully operate.
5. Use institutional networks, then audit them for bias and conflict. University, corporate, and portfolio ties require independent validation.
6. Separate company success, securities, liquidity, and return. Value, financing, shares, gross sales, proceeds, and net return are distinct.
7. Hold structural winners, but re-underwrite cycle, price, and concentration. Duration does not excuse valuation or fund-size mistakes.

[researcher inference · strong] full field guide

11. Evolution and contradictions

[researcher inference · strong]

The largest internal tension is scale versus discipline. Breyer warns against excess capital and sector tourists, yet Accel VIII reportedly became oversized. He favors team adaptability, but public records rarely freeze the prospective team score. He values regulation as a constraint, yet Prosper failed the registration perimeter. He claims contrarian patience, but no public price/ownership hurdle shows when patience becomes overpayment. [researcher inference · strong]

12. Failures and limits

The evidence-complete adverse set includes Accel VIII/prior weak funds, the missed Google opportunity, Prosper's registration failure, the AOL–Time Warner forecast, and Harbor's unknown acquisition economics. Clinkle remains an unresolved adverse lead because no verified write-off/security ledger was found. [researcher inference · strong] failure ledger

Method limits:

• Broad networks may generate access rather than differentiated judgment.
• Famous institutions and repeat founders may weaken independent search.
• “Team over strategy” can become unfalsifiable without dated assessments.
• A simple wedge can conceal regulation, safety, and buyer complexity.
• Family-office patience can conceal opportunity cost without explicit return thresholds.
• Published winners create survivorship and rationale-selection bias.
• Facebook can dominate economics and reputation while obscuring ordinary results.

[researcher inference · strong]

No record proves a career-wide hit rate, loss rate, median multiple, fund return, personal cash return, value-add effect, or conflict-free allocation history. [researcher inference · insufficient public record]

13. Historical operating environment

Breyer's method formed in a US venture market moving from semiconductor/network specialists toward Internet platforms, with smaller relationship-driven partnerships and public exit windows. The late-1990s capital surge rewarded access and speed while punishing fund-size and price discipline after 2000. [researcher inference · moderate] PBS note

The 2005 Facebook decision preceded mature mobile distribution, modern social advertising, current privacy regimes, and today's network scale. Circle began before stablecoins had today's institutional/legal framing. Healthcare AI now operates amid stronger models but higher evidence, privacy, integration, and safety expectations. [researcher inference · strong]

Vehicle conditions changed too: Accel partnership capital, a personal trust, IDG-Accel, public boards, and a post-2006 family office offer different duration and authority. A historical lesson transfers only after replacing the original vehicle, information, regulation, and price assumptions with current ones. [researcher inference · strong]

14. Modern VC translation (as of 2026-08-21)

What transfers

• Prepare an explicit theme map before sourcing.
• Demand a simple measurable workflow/use wedge.
• Score founder update and recruiting behavior prospectively.
• Separate technical, regulatory, workflow, commercial, and economic milestones.
• Pair domain and technical diligence while preserving independent channels.
• Map every person, vehicle, conflict, and decision right.
• Preserve duration for genuine compounding, with price/concentration limits.

[researcher inference · strong]

What changed

AI model capability is more accessible, so differentiated data, workflow distribution, evaluation, safety, and cost matter more than a generic model claim. Universities remain important but talent and models diffuse faster. Regulation is more fragmented across privacy, AI, financial, and clinical regimes. Later-stage/private liquidity and large funds can blur realization. [researcher inference · moderate]

A 2026 implementation

For each deal, preserve a dated “before” packet: thesis before contact, source/relationship map, user behavior/workflow evidence, independent domain references, data-rights schedule, regulatory authority map, model/evidence tests, buyer/economics, team update examples, ownership/price/reserves, vehicle allocation, conflicts/recusals, and kill milestones. Reopen it at every follow-on. [researcher inference · strong]

Misuse guards

Do not call a warm introduction proprietary sourcing, an institutional affiliation validation, a financing a success, an IPO price a return, retained shares realized, a famous founder adaptive, or long duration patient without an opportunity-cost threshold. [researcher inference · strong]

15. Comparison with completed peers

[researcher inference · moderate] Breyer is the broadest network institutionalist, Casado the narrowest technical specialist, Stephens the most mission/government explicit, and Doriot the closest long-duration governance analogue. The records cannot support a synthetic return ranking because vehicles, eras, marks, archives, and denominator quality differ. Casado underlying market-annealing essay · Stephens underlying interview

16. The strongest case against greatness

The strongest skeptical explanation is that Breyer's visible record combines one dominant Facebook outlier, Accel's platform and partner labor, elite network access, family/personal capital flexibility, favorable power-law luck, and selected outcome publishing. Chien and Efrusy supplied meaningful Facebook work; the Accel partnership supplied capital and approval; critics questioned credit and a personal side allocation; earlier funds reportedly lost money; and no complete denominator tests ordinary selection. [researcher inference · strong] Forbes 2011

Circle prevents the simplest “only Facebook” dismissal, but not the full skeptic case. The filing proves affiliated shares and liquidity, while cost basis, allocation, fees, and net return remain absent. Redback/Foundry outcome claims are reported, and the 2011/2025 >100x lists differ. Healthcare AI supplies a coherent doctrine but mostly private/unknown outcomes. [researcher inference · strong] Circle filing · denominator

The greatness claim becomes stronger only with contemporaneous IC/pass/follow-on records, partner credit, full cost/proceeds schedules, vintage fund returns, and a complete adverse ledger. Until then, the evidence supports an unusually consequential senior sponsor and network builder—not lone genius or audited universal outperformance. [researcher inference · insufficient public record]

17. Analyst study guide, glossary, and go deeper

Recommended order

1. Read the profile for chronology and the testable edge.
2. Read PBS 2001, Stanford 2005, and HBS 2008 before later winner narratives.
3. Run the Facebook, Circle, and Prosper labs without reading reveals.
4. Read the AI essay, Paige, and current co-led healthcare doctrine.
5. Test every outcome against the vehicle map and review.

Glossary

• Prepared mind: explicit category work completed before a company meeting.
• Simple wedge: bounded initial product whose use/value can be observed.
• Adaptive team: founders who update, recruit, and preserve truth under change.
• Sequential risk removal: capital tied to the next binding uncertainty.
• Institutional interpreter: translator among technology, users, regulators, incumbents, and capital.
• Vehicle fit: match among duration, ownership, reserves, authority, incentives, and liquidity.
• Deemed beneficial ownership: filing concept that can combine control across entities; not identical to personal economics.
• Gross liquidity: shares sold × offer price before costs; not net return.
• Power-law test: measure how much of visible value depends on the top outcome.
• No-hindsight lab: freeze formation-date evidence before reading results.

Reusable questions

• What was known before the introduction, and who produced that knowledge?
• What simple behavior/workflow falsifies the product claim?
• Which dated founder updates demonstrate learning?
• Who regulates, validates, integrates, buys, and benefits?
• Which data rights and independent tests survive commercialization?
• What ownership, price, reserve, and partner-capacity assumptions fit the vehicle?
• Who discovered, sponsored, approved, governed, followed on, and exited?
• Which personal/fund/board interests require disclosure or recusal?
• What happens if the top outcome is removed?
• Which missing artifact would most change the conclusion?

[researcher inference · strong] Go deeper through the source map, writings, talks, investment forensics, and mental-model field guide.

Pareto 80/20 — what drove the record

Three rankings are required because company value, investor proceeds, and learning value are different measures.

Economic contribution — observable evidence, not audited investor return

The rank is reproducible only as a public-evidence proxy: disclosed Breyer-linked liquidity first, reported realized multiple second, company consideration/mark third. It is not a personal or fund MOIC ranking. [researcher inference · strong]

Historical importance

Profile chronology

Learning value

Failure ledger

Evidence rules and searched denominator

An item enters the bounded public scope if it appears as one of 110 unique company headings on Breyer Capital's live portfolio page or is one of 12 additional materially researched, public Breyer-linked companies absent from that page. The official headings were case-insensitively exact-deduplicated; repeated section appearances collapse to one company; “Contact Us” and “Featured Companies” are navigation headings and are excluded. The union below is a reproducible disclosed scope, not a career universe. [contemporaneous record · strong] Breyer Capital portfolio

“Researched” means at least one primary/filing or material independent route was opened beyond a name. “Sufficiently reconstructed” requires a bounded formation rationale, role/vehicle evidence, a material outcome/adverse fact, and an explicit economics limit. Financing, valuation, company consideration, affiliated holdings, shares sold, investor proceeds, and return remain separate fields. [researcher inference · strong]

Bounded 122-name union inventory

Official portfolio inventory — 110 unique company headings

The five-column table is alphabetized down each numbered block. Names preserve the live page's displayed spelling; aliases used in the research ledger are reconciled below. [contemporaneous record · strong]

Additional researched names absent from the official page — 12

The 25-row researched ledger is a strict subset of this union: 13 official-page lineages—Meta / Facebook (ledger “Facebook”), Circle, Etsy, Prosper, Spotify, Legendary, Marvel, Walmart.com, Kensho, Subtle Medical, Iterative Scopes, Datalogix, and Open Evidence (ledger “OpenEvidence”)—plus all 12 additional names numbered 111–122. [researcher inference · strong]

Counted search funnel

The bounded disclosed-scope reconstruction ratio is 7 / 122 = 5.738%, rounded 5.7%. It cannot be extrapolated to the career because the official page says it presents only “some” past and present investments, mixes vehicles, and omits stealth, write-offs, passes, and unknown legacy deals. Career-wide coverage fraction: not computable. [researcher inference · strong] source coverage

Comprehensive public person-linked ledger

[researcher inference · strong] The ledger is an attribution-control sample, not a “greatest hits” list. Where the source is only the official portfolio, it establishes a vehicle-level marketing association and no more. Official portfolio · filing/source controls

Vehicle, economics, and incentive boundaries

[researcher inference · strong] Facebook vehicle footnotes · Walmart recusal · Forbes fund terms

People, capital, and influence map

[researcher inference · strong] Forbes Facebook team account · Circle interview · current healthcare doctrine

Historical and base-rate context

The 1990s network-software period offered rapidly expanding enterprise/Internet adoption, public-market exits, and severe valuation dispersion. By 1999–2001, excess venture supply and large funds increased the danger that correct themes produced weak prices and too many companies. Breyer described this at the time; Accel VIII supplies a reported adverse portfolio control. [contemporaneous record · moderate] PBS note · Forbes 2011

Facebook's 2005 environment combined nascent social networks, rapid campus engagement, unclear identity/privacy policy, uncertain monetization, and strong competing platforms. Today's scaled-network economics were not known. [researcher inference · moderate] Stanford 2005 note

Circle's 2013 environment combined Bitcoin-era infrastructure promise with extreme regulatory, adoption, custody, liquidity, and business-model uncertainty. The later stablecoin/IPO outcome must not leak into the formation test. [researcher inference · moderate] Circle interview note

Healthcare AI in 2018–2026 sits after deep-learning advances but faces long clinical validation, privacy, FDA, reimbursement, hospital integration, and model-generalization cycles. A family office may fit duration, but no cohort economics establish that the strategy beats these base-rate constraints. [researcher inference · strong] healthcare writings synthesis

No-hindsight decision labs

Each lab freezes the investable evidence at the decision period. Do the questions before reading the reveal.

Decision lab 1 — Facebook, 2005

Formation-date evidence. Accel had researched social networks; Facebook offered a simple campus product with intense repeat use; Chien/Efrusy supplied discovery/pursuit; Breyer was a senior sponsor. [contemporaneous record · strong] Stanford 2005 note · Forbes team account

Unknowns then. Durable non-campus adoption, identity norms, moderation/privacy burden, monetization, competitive response, founder control, and terminal platform economics.

Alternatives. Other social networks, consumer-Internet opportunities, or waiting for monetization proof.

Risks. Fad/retention decay, platform displacement, governance conflict, regulatory harm, high implied price, and partner/person allocation conflict.

Provisional underwriting. Invest only if cohort engagement is exceptional, campus expansion repeats, the founder recruits and learns, governance rights are clear, and ownership justifies the binary risk. Size reserves against explicit geographic/product milestones; document team/vehicle allocation before approval. [researcher inference · moderate]

Stop before reveal — reader exercise.

• Would you invest then? Choose invest, pass, or stage-and-observe using only the 2005 evidence.
• What would kill the deal? Specify the retention/expansion failure, governance breach, or competitive response that ends the thesis.
• What ownership/terms fit the risk? State target ownership, board rights, reserve budget, allocation/conflict disclosure, and price ceiling.
• What evidence unlocks the next check? Name the campus-expansion cohort, repeat-use threshold, recruiting proof, and accountable partner review.

Reveal. Breyer joined the board in April 2005; Accel entities and his trust held distinct securities; the IPO created a dominant visible outcome. This does not validate every ex-ante premise or lone attribution. [contemporaneous record · strong] Facebook filing

Decision lab 2 — Circle, 2013

Formation-date evidence. Jeremy Allaire was a known repeat founder from Macromedia; digital-currency infrastructure could become an Internet-native financial rail; Breyer describes relationship-based conviction. The account is retrospective, not a 2013 memo. [retrospective witness · moderate] Circle note

Unknowns then. Regulatory classification, banking access, custody/security, stable value, consumer versus infrastructure model, token liquidity, unit economics, and institutional adoption.

Alternatives. Exchanges, custody providers, traditional fintech rails, or waiting for regulatory clarity.

Risks. Existential enforcement, fraud/security, reserve confidence, volatile demand, business-model pivots, follow-on intensity, and reputational spillover.

Provisional underwriting. Back the founder only with a regulatory route map, named banking/compliance owners, security controls, staged product milestones, and vehicle reserves adequate for a decade-long path. [researcher inference · strong]

Stop before reveal — reader exercise.

• Would you invest then? Choose invest, pass, or stage-and-observe using only the 2013 evidence.
• What would kill the deal? Name the regulatory, banking, custody, security, or adoption event that invalidates the thesis.
• What ownership/terms fit the risk? State ownership, governance, reserve rights, legal covenants, and duration-adjusted price.
• What evidence unlocks the next check? Define the lawful product wedge, repeat use, named banking/compliance control, and security proof.

Reveal. Circle evolved and completed a 2025 IPO; affiliated holders sold 3,355,468 shares at $31 and retained 13,360,862. Return and personal/vehicle allocation remain unproved without cost basis. [contemporaneous record · strong] Circle 424B4

Decision lab 3 — Prosper, 2006

Formation-date evidence. Peer-to-peer credit promised a new marketplace connecting borrowers and lenders; Chris Larsen articulated regulation as a potential barrier/defensibility issue. Breyer's public segment is a question, not authorship of Larsen's answer. [contemporaneous record · strong] regulatory transcript note

Unknowns then. Whether notes required registration, state-law exposure, lender/borrower protection, credit-cycle losses, servicing continuity, and whether marketplace growth could pause safely.

Alternatives. Partner with regulated incumbents, build only enabling software, or wait for a formal offering path.

Risks. Illegal securities issuance, shutdown/quiet period, class actions, state penalties, adverse selection, credit deterioration, and reputational harm.

Provisional underwriting. Pass until securities counsel and regulators validate the offering path, or invest with a hard pre-launch legal gate and capital explicitly conditioned on registration. Regulation is a moat only after lawful operation is demonstrated. [researcher inference · strong]

Stop before reveal — reader exercise.

• Would you invest then? Choose invest, pass, or stage-and-observe using only the 2006 evidence.
• What would kill the deal? Specify the unresolved registration, state-law, servicing, credit, or customer-protection issue.
• What ownership/terms fit the risk? State launch conditions, stop authority, downside protection, reserves, and board/legal information rights.
• What evidence unlocks the next check? Identify the written legal/regulatory clearance, compliant offering route, servicing-continuity test, and loss data.

Reveal. The SEC found Prosper sold unregistered securities from 2006 through 2008. The company entered a quiet period; filings disclose state settlement and litigation exposure. Breyer later returned as an investor after remediation. The finding is against Prosper, not Breyer. [contemporaneous record · strong] SEC order · Prosper S-1/A · 2015 company release

Decision lab 4 — Paige, 2018

Formation-date evidence. Digitized pathology could support machine learning; a leading cancer center supplied domain/data adjacency; the team bridged medical and technical expertise. [investor-stated · strong] Paige note

Unknowns then. Commercial data rights, external representativeness, FDA path, prospective clinical utility, pathologist adoption, procurement, integration, and gross margin.

Alternatives. Radiology AI, research tools, workflow software without diagnosis, or waiting for multi-site validation.

Risks. Institution-specific overfit, slow digitization, regulatory delay, liability, buyer concentration, and capital intensity.

Provisional underwriting. Invest only if data rights survive commercialization, an external-validation plan precedes platform expansion, and milestones separate algorithm accuracy, workflow use, regulatory evidence, and paid adoption. [researcher inference · strong]

Stop before reveal — reader exercise.

• Would you invest then? Choose invest, pass, or stage-and-observe using only the 2018 evidence.
• What would kill the deal? Define the data-rights, external-validation, regulatory, workflow, or procurement failure.
• What ownership/terms fit the risk? State ownership, data/control covenants, board rights, reserve budget, and price for a long evidence cycle.
• What evidence unlocks the next check? Require multi-site validation, commercial data rights, regulatory progress, pathologist use, and paid deployment.

Reveal. Public evidence in this project establishes continuing company/portfolio presence, not a realized investor return or proof that the formation gates were met. [researcher inference · insufficient public record] official portfolio

Decision lab 5 — Iterative Scopes, 2021

Formation-date evidence. Endoscopy generates video; specialists can label findings; AI may improve detection and clinical-trial operations; the company named clinical collaborators and evidence ambitions. [investor-stated · strong] Iterative note

Unknowns then. False-positive/negative harm, prospective utility, workflow latency, data ownership, reimbursement, hospital/pharma buyer split, and sales-cycle length.

Alternatives. Non-AI endoscopy workflow, pharma trial software, other imaging/control points, or wait for prospective evidence.

Risks. Dual-market dilution, regulatory delay, clinician resistance, weak prospective lift, fragmented data, and reserve burden.

Provisional underwriting. Stage capital through retrospective performance, prospective clinical utility, workflow adoption, and paid repeat deployment. Keep diagnostic and trial economics separate. [researcher inference · strong]

Stop before reveal — reader exercise.

• Would you invest then? Choose invest, pass, or stage-and-observe using only the 2021 evidence.
• What would kill the deal? Define the clinical endpoint, workflow-latency, buyer, data-rights, or dual-market failure.
• What ownership/terms fit the risk? State ownership, tranche/reserve plan, governance, information rights, and valuation ceiling.
• What evidence unlocks the next check? Require prospective utility, workflow adoption, a named budget owner, and paid repeat deployment.

Reveal. The public record documents the original rationale and continuing association, not a complete milestone history or realized return. [researcher inference · insufficient public record] official portfolio

Decision lab 6 — Accel VIII, 1999 portfolio construction

Formation-date evidence. Internet adoption was real, venture inflows and valuations were rising, and Accel had network-software experience. Public retrospective reporting says the fund initially reached $1.2 billion. [retrospective witness · moderate] Forbes 2011

Unknowns then. Bubble duration, ownership available at scale, partner bandwidth, reserve demand, exit windows, and whether a larger fund forced stage/style drift.

Alternatives. Cap fund size, return capital, split strategies, slow deployment, or require higher proof at later-stage prices.

Risks. Fee incentive, overcapitalized companies, crowded sectors, weak governance bandwidth, valuation compression, and correlated portfolio failure.

Provisional underwriting. Size the fund from opportunity count × target ownership × reserve policy, not demand. Require a base-rate case in which prices normalize and exit windows close. [researcher inference · strong]

Stop before reveal — reader exercise.

• Would you accept or cut then? Choose the proposed fund, a smaller fund, a split strategy, or no commitment using only 1999 evidence.
• What would kill the fund-size case? Specify opportunity-count, price, partner-capacity, reserve, concentration, or exit-window limits.
• What ownership/terms fit the risk? State target ownership, check/reserve ratios, fee/carry alignment, pace limits, and concentration cap.
• What evidence unlocks the next commitment? Require a bottom-up company pipeline, partner workload, downside vintage model, and deployment schedule.

Reveal. Forbes reported the fund was reduced to $680 million and lost money, following weak vintages and LP departures. The report is not an audited LP statement, but it is a real adverse portfolio control. [retrospective witness · moderate] Forbes 2011

Portfolio forensics and adversarial tests

Ex-Facebook test

Removing Facebook leaves a meaningful but materially weaker public case. Circle has filing-controlled liquidity; Redback and Foundry have reported >100x outcomes; Spotify, Etsy, Kensho, and others supply company-level success. But no audited security cash flows or complete loss denominator establish elite net performance without Facebook. [researcher inference · strong]

Reported-list consistency test

The 2011 profile associates reported >100x results with Redback, Foundry, and RealNetworks. The 2025 list names Facebook, Circle, Redback, and Foundry. The difference may reflect selection, realization timing, definition, or memory; the public record cannot resolve it. [retrospective witness · moderate] Forbes 2011 · Forbes 2025

Role-sensitivity test

Facebook is re-described after allocating discovery to Chien, pursuit to Efrusy, senior sponsorship/board work to Breyer, company building to Zuckerberg/team, and capital/approval to Accel. The deal remains important, but the claim changes from lone picker to senior institutional sponsor. [researcher inference · strong]

Vehicle-sensitivity test

Combining Accel funds, Breyer trust, Breyer Capital, IDG-Accel, and board compensation inflates both breadth and apparent economics. Separating them removes the false inference that every official portfolio entry or beneficially owned share is a personal Breyer investment. [researcher inference · strong]

Survivorship/publication test

The live portfolio says “some” past and present investments; official writings discuss selected positive cases; no full pass/write-off ledger exists. Therefore win rate, loss rate, median multiple, and share of value from the top one or two outcomes cannot be computed. [researcher inference · insufficient public record]

Time and realization test

Private marks, IPO shares retained, company consideration, and gross shares sold are not distributions. Circle's sale is a real liquidity event, but retained shares, lockups, taxes, fees, cost basis, and vehicle allocation remain necessary for return. [researcher inference · strong]

Failures, misses, ordinary outcomes, and anti-portfolio

Accel VIII and prior weak funds

• Original thesis/evidence — Internet adoption was real and Accel had network-software experience; the prospective fund memo is unavailable. [researcher inference · moderate] PBS note
• Warning signs — Breyer contemporaneously identified excess capital, weak entrants, and sector tourists; 2011 reporting adds prior weak funds and LP departures. [retrospective witness · moderate] Forbes 2011
• Capital response — The fund reportedly began at $1.2b and was cut to $680m; company deployment, reserves, and write-down timing remain unknown. [retrospective witness · moderate]
• Governance — Accel partnership governance applied; Breyer's vote, veto, and sole responsibility are unknown. [researcher inference · insufficient public record]
• Stop/exit — No security-level stop, sale, or write-off schedule was found. [researcher inference · insufficient public record]
• Outcome — Forbes reported that the vintage lost money; no audited LP statement was located. [retrospective witness · moderate]
• Claimed lesson — Secular correctness does not rescue fund size, price, partner capacity, or correlated deployment. [researcher inference · strong]
• Later process change — Accel IX was smaller and later contained Facebook, but public evidence does not prove a formal rule change caused the outcome. [researcher inference · moderate]
• Attribution — This is an Accel partnership/vintage outcome, not a Breyer-only cash loss. [researcher inference · strong]
• Counterfactual uncertainty — A smaller fund, slower pace, or different ownership targets may have helped, but no contemporaneous alternative model permits quantification. [researcher inference · insufficient public record]

Google miss

• Original thesis/evidence — The 2011 profile reports that Accel missed/passed Google; the proposal, product evidence, price, and thesis are unavailable. [retrospective witness · moderate] Forbes 2011
• Warning signs — Unknown; no contemporaneous pass memo or warning log was found. [researcher inference · insufficient public record]
• Capital response — No capital was invested in the reported opportunity; counterfactual check and ownership are unknown. [researcher inference · insufficient public record]
• Governance — Exact sourcing partner, diligence owner, IC participants, dissent, and decision rights are unknown. [researcher inference · insufficient public record]
• Stop/exit — The reported action was a pass/miss, not an owned position with an exit. [researcher inference · moderate]
• Outcome — Google became an exceptional company; no counterfactual Accel/Breyer return can be calculated. [researcher inference · strong]
• Claimed lesson — Prepared minds still miss outliers; preserve dated pass reasons and later compare them. [researcher inference · strong]
• Later process change — No public record ties the miss to a specific later sourcing or IC change. [researcher inference · insufficient public record]
• Attribution — The source reports an Accel miss; it does not establish a Breyer-only decision. [retrospective witness · moderate]
• Counterfactual uncertainty — Access, allocation, ownership, dilution, follow-ons, and holding behavior make the forgone return unknowable. [researcher inference · strong]

Prosper registration failure

• Original thesis/evidence — Peer-to-peer credit promised a borrower/lender marketplace; Chris Larsen, not Breyer, supplied the substantive regulatory-defensibility answer. [contemporaneous record · strong] Stanford note
• Warning signs — Securities-registration, state-law, servicing, customer-protection, and credit-cycle risks were binding and incompletely resolved. [researcher inference · strong]
• Capital response — Accel/trust interests are filed and Breyer later returned after remediation; intervening checks, write-downs, and reserves are unknown. [contemporaneous record · moderate] Prosper 2011 filing
• Governance — Breyer had a board-era association; counsel advice, IC roles, and authority to halt issuance remain unknown. [researcher inference · insufficient public record]
• Stop/exit — Prosper entered a quiet period and changed its offering path; no Breyer/Accel security-level exit schedule was found. [contemporaneous record · strong]
• Outcome — The SEC found Prosper sold unregistered securities; issuer filings disclose settlements/litigation; the company later continued. [contemporaneous record · strong] SEC order · S-1/A
• Claimed lesson — Regulation is a right-to-operate gate before it can become a moat. [researcher inference · strong]
• Later process change — Returning after remediation is consistent with updated underwriting, but no documented Breyer process-change memo was found. [researcher inference · moderate] 2015 release
• Attribution — The SEC order names Prosper, not Breyer; no investor misconduct or write-off is inferred. [contemporaneous record · strong]
• Counterfactual uncertainty — Earlier registration, a regulated partner, or delayed launch might have reduced harm, but cost, timing, and viability are unknown. [researcher inference · insufficient public record]

AOL–Time Warner forecast

• Original thesis/evidence — Breyer expected a powerful integrated Internet/media platform while correctly identifying broader bubble excess. [investor-stated · strong] PBS note
• Warning signs — Integration, price, organizational complexity, and bubble-era assumptions could invalidate the company-specific forecast. [researcher inference · moderate]
• Capital response — No Breyer/Accel investment linked to this forecast is claimed; capital response is not applicable on the available record. [researcher inference · insufficient public record]
• Governance — Breyer had no documented governance role in the transaction in this package. [researcher inference · insufficient public record]
• Stop/exit — Not applicable as an investment case; no owned security is reconstructed. [researcher inference · insufficient public record]
• Outcome — The combination did not validate the favorable integration forecast. [researcher inference · strong]
• Claimed lesson — Separate a correct structural technology thesis from company, price, and integration underwriting. [researcher inference · strong]
• Later process change — No public statement ties this forecast miss to a formal later rule change. [researcher inference · insufficient public record]
• Attribution — The forecast is Breyer's direct speech; the corporate outcome belongs to AOL/Time Warner management and market conditions. [researcher inference · strong]
• Counterfactual uncertainty — Alternative terms, timing, or integration could have differed; no defensible investor-return counterfactual exists. [researcher inference · strong]

Harbor acquisition — ordinary-outcome boundary

• Original thesis/evidence — Breyer Capital cited a known operator, insurance opportunity, and sequential scope. [investor-stated · strong] Harbor note
• Warning signs — Decision-date commercial, regulatory, capital, and distribution kill criteria were not published. [researcher inference · insufficient public record]
• Capital response — Initial/follow-on checks, reserves, ownership, dilution, and vehicle allocation are unknown. [researcher inference · insufficient public record]
• Governance — Board rights, recruiting work, and decision authority are unknown. [researcher inference · insufficient public record]
• Stop/exit — A public acquisition association exists; sale process and investor consent/influence are unknown. [researcher inference · insufficient public record]
• Outcome — Purchase price, investor proceeds, and cost basis were not found. [researcher inference · insufficient public record]
• Claimed lesson — “Acquired” is not a return category; ordinary outcomes require security economics. [researcher inference · strong]
• Later process change — No public postmortem or process-change record was found. [researcher inference · insufficient public record]
• Attribution — The investment rationale is Breyer Capital's publication; founder/team and acquirer retain operating/outcome credit. [researcher inference · strong]
• Counterfactual uncertainty — Without price, proceeds, timing, or alternatives, profit/loss and a better exit path are unknowable. [researcher inference · strong]

Clinkle — unresolved adverse lead

• Original thesis/evidence — Public association exists, but no contemporaneous Breyer memo or reliable formation rationale was located in the processed routes. [researcher inference · insufficient public record]
• Warning signs — No source-quality-controlled warning chronology was sufficient for a factual postmortem. [researcher inference · insufficient public record]
• Capital response — Check, security, vehicle, follow-on, reserve, write-down, and recovery are unknown. [researcher inference · insufficient public record]
• Governance — Sponsorship, IC, board, recruiting, and stop authority are unknown. [researcher inference · insufficient public record]
• Stop/exit — No verified Breyer security-level exit or write-off was found. [researcher inference · insufficient public record]
• Outcome — A specific cash loss or multiple cannot be claimed. [researcher inference · insufficient public record]
• Claimed lesson — Adverse reputation is a retrieval lead, not a substitute for attribution and economics. [researcher inference · strong]
• Later process change — None is documented. [researcher inference · insufficient public record]
• Attribution — Association does not establish Breyer control or sole responsibility. [researcher inference · strong]
• Counterfactual uncertainty — Without thesis, role, cash flows, or decision records, no invest/pass or recovery counterfactual is defensible. [researcher inference · insufficient public record]

Terminal uncertainty and retrieval ledger

[researcher inference · insufficient public record] These terminal statuses mean the scoped public routes are exhausted, not that the private facts do not exist. No archive or person was contacted; no missing value is estimated.

Bottom line

The defensible investment record is a team- and vehicle-produced power-law career. Facebook dominates the visible economics; Circle is the strongest second filing-controlled outcome; Redback and Foundry are important reported historical wins; Prosper and Accel VIII are real adverse controls. The reusable method is prepared thematic search, simple high-engagement entry, adaptive-team selection, milestone capital, specialist/institutional validation, recruiting, governance, and patience. [researcher inference · strong]

The limits are equally important: only 7 of 122 bounded public names are sufficiently reconstructed; the official universe is selected and incomplete; return claims lack audited cash flows; and related vehicles/roles can easily be conflated. Career-wide coverage fraction: not computable. [researcher inference · insufficient public record] review · philosophy

Coverage and evidence base

The priority written corpus contains nine items: nine processed, zero partial, zero inaccessible. Every ★★★ and ★★ writing in the source map has an exact-schema note and a source-by-source analysis below. Two Austin posts visible only through the Medium profile index are ★ metadata leads and do not enter the denominator. [researcher inference · strong]

The corpus is unusually useful because it spans one general AI framework, three healthcare-thesis statements, four contemporaneous investment announcements, and one fintech thesis. It is also selected publication: the author explains investments he chose to announce and does not publish an equivalent loss, pass, or full-deal ledger. [researcher inference · strong]

Literal priority-source reconciliation

No duplicate mirror is counted as an additional work. The 2022 healthcare page's old URL slug and newer displayed title are one item. The live healthcare principles page is a different, current firm artifact and remains separately attributed. [researcher inference · strong]

Source-by-source analysis

1. Some Thoughts on AI Investing — the general filter set

The 2019 essay is the cleanest general statement. Breyer prefers domain-specific AI over undifferentiated horizontal claims and asks whether a company combines specialized expertise, proprietary or advantaged data, responsible practice, and an interdisciplinary team. [investor-stated · strong] note, takeaways and philosophy

The mechanism is practical: a vertical workflow supplies a bounded problem, data supplies learning and defensibility, experts specify what correctness means, and a mixed team turns research into deployment. The method screens for both technical possibility and institutional adoption. [researcher inference · moderate] full essay

The essay leaves hard questions open. It does not define minimum dataset rights, evidence thresholds, pricing, sales cycles, model drift controls, or the economics that justify domain-heavy teams. “Responsible” is a criterion without a published audit procedure. [researcher inference · strong] note, contradictions and open questions

2. Healthcare AI Thesis — 2022 update: portfolio pattern recognition

The 2022 body moves from general filters to a portfolio map. It emphasizes clinical-workflow integration, differentiated healthcare data, regulatory/evidence demands, and teams that connect technologists with medical specialists. Universities and existing portfolio relationships function as discovery and validation channels. [investor-stated · strong] 2022 note

Its evidence is strongest as contemporaneous description of what Breyer Capital believed. Named holdings show consistency but not causal validation: a list of selected investments cannot establish hit rate or return, and portfolio adjacency can amplify correlated data, reimbursement, and deployment risks. [researcher inference · strong] source map, denominator boundary

3. Healthcare AI Thesis — 2023 update: narrower applications and shared credit

The 2023 essay narrows attention to concrete areas including diagnostics, clinical trials, drug discovery, workflow productivity, and patient engagement. It also makes collaborators and specialist inputs more visible, which is analytically valuable: the doctrine is becoming an institutional method rather than a lone-investor intuition. [investor-stated · strong] 2023 note, takeaways and biography evidence

The strongest improvement is problem specificity. The main unresolved issue is commercialization: clinical merit, FDA clearance, or access to data does not by itself establish buyer urgency, budget ownership, reimbursement, integration cost, or attractive venture returns. [researcher inference · strong] 2023 note, contradiction section

4. Current healthcare and life-sciences principles — co-led doctrine

The current firm page synthesizes clinical need, evidence, workflow, data, responsible AI, and interdisciplinary company building. It is explicitly co-led with physician-investor Mustafa Cheatham; collective examples and conclusions must not be restated as Breyer's sole authorship. [contemporaneous record · strong] current-thesis note

Its modern value is a fuller diligence stack: clinical problem, technical performance, regulatory path, health-system implementation, safety, and team composition. Its weakness is measurement. The page still does not publish thresholds, failed validations, portfolio-wide follow-on decisions, or cohort economics. [researcher inference · strong]

5. Investing in the Future of Fintech — networks as both thesis and sourcing engine

The fintech essay combines a broad inclusion/digital-infrastructure thesis with a sourcing map: founders refer founders, IDG and global partners extend geography, and portfolio/co-investor ties create warm paths. [investor-stated · strong] fintech note

That network can reduce search cost and improve reference depth, but it can also create homophily, adverse-selection blind spots, and repeated exposure to the same market assumptions. The essay offers no cold-source comparison or evidence that network-referred companies outperform. [researcher inference · strong]

The most reusable idea is to underwrite the institutional rail rather than a transient interface. The misuse is to call every regulated financial app infrastructure without proving licenses, compliance ownership, risk capital, unit economics, and durable distribution. [researcher inference · moderate]

6. Introducing Paige.AI — data and domain at the point of diagnosis

The 2018 Paige announcement is valuable because it freezes the thesis before later outcomes: pathology is a high-skill diagnostic workflow; digitized slides can become a differentiated learning asset; Memorial Sloan Kettering provides domain/data adjacency; and the team bridges medicine and machine learning. [investor-stated · strong] Paige note

The no-hindsight diligence questions are data rights, representativeness, pathologist workflow, regulatory evidence, hospital procurement, and whether the institution relationship survives commercialization. The announcement does not answer all of them, and later financing cannot retroactively prove the initial underwriting. [researcher inference · strong]

7. Introducing Subtle Medical — workflow control before grand platform claims

The Subtle Medical announcement focuses on a narrow control point: improve medical-image quality and/or acquisition efficiency within existing radiology workflows. It emphasizes Stanford-linked technical/clinical collaboration and a product that could add value without replacing the scanner fleet. [investor-stated · strong] Subtle note

The attractive mechanism is deployment leverage; the risk is that an apparent software layer depends on scanner vendors, protocol variation, clinical validation, reimbursement, and hospital integration. The essay states the opportunity more clearly than the kill criteria. [researcher inference · moderate]

8. Announcing Iterative Scopes — sequential clinical proof

The Iterative Scopes post targets gastrointestinal endoscopy and clinical trials, where video data, specialist labels, workflow insertion, and pharma/provider demand may reinforce one another. It names evidence generation and clinical collaboration rather than relying only on generic AI language. [investor-stated · strong] Iterative note

This is the corpus's strongest example of a staged proof path: algorithm performance must become clinical utility, trial/product adoption, and economic value. The missing controls remain data ownership, false-negative/false-positive costs, buyer split, and reserve requirements across a long healthcare cycle. [researcher inference · moderate]

9. Why We Invested in Harbor — known operator, bounded entry, unclear economics

The Harbor essay is relationship-led. It emphasizes the team's prior operating record and an initial insurance distribution/product wedge, consistent with Breyer's preference for adaptive people over fixed plans. [investor-stated · strong] Harbor note

Harbor is also a useful ordinary-outcome boundary. The public record indicates an acquisition, but purchase price, investor proceeds, cost basis, and vehicle split were not found. A named acquisition is not evidence of a good venture return. [contemporaneous record · strong] Harbor note, contradictions and open questions

Cross-source synthesis

The common causal stack

Across the nine works, the recurring stack is:

1. Find an important but bounded institutional workflow.
2. Pair deep domain authority with technical builders.
3. Secure lawful, useful, representative data or another proprietary input.
4. Insert into an existing control point before expanding the platform claim.
5. Validate with named institutions, specialists, and measurable evidence.
6. Use founder and university networks for discovery, references, and recruiting.
7. Preserve room for the team to change strategy as evidence arrives.

This is a synthesis, not a verbatim seven-step method published by Breyer. [researcher inference · moderate] AI note · Paige note · current healthcare note

What remains stable

Team quality is the stable center. The healthcare corpus adds domain pairing to the earlier team-over-strategy idea; fintech adds network referrals; the investment posts show the principle in named contexts. Simple entry wedges and institutional access also persist. [researcher inference · strong] HBS spoken comparison · Subtle note

What changes

The doctrine evolves from a 2019 screen into a 2026 operating system. “AI plus data plus domain” becomes clinical evidence, workflow integration, responsible deployment, and specialist co-leadership. The evolution is partly a response to sector complexity; it is not proof that the current version was used in earlier deals such as Facebook or Redback. [researcher inference · strong]

Strongest counterarguments

• Selection bias: all nine works are positive or thesis-forward; no pass memo or write-off essay appears.
• Authority bias: elite universities and medical centers may improve validation but can weaken customer/base-rate independence.
• Data ambiguity: access is not ownership, portability, lawful reuse, representativeness, or durable advantage.
• Workflow ambiguity: clinical usefulness does not identify buyer, budget, integration cost, or distribution.
• Outcome leakage: later prestige and financing can make an initially incomplete thesis appear inevitable.
• Attribution: the current doctrine is explicitly co-led; portfolio companies and specialist collaborators contributed knowledge.

[researcher inference · strong] source-map limits

Evolution of the written philosophy

[researcher inference · strong]

Decision playbook and modern translation

Screen 1 — identify the real workflow owner

Ask who performs the task, who suffers the error, who buys, who integrates, who regulates, and who captures the savings. A product is not a vertical platform until those parties align. Use now: AI agents in medicine, finance, law, and industrial operations. Misuse: mapping an impressive demo onto a workflow without budget authority. [researcher inference · moderate] current healthcare note

Screen 2 — decompose “proprietary data”

Record source, rights, consent, representativeness, freshness, labeling, portability, security, and the counterfactual available to competitors. Use now: multimodal clinical data and regulated enterprise agents. Misuse: treating an academic affiliation or temporary API as a moat. [researcher inference · strong] Paige note

Screen 3 — test interdisciplinary operating behavior

Do not merely count impressive resumes. Observe who resolves disagreement, owns clinical/technical tradeoffs, recruits missing capability, and can change strategy without abandoning truth. Use now: scientific and domain AI teams. Misuse: assembling credentialed advisers without operating authority. [researcher inference · moderate] AI note

Screen 4 — stage proof in the right order

Define technical, workflow, regulatory, commercial, and economic milestones separately. Release capital as the company removes the next binding risk, while retaining a real cash buffer. Use now: long-validation companies. Misuse: funding “18 months of runway” without a decision-changing milestone. [researcher inference · strong] Iterative note · burn-rate talk note

Screen 5 — audit the network

Map who introduced the company, who supplied validation, who benefits, what independent channel was searched, and whether related roles require recusal. Use now: dense university/founder/co-investor ecosystems. Misuse: treating warm access as neutral validation. [researcher inference · strong] fintech note · Walmart proxy control

Screen 6 — keep vehicle fit explicit

Long clinical cycles require reserves, patience, governance, and follow-on authority. A family office may tolerate duration differently from a fund, but the public record does not reveal Breyer Capital's allocation or return thresholds. Use now: distinguish strategic patience from undisciplined duration. [researcher inference · strong]

Best five writings for study

1. Some Thoughts on AI Investing — the most compact general filter set.
2. Introducing Paige.AI — the best decision-date domain/data/team case.
3. Healthcare AI Thesis — 2023 — the clearest sector narrowing and collaborator attribution.
4. Investing in the Future of Fintech — the clearest sourcing-network map and its implicit risks.
5. Current healthcare/life-sciences principles — the most mature modern doctrine, correctly treated as co-led.

Read Why We Invested in Harbor immediately after the best five as an ordinary-outcome control: clear rationale plus acquisition does not equal a known venture return. [researcher inference · strong]

Evidence still missing

No accessible writing in the priority corpus supplies a complete declined-deal memo, investment-committee record, reserve model, valuation discipline, ownership target, loss/write-off analysis, portfolio-wide implementation scorecard, data-rights checklist, clinical evidence threshold, conflict policy, or realized return. The most useful next artifacts would be contemporaneous pass memos, follow-on memos, Breyer Capital allocation records, and a cohort table that includes every healthcare-AI company—not only selected examples. [researcher inference · strong]

Coverage and evidence base

The priority spoken corpus contains nine items: eight full transcripts or full Breyer-bounded segments and one partial Forbes video/clip corpus; all nine are processed into exact-schema notes and source-by-source analysis. The Stanford archive contains additional mixed-speaker appearances, but this synthesis uses only speech identifiable as Breyer's. [contemporaneous record · strong] source-map talk denominator

The corpus spans May 2001 to September 2025. It captures bubble-era warnings before Facebook, contemporaneous Facebook discussion in 2005, founder and fund doctrine in 2008, post-Accel sourcing in 2017, a long Circle/Facebook reflection in 2023, and partial outcome claims in 2025. Direct speech is strong evidence that Breyer said something at a date; it is not automatically proof that he used the rule consistently or caused an outcome. [researcher inference · strong]

Literal priority-source reconciliation

The long Stanford From Harvard to the Facebook program is supporting biography because Zuckerberg supplies most substantive answers. The Emotional Roller Coaster is primarily Chris Larsen. Neither is counted as an additional Breyer doctrine source. [contemporaneous record · strong] Stanford contributor archive

Source guide and analysis

1. PBS FRONTLINE (May 2001) — bubble diagnosis under pressure

This is the best early primary source because it predates the later signature deals. Breyer distinguishes a valid Internet adoption curve from unsustainable valuations, weak entrants, sector tourists, and excess capital. He argues that companies should remove risks in sequence rather than spend simply because financing is available. [investor-stated · strong] PBS note, takeaways and philosophy

The interview shows both insight and calibration limits. The ecosystem diagnosis aged better than his positive AOL–Time Warner integration forecast. A reusable process must therefore separate a correct structural thesis from company, price, and integration underwriting. [researcher inference · strong] PBS note, contradiction section

2. Stanford, Investing in New Ideas (October 2005) — prepared mind, not lone discovery

Breyer says Accel had researched social networks before meeting Facebook and describes a simple product with intense user engagement. The timing makes this strong evidence that category preparation and product-use signals were part of the public rationale in 2005. [investor-stated · strong] Stanford investing note

It does not identify every partner's sourcing or approval role, deal alternatives, price sensitivity, or personal-versus-fund allocation. Those require the team and filing controls in investments.md. [researcher inference · strong]

3. Stanford, Listening Skills (October 2005) — adaptive learning as founder evidence

In the bounded segment, Breyer treats listening and experimentation as signs that an entrepreneur can update without losing conviction. This supports the later “team over fixed strategy” doctrine, but it is a qualitative heuristic rather than a scored reference method. [investor-stated · strong] listening note

The failure mode is agreeable storytelling: investors can interpret any change as adaptability after success and as lack of focus after failure. A modern version needs dated hypotheses and update logs. [researcher inference · moderate]

4. Stanford, Regulatory Constraints (December 2006) — a speaker-attribution trap

Breyer asks about regulation and defensibility; Prosper founder Chris Larsen supplies the substantive answer. The transcript proves Breyer was probing the issue, not that every moat claim or answer was his doctrine. [contemporaneous record · strong] regulatory note

Prosper's later SEC order is the essential counterweight. The SEC found Prosper—not Breyer—sold unregistered securities from 2006 through 2008. Regulation can raise barriers only after the company has a valid legal operating path. [contemporaneous record · strong] SEC order

5. Stanford, Managing the Burn Rate (December 2006) — milestone capital plus buffer

The bounded exchange distinguishes money available from money justified by the next risk-reduction plan. It also recognizes that a company needs enough buffer to avoid financing from a position of immediate distress. [investor-stated · strong] burn-rate note

The public segment does not specify reserve ratios, bridge authority, insider-round pricing, or what happens after a miss. Those are the variables that turn a sensible slogan into portfolio discipline. [researcher inference · strong]

6. HBS One-on-One (March 2008) — the fullest pre-crisis venture doctrine

Breyer says early-stage teams matter more than a fixed strategy and favors simple ideas capable of winning a large share of an important market. He describes fund construction across stages, global/China expansion, a contrarian cycle, and the way carried interest can bias behavior. [investor-stated · strong] HBS note

The interview matters because it predates the full Facebook outcome and the financial crisis. Yet it remains an edited alumni Q&A without a decision ledger. The exact allocation, pricing, reserves, vetoes, and ordinary outcomes are absent. [contemporaneous record · moderate]

7. TechCrunch (November 2017) — sourcing after the Accel operating transition

Breyer emphasizes universities, machine learning, younger technical networks, and deliberate search for unusually ambitious founders. The interview connects the 2005 campus pattern to a post-Accel sourcing strategy. [investor-stated · moderate] TechCrunch note

The title's “next Mark Zuckerberg” framing creates outcome leakage: it centers an iconic winner while providing no prospective comparison between selected, passed, and failed teams. [researcher inference · strong]

8. Circle Money Movement (June 2023) — repeat founder, board practice, and present thesis

The Circle transcript is the richest later source. Breyer describes knowing Jeremy Allaire from Macromedia, backing Circle as a repeat-founder/product-infrastructure bet, opening board meetings with adverse issues, treating recruiting as central, and using multigenerational/interdisciplinary networks for current AI sourcing. [retrospective witness · strong] Circle note, 33:00–59:00 and 1:16:00–1:23:00

The role and memory claims are Breyer's perspective. The Circle prospectus independently controls securities and sale data but cannot validate the original decision narrative or board contribution. [researcher inference · strong] Circle 424B4

9. Forbes video and clips (September 2025) — outcome claims with a partial transcript

In the visible corpus Breyer reports four realized investments above 100x and emphasizes patience and public-market liquidity. A contemporaneous Forbes profile names Facebook, Circle, Redback, and Foundry. These are self-reported returns without cost-basis, security, vehicle, fee, or cash-flow schedules. [investor-stated · moderate] Forbes video note · Forbes profile, outcome paragraphs

The partial access boundary matters: no argument relies on unseen audio, and the 2025 list is not silently merged with the different 2011 reported list. [contemporaneous record · strong]

Recurring frameworks and strongest counterarguments

Prepare deeply, decide simply

The spoken record moves from sector research to a simple product/use judgment. Facebook is the strongest named example. The counterargument is that the category work may be institutional Accel labor and the simple-wedge rule may be reconstructed around a winner. [researcher inference · moderate] Stanford investing note · Forbes team attribution

Team over static strategy

Listening, learning, recruiting, and adaptive founders recur from 2005 through 2023. The counterargument is unfalsifiability: without dated team assessments and comparable failures, later pivots can always be called adaptive. [researcher inference · strong] listening note · HBS note · Circle note

Stage capital against risk

The 2001 and 2006 sources separate available cash from justified burn. The counterargument is Accel VIII: fund size and cycle exposure reportedly overwhelmed discipline at the portfolio level. [researcher inference · strong] PBS note · Forbes adverse record

Use regulation and institutions as both filter and moat

China, fintech, Prosper, and healthcare all require institutional interpretation. The counterargument is that institutional complexity can destroy rather than defend the business; Prosper is a direct legal-perimeter example. [researcher inference · strong] HBS note · Prosper S-1/A, legal proceedings

Be contrarian across cycles, not careless on price

Breyer distinguishes long-duration technology adoption from short-cycle sentiment and says difficult fundraising periods can be attractive. The counterargument is that “contrarian” can rationalize overfunding; no public valuation/ownership rule explains when price breaks the thesis. [researcher inference · strong] PBS note · HBS note

Build multigenerational, interdisciplinary networks

By 2017–2023, campus links, younger technical talent, repeat founders, medical specialists, and family-office continuity form the sourcing story. The counterargument is access dependence, homophily, related-role conflicts, and difficulty separating neutral validation from relationship reinforcement. [researcher inference · strong] TechCrunch note · Circle note

Evolution of spoken doctrine

[researcher inference · strong]

Biography, deals, and role evidence

[researcher inference · strong]

Five highest-value appearances

1. PBS FRONTLINE, 2001 — earliest complete process source and a named failed forecast.
2. HBS One-on-One, 2008 — fullest pre-crisis statement of team, fund, cycle, global, and incentive doctrine.
3. Stanford, Investing in New Ideas, 2005 — best contemporaneous Facebook/prepared-mind evidence.
4. Circle, The Power of Courage, 2023 — richest repeat-founder, board, recruiting, and current-method reflection.
5. Stanford, Managing the Burn Rate, 2006 — concise capital-versus-risk distinction that transfers cleanly.

Pair the top five with the Prosper regulatory segment and SEC order so the moat claim is tested against a real adverse outcome. [researcher inference · strong]

Evidence still missing

No accessible talk supplies a full Facebook 2005 decision memo, alternative set, valuation walk, ownership target, IC vote, partner-allocation record, reserve history, founder-reference script, postmortem, complete loss list, or portfolio return. The Forbes 2025 item lacks a full stable transcript. Searches of Stanford/STVP, PBS, HBS, Circle, TechCrunch, Forbes, Breyer Capital, SEC filings, and title variants produced no independent full record of those items. [researcher inference · insufficient public record]

Model 1 — prepare narrowly, then decide on a simple high-use wedge

• Rule — Build a category map before the company appears; then require a simple product whose use or workflow value is visible. [researcher inference · strong] Stanford 2005 note
• Evidence status — Strong as a stated rule and contemporaneous Facebook rationale; moderate as a career-wide implemented edge because no pass/control set exists. [investor-stated · strong]
• Provenance and origin — The 2001 bubble interview separates enduring Internet adoption from weak companies, and the 2005 Stanford segment describes prior social-network research. [investor-stated · strong] PBS note · Stanford note
• Mechanism and reasoning — Preparation narrows search and identifies important unknowns; simple use constrains theme enthusiasm with behavior. The combination reduces charismatic-founder and fashionable-sector error. [researcher inference · moderate]
• Historical and vehicle context — Accel's partnership supplied research, referrals, approval, and capital. “Prepared mind” should not be individualized when partners produced the lead or analysis. [retrospective witness · moderate] Forbes team account
• When to use — New categories with sparse data but observable early behavior: developer tools, network products, financial rails, clinical workflows, and vertical AI. [researcher inference · moderate]
• How to apply — Before contact, write the adoption thesis, incumbents, user pain, disconfirming facts, and valuation range. After contact, record frequency, retention/workflow lift, expansion, buyer, and what changed. [researcher inference · strong]
• Examples — Three bounded cases. [researcher inference · moderate]
  • Documented application: Facebook category preparation plus campus use in 2005. Stanford note
  • Retrospective application: Circle as a repeat-founder financial-infrastructure thesis. Circle note
  • Documented application: Paige as a pathology/data/workflow wedge. Paige note
• Evidence — Independent contemporaneous speech (PBS/Stanford), a decision-date Paige writing, and Facebook's filing-confirmed board/ownership chronology support different parts of the model. [contemporaneous record · strong] Facebook S-1/A
• Limits, failure modes, and exception ledger — Correct themes can coexist with poor companies, prices, and integration; AOL–Time Warner is a forecast miss, and Accel VIII is a portfolio-level adverse control. A simple interface can hide regulation or weak economics. [retrospective witness · moderate] Forbes adverse record
• Evolution and contradictions — The model evolves from Internet adoption and consumer use to institutional workflow/data. It contradicts itself when “simple” becomes post-hoc or category work is credited to one partner. [researcher inference · strong]
• Modern VC translation (as of 2026-08-21) — For AI, require a dated pre-meeting thesis plus deployment behavior, evaluation, cost, workflow owner, and independent reference. Model access alone is not preparation or a moat. [researcher inference · strong]

Model 2 — back learning velocity and recruiting, not plan fidelity

• Rule — Prefer founders who listen, update, recruit missing talent, and preserve truth under disagreement over founders who merely defend a static plan. [investor-stated · strong] Stanford listening note
• Evidence status — Strong as repeated direct doctrine from 2005–2023; moderate as causal selection evidence because prospective scores and failures are missing. [investor-stated · strong]
• Provenance and origin — The 2005 Stanford segment links listening to learning; the 2008 HBS interview states team over strategy; the 2023 Circle account emphasizes recruiting and bad-news-first boards. [investor-stated · strong] HBS note · Circle note
• Mechanism and reasoning — Early strategy is uncertain. Teams with high learning velocity can absorb customer, technical, and regulatory evidence; recruiting expands the option set; candid boards shorten error persistence. [researcher inference · moderate]
• Historical and vehicle context — Board access and reserves let an investor help a team update. A personal/family-office relationship may persist longer than a closed-end fund, but public allocation rules are unavailable. [researcher inference · insufficient public record] vehicle boundaries
• When to use — Pre-product-market fit, regulated markets, platform transitions, and technical/scientific companies where strategy will change but integrity and mission should not. [researcher inference · moderate]
• How to apply — Ask for three dated belief changes, references on bad-news behavior, recruiting wins/losses, unresolved role gaps, and a live disagreement. Score update quality before outcomes; revisit at follow-on. [researcher inference · strong]
• Examples — Four bounded cases, including an adverse boundary. [researcher inference · moderate]
  • Retrospective application: Facebook recruiting and board practice in Breyer's account. Circle note
  • Retrospective application: Repeat-founder conviction in Jeremy Allaire/Circle. Circle note
  • Documented application: Harbor's known-operator rationale. Harbor note
  • Exception: Prosper shows the boundary: an adaptive team and later remediation cannot substitute for a lawful operating path, and the SEC finding is against the company rather than proof of Breyer recruiting or governance efficacy. Prosper SEC order · regulatory note
• Evidence — Stanford and HBS provide independent dated doctrine; Harbor provides a decision-date writing; Circle provides later practice memory. None supplies a portfolio control group. [contemporaneous record · strong] Stanford transcript · HBS Q&A
• Limits, failure modes, and exception ledger — The model is vulnerable to hindsight: success makes pivots look adaptive, failure makes them look unfocused. Repeat relationships can create overtrust. Harbor's undisclosed acquisition economics prevent outcome validation; Prosper shows that later adaptation cannot cure an invalid initial right-to-operate path. [researcher inference · insufficient public record]
• Evolution and contradictions — “Listening” becomes team-over-strategy, then recruiting and multigenerational collaboration. The contradiction is the absence of a published rule for when too many updates mean weak conviction. [researcher inference · strong]
• Modern VC translation (as of 2026-08-21) — Preserve a founder-update ledger: initial belief, new evidence, changed action, owner, cost, and result. Use it for follow-ons and references; never infer learning velocity from a polished narrative alone. [researcher inference · strong]

Model 3 — capital is a sequence of risk purchases

• Rule — Release capital against the next binding risk and preserve a survival buffer; do not spend merely because money is available. [investor-stated · strong] PBS note
• Evidence status — Strong as contemporaneous 2001/2006 doctrine; moderate as implemented portfolio practice because reserve and tranche records are absent. [investor-stated · strong]
• Provenance and origin — Breyer described staged risk removal during the dot-com correction; the Stanford burn exchange distinguishes justified milestones from cash availability. [investor-stated · strong] burn-rate note
• Mechanism and reasoning — Capital buys information and option value. Technical, product, legal, commercial, and organizational uncertainties should be sequenced so a failed gate limits sunk cost and a passed gate improves financing power. [researcher inference · strong]
• Historical and vehicle context — Large funds and abundant capital can weaken the constraint; the reported Accel VIII experience is the portfolio-level counterexample. A family office may extend duration but still needs opportunity-cost gates. [retrospective witness · moderate] Accel VIII lab
• When to use — Long-validation AI, clinical, fintech, infrastructure, hardware, and market-creation bets where one proof materially changes the next risk or financing terms. [researcher inference · strong]
• How to apply — List risks in dependency order; define evidence, owner, cost, date, and kill condition; budget base and downside cases; separate company runway from investor reserves; precommit what a miss changes. [researcher inference · strong]
• Examples — Three bounded cases. [researcher inference · strong]
  • Researcher-applied illustration: Iterative Scopes—retrospective performance → prospective utility → deployment → paid repeat use. Iterative note
  • Researcher-applied illustration: Paige—data rights → external validation → regulatory path → procurement. Paige note
  • Exception: Accel VIII—reported fund scale and losses despite bubble warnings. Forbes 2011
• Evidence — PBS and Stanford are independent primary speech; Iterative/Paige are decision-date writings; Forbes supplies adverse fund reporting. No capital-call or reserve ledger validates execution. [contemporaneous record · strong]
• Limits, failure modes, and exception ledger — Milestones can reward easy metrics, underfund platform work, or create financing cliffs. The “buffer” can become unbounded runway. A binary legal risk may need resolution before any launch, as Prosper shows. [contemporaneous record · strong] Prosper order
• Evolution and contradictions — The rule begins as dot-com capital discipline and translates into clinical evidence sequencing. It is contradicted by any fund-size decision that precedes a bottom-up opportunity/capacity model. [researcher inference · strong]
• Modern VC translation (as of 2026-08-21) — Require milestone budgets to state what decision the evidence changes, not only KPI targets. For foundation-model products, include inference economics, evaluation drift, data rights, security, and workflow adoption. [researcher inference · strong]

Model 4 — regulation is a gate before it can be a moat

• Rule — Give no defensibility credit to regulation until the company has a lawful, operable path and named authority for stopping unsafe or noncompliant activity. [researcher inference · strong] Prosper order
• Evidence status — Strong as a researcher-derived rule grounded in direct adverse evidence; provisional as an investor-authored universal maxim because Larsen, not Breyer, supplies the 2006 substantive answer. [researcher inference · strong]
• Provenance and origin — Breyer asked how regulation creates defensibility in the 2006 Prosper discussion. The subsequent SEC order and issuer filing show the legal perimeter failed. [contemporaneous record · strong] regulatory note · Prosper S-1/A
• Mechanism and reasoning — Licensing, evidence, compliance systems, and institutional trust can deter entrants only after they enable service. Before then they create shutdown, liability, delay, capital, and reputation risk. [researcher inference · strong]
• Historical and vehicle context — Prosper involved Accel/trust interests; Circle required long-duration financial/regulatory navigation; healthcare AI adds FDA, privacy, clinical evidence, and health-system controls. The governing vehicle must survive delay and fund compliance work. [researcher inference · strong] Circle note · current healthcare note
• When to use — Fintech, crypto, clinical AI, insurance, marketplaces handling securities/credit, and products whose deployment changes regulated decisions. [researcher inference · strong]
• How to apply — Map statutes/regulators, licenses/registrations, counsel assumptions, stop authority, customer continuity, incident handling, capital needs, evidence thresholds, and a launch gate. Seek independent counsel and regulator/customer references. [researcher inference · strong]
• Examples — Three bounded cases. [researcher inference · strong]
  • Exception: Prosper sold unregistered securities and paused. SEC order
  • Retrospective application: Circle navigated a long infrastructure/regulatory path. Circle note
  • Researcher-applied illustration: Paige/Iterative require regulatory and clinical-evidence gates before moat credit. Paige note · Iterative note
• Evidence — The SEC order and issuer filing independently establish Prosper's legal outcome; Circle direct speech and healthcare writings show recurring institutional relevance. No private legal diligence files were found. [contemporaneous record · strong]
• Limits, failure modes, and exception ledger — Compliance can be necessary but non-differentiating; incumbents may have the advantage; rules can change; legal opinions may be conditional; a lawful product may still have weak unit economics or social harm. [researcher inference · strong]
• Evolution and contradictions — The 2006 inquiry becomes a negative lesson, while later fintech/healthcare doctrine incorporates more explicit evidence and responsibility. Returning to Prosper after remediation may show update discipline, but does not validate the original launch. [researcher inference · moderate] Prosper 2015 release
• Modern VC translation (as of 2026-08-21) — Add a “right to operate” memo to every regulated AI deal. No launch or next check without resolved authority, evidence, data, monitoring, and customer-continuity gates. [researcher inference · strong]

Model 5 — use institutional networks, then audit them for bias and conflict

• Rule — Use partners, repeat founders, universities, global firms, portfolio companies, and boards to source and validate—but record interests, independent channels, and recusals. [researcher inference · strong] fintech note · Walmart proxy
• Evidence status — Strong that the network exists and influenced sourcing; moderate that it produces superior selection; insufficient public record for a cold-source or conflict-adjusted outcome comparison. [researcher inference · insufficient public record]
• Provenance and origin — Facebook partner sourcing, IDG-Accel China, fintech referrals, university links, and medical collaborators recur across two decades. [retrospective witness · moderate] Forbes 2011 · TechCrunch note
• Mechanism and reasoning — Dense networks reduce search and reference cost, reveal talent, enable recruiting, and translate institutions. The same density can create homophily, mutual validation, allocation disputes, and information conflicts. [researcher inference · strong]
• Historical and vehicle context — Accel, personal trust, Breyer Capital, IDG-Accel, public boards, and philanthropy/universities confer different duties and beneficiaries. Treating them as one “network” hides decision rights and economics. [contemporaneous record · strong] Facebook filing
• When to use — Difficult expert markets, new geographies, repeat founders, university-origin technology, recruiting-intensive companies, and deals requiring incumbent integration. [researcher inference · moderate]
• How to apply — Map introducer, prior relationship, economic ties, board/fund roles, validator incentives, missing population, cold channel, and recusal. Require at least one independent customer/domain and one independent technical reference. [researcher inference · strong]
• Examples — Three bounded cases. [researcher inference · strong]
  • Documented application: Chien/Efrusy/Breyer/Accel form a team network in Facebook. Forbes 2011
  • Exception: Kosmix/Walmart records recusal and separate economics. Walmart proxy
  • Documented application: Paige/healthcare use a university and specialist network. Paige note
• Evidence — Journalism supplies Facebook partner roles, SEC filings control securities/recusal, and direct writings/speech supply the sourcing doctrine. No full relationship or referral-performance database exists. [contemporaneous record · strong] current healthcare note
• Limits, failure modes, and exception ledger — Warm access can amplify status and selection bias; university data may not transfer; global partners may own the local edge; related boards can require recusal; personal/fund opportunity allocation can attract criticism. [retrospective witness · moderate] Forbes criticism passages
• Evolution and contradictions — The network expands from Accel/technology executives to China, universities, family office, clinicians, and younger technical talent. Growth improves reach while increasing conflict and attribution complexity. [researcher inference · strong]
• Modern VC translation (as of 2026-08-21) — Build a relationship graph beside the investment memo: edge type, evidence supplied, beneficiary, conflict, independent corroboration, recusal, and data-use permission. Audit outcomes by source channel. [researcher inference · strong]

Model 6 — separate company success, securities, liquidity, and return

• Rule — Never substitute company valuation, financing, transaction consideration, beneficial ownership, shares sold, gross proceeds, or retained shares for cost-basis-adjusted investor return. [researcher inference · strong] Circle 424B4
• Evidence status — Strong as an accounting/evidence rule grounded in filings; insufficient public record for Breyer's career/fund/personal net returns. [researcher inference · insufficient public record]
• Provenance and origin — Facebook, Circle, Etsy, Kosmix, Marvel, Foundry, and Kensho expose different metrics and entity boundaries. Public profiles often compress them into “wins.” [researcher inference · strong] investment ledger
• Mechanism and reasoning — Return requires security, quantity, dates, cost, dilution, distributions/sales, fees/carry, taxes where relevant, and vehicle ownership. Missing any element limits the claim. [researcher inference · strong]
• Historical and vehicle context — Accel funds, personal trust, Breyer Capital, IDG-Accel, and director compensation have different beneficiaries. SEC deemed ownership may group controlled entities without assigning economic interest personally. [contemporaneous record · strong] Facebook S-1/A
• When to use — Every performance claim, IPO, tender, secondary, acquisition, private mark, board-share disclosure, and “greatest trade” ranking. [researcher inference · strong]
• How to apply — Build a security/vehicle cash-flow table; label reported, calculated, or unknown; distinguish gross and net; do not allocate affiliated sales without footnote support; show denominator and removed-outlier sensitivity. [researcher inference · strong]
• Examples — Three bounded cases. [contemporaneous record · strong]
  • Documented application: Facebook combines trust and six Accel entities but footnotes separate them. Facebook filing
  • Documented application: Circle's 3,355,468 shares × $31 ≈ $104.0m gross, not 100x proof. Circle filing
  • Documented application: Etsy separates 26,492,396 Accel shares from 2,943,533 personal/Breyer Capital shares. Etsy filing
• Evidence — Three issuer filings independently control the main security claims; Walmart controls Kosmix interests/recusal; journalism supplies explicitly reported return claims rather than audited data. [contemporaneous record · strong] Walmart proxy
• Limits, failure modes, and exception ledger — Filings can still omit private cost basis, fund distributions, fees/carry, taxes, side letters, and beneficial allocation. Gross arithmetic is precise but may be economically incomplete. [researcher inference · strong]
• Evolution and contradictions — The visible record shifts from journalist-reported fund/deal multiples to issuer-controlled securities, yet 2011 and 2025 >100x lists differ. Preserve both rather than harmonize. [retrospective witness · moderate] Forbes 2025
• Modern VC translation (as of 2026-08-21) — Require portfolio dashboards to show company mark, cost, realized gross, unrealized value, fund/vehicle ownership, fees/carry, and net LP return separately. Publish no multiple without dates and cash-flow scope. [researcher inference · strong]

Model 7 — hold structural winners, but re-underwrite cycle, price, and concentration

• Rule — Look through sentiment to structural adoption and hold compounding winners, while repeatedly testing price, fund size, ownership, concentration, and exit liquidity. [researcher inference · strong] PBS note · HBS note
• Evidence status — Strong as stated contrarian/patience doctrine; moderate as portfolio execution; insufficient public record for audited duration-adjusted results. [researcher inference · insufficient public record]
• Provenance and origin — The 2001 interview distinguishes adoption from the bubble; the 2008 HBS interview discusses cycle opportunity; later Circle/Forbes accounts emphasize long holding. [investor-stated · strong] Forbes video note
• Mechanism and reasoning — Sentiment dislocation can improve access and price; long holding preserves convexity and avoids premature sale. Re-underwriting prevents “long term” from becoming a defense of deteriorating fundamentals or oversized funds. [researcher inference · strong]
• Historical and vehicle context — Closed-end Accel funds, a personal trust, and Breyer Capital have different duration/liquidity constraints. Retained public shares do not prove identical holding decisions across beneficiaries. [contemporaneous record · strong] Circle filing
• When to use — Secular adoption with volatile financing/markets, durable network or infrastructure economics, and companies whose proof cycle exceeds current sentiment. [researcher inference · moderate]
• How to apply — Re-underwrite market size, cohort/workflow behavior, competitive position, governance, cash needs, valuation, ownership, reserve capacity, liquidity, and concentration at each financing or market regime shift. [researcher inference · strong]
• Examples — Three bounded cases. [researcher inference · strong]
  • Researcher-applied illustration: Facebook holdings and sales remained separated through IPO. Facebook filing
  • Researcher-applied illustration: Circle affiliates sold part and retained most IPO shares. Circle filing
  • Exception: Accel VIII reportedly scaled into bubble-era adversity. Forbes 2011
• Evidence — PBS/HBS are independent primary doctrine; Facebook/Circle filings control securities; Forbes supplies adverse fund reporting and later return claims. Holding causality and net returns remain unobserved. [contemporaneous record · strong]
• Limits, failure modes, and exception ledger — Structural theses can be wrong; price can absorb all upside; concentration can threaten a fund; regulation/technology can change; private marks can delay recognition; taxes and liquidity may justify sales. [researcher inference · strong]
• Evolution and contradictions — Bubble skepticism becomes contrarian cycle language and later patience. The contradiction is that a reported oversized fund and bad corporate forecast occurred alongside the doctrine. [researcher inference · strong]
• Modern VC translation (as of 2026-08-21) — Use a hold/sell memo with vehicle-specific constraints and an ex-top-one portfolio test. Separate belief in AI/crypto/healthcare adoption from the company, price, dilution, and liquidity case. [researcher inference · strong]

Evidence gaps and candidate models

[researcher inference · insufficient public record] These are hypotheses for future retrieval, not hidden additional models. The seven promoted models each have twelve components, at least three bounded applications/exceptions, and multiple independent evidence routes; none is represented as a universal Breyer algorithm. source map

What the evidence supports

The supported conclusion is narrower than a hero narrative: Breyer is an unusually consequential senior sponsor, network builder, and institutional interpreter whose strongest visible cases are Facebook and Circle. Direct speech supports prepared thematic search, adaptive-team selection, staged risk, recruiting, and institutional networks; filings support specified roles, entity-separated securities, and bounded liquidity. [researcher inference · strong] profile synthesis · investment bottom line

The record does not prove lone sourcing, audited career outperformance, a complete denominator, investor-causal value creation, or personal ownership of every security he beneficially controlled. Facebook remains the dominant visible economic outcome; Circle is the strongest ex-Facebook filing control; Redback/Foundry remain reported rather than cash-flow reconstructed; Accel VIII and Prosper remain adverse evidence. [researcher inference · strong] Pareto analysis · ex-Facebook test

What could falsify the thesis

• A complete company/pass/write-off ledger showing no better sourcing or selection outside privileged networks.
• Accel/Breyer Capital vintage cash flows showing weak net DPI/TVPI/MOIC after fees, reserves, and loss concentration.
• Contemporaneous Facebook/Circle records assigning the decisive thesis, negotiation, governance, or recruiting work elsewhere.
• A channel-adjusted cohort showing university/repeat-founder/warm leads underperform cold search.
• Healthcare-AI follow-on and outcome records showing the published domain/data/evidence filters were not used or did not discriminate.
• Allocation records showing unresolved personal/fund/board conflicts materially drove opportunity assignment.
• Independent founder/board records contradicting the claimed bad-news, recruiting, or long-duration contribution.

Conversely, contemporaneous IC/pass/follow-on memos, complete partner-role data, security-level cost/proceeds, audited fund returns, and a full adverse denominator would materially strengthen the thesis. [researcher inference · strong]

Attribution and denominator audit

Facebook is decomposed into Chien discovery, Efrusy pursuit, Breyer senior sponsorship/negotiation/board/personal-trust investment, Accel partnership approval/fund capital, and Zuckerberg/team company building. Circle, Etsy, Prosper, Kosmix, Legendary, Marvel, Walmart.com, and IDG-Accel receive equivalent entity/role controls. Unknown approval, allocation, follow-on, and exit roles remain unknown. [researcher inference · strong] people map · vehicle table

The bounded public universe is a listed, deduplicated union of 110 official portfolio names plus 12 additional materially researched names absent from that page. All 122 were name/source searched; 25 received material research; seven were sufficiently reconstructed; 115 remain unresolved within the bounded set. Every researched-ledger name resolves to an official alias or one of the 12 extras. The official page says it lists only “some” past and present investments and mixes vehicles. The 7 / 122 = 5.738%, rounded 5.7% reconstruction ratio describes this disclosed set only; career-wide coverage fraction: not computable. [researcher inference · strong] exact union · funnel · source coverage

Transfer limits for modern VCs

The transferable practices are a dated prepared-mind memo, a simple observable use/workflow wedge, prospective founder-update evidence, risk-sequenced capital, independent institutional/domain validation, and explicit people/vehicle/conflict maps. They transfer as process hypotheses, not as guarantees of selection or return. [researcher inference · strong] modern translation

Conditions have changed: AI model access diffuses quickly; data rights, evaluation, workflow distribution, security, safety, and inference economics matter more; regulation is fragmented; large funds/private liquidity blur realization; and universities or famous specialists can amplify authority bias. Do not copy Breyer's network breadth, holding period, or family-office concentration without matching vehicle duration, ownership, reserves, independent search, and opportunity-cost thresholds. [researcher inference · moderate]

Unresolved evidence

The highest-value missing artifacts are Accel LP returns and allocation agreements; Facebook 2005 IC/board and partner-role records; Circle's full financing/cost-basis history; Breyer Capital IC, reserve, loss/write-off, succession, and personal/fund allocation records; Redback/Foundry security cash flows; Prosper board/counsel records; IDG-Accel partner/vehicle attribution; healthcare-AI pass/follow-on/cohort data; and independent founder accounts of board contribution. Terminal uncertainty and retrieval ledger [researcher inference · insufficient public record]

No audited career/fund DPI, TVPI, MOIC, IRR, carry, personal cash-flow, hit rate, loss rate, median multiple, or value-add effect was located. Clinkle remains an adverse lead without a verified write-off; Legendary and Marvel remain transaction/board cases without reconstructed VC returns. These are bounded absences, not invitations to estimate. [researcher inference · insufficient public record] source coverage

Corrections made during adversarial review

• Reconciled conflicting Accel dates to “joined in 1987, partner by 1990,” while preserving 1985/1987/1990 source conflict. Profile formation
• Split Facebook discovery, pursuit, sponsor/board, company-building, fund capital, and trust investment; removed lone-attribution shorthand. People map
• Kept Facebook's 201,378,349 combined beneficial shares separate from 11,703,132 trust shares and aggregate IPO sales; no share count became personal return. Facebook control
• Calculated Circle's disclosed gross sale as shares × offer price while explicitly refusing to call it net proceeds or 100x proof. Circle control
• Separated Etsy's Accel and personal/Breyer Capital holdings and retained partner/team credit. Ledger
• Treated Walmart.com as an Accel-era joint venture; did not backdate Breyer Capital before July 2006. Profile boards section
• Treated Marvel as a board-compensation case, not a reconstructed VC investment; treated Legendary company/transaction values as insufficient for an investor multiple. Ledger
• Preserved IDG pre-2005 versus IDG-Accel post-2005 boundaries. Vehicle table
• Named Prosper—not Breyer—as the subject of the SEC order and used the case as a governance/regulatory lesson, not a misconduct allegation. Prosper lab
• Preserved different 2011 and 2025 >100x reported lists rather than harmonizing them. Forensics
• Normalized every material evidence label to one allowed attribution and one allowed support value. Label audit

[researcher inference · strong]

Corrected denominator, profile, and model re-audit

[researcher inference · strong]

Twenty-point acceptance matrix

Central-claim citation audit trail

Ordinary-claim citation audit trail

The sample was fixed at two ordinary claims from each of seven pre-review reader documents. These are not the central claims above.

Independent deal and team cross-checks

[researcher inference · strong] Six cases exceed the required five independent deal/team checks; issuer/regulator filings control roles, entities, securities, and legal outcomes over marketing shorthand.

Arithmetic and denominator checks

[researcher inference · strong] No arithmetic uses undisclosed cost basis, tax, carry, or allocation. Investment denominator

Citation, label, and link QA

Every root file begins with H1 followed immediately by a substantive H2; investments.md begins with the exact Pareto heading. The 18 notes each contain the required metadata line and six H2 sections. No full transcript is stored. [researcher inference · strong]

All detected material evidence labels conform to one allowed attribution—investor-stated, contemporaneous record, documented behavior, retrospective witness, or researcher inference—and one allowed support value—strong, moderate, provisional, or insufficient public record. [researcher inference · strong]

Local Markdown targets and heading anchors were mechanically checked. External spot checks covered the Facebook S-1/A, Circle 424B4, Walmart proxy, Prosper SEC order, HBS interview, PBS transcript route, Circle transcript, Etsy S-1/A, and official portfolio. Access boundaries in notes control where a server blocks direct retrieval or a transcript is partial. [researcher inference · strong]

Quotes are confined to the note layer and each is no more than 25 words; reader documents primarily paraphrase. Numbers are adjacent to filings or explicitly tagged reported. No company valuation, financing, acquisition consideration, affiliated sale, or board stock becomes an investor return. [researcher inference · strong]

Reader conclusion

Study Breyer for the combination of thematic preparation, simple-use discipline, adaptive-team selection, staged risk, institutional networks, recruiting, governance, and duration. Study the same record to learn why partner credit, vehicle boundaries, regulation, fund size, and power-law denominators are not footnotes—they determine what the success claim means. [researcher inference · strong] study guide

The analysis remains useful because it preserves uncertainty rather than concealing it. Facebook remains the dominant visible outcome; Circle is the strongest ex-Facebook filing control; Redback/Foundry remain reported; Accel VIII and Prosper remain adverse; and career-wide coverage and audited return remain unavailable. [researcher inference · strong]

Source-map boundary and retrieval priorities

This map separates direct voice, issuer filings, firm/vehicle claims, journalism, transaction controls, and skeptical evidence. Checkboxes mean processed into the package, not endorsement. Stars govern required parity: every ★★★ and ★★ writing or talk below has a matching note and source-by-source treatment. Company values, share counts, fund ownership, Breyer's personal economic interest, and realized cash are never interchangeable.

Official identity, institutions, archives, and vehicle boundaries

• ★★★ | Jim Breyer biography | https://breyercapital.com/jim-breyer/ | live 2026 | official biography | Current Breyer Capital role, stated interests, boards, education, and self-reported outcome claims.
• ★★★ | Breyer Capital portfolio | https://breyercapital.com/portfolio/ | live 2026 | official portfolio page | Reproducible 110-company official-page denominator after deduplicating headings and excluding two navigation headings; bounded research union is 122; explicitly says “some” past/present investments.
• ★★★ | Breyer Capital home | https://breyercapital.com/ | live 2026 | official website | Current featured portfolio and “over 40” successful offering/M&A claim.
• ★★★ | Stanford eCorner contributor archive | https://ecorner.stanford.edu/contributor/jim-breyer/ | migrated archive, live 2026 | official media index | Enumerates seven Stanford appearances and identifies mixed-speaker boundaries.
• ★★ | Stanford eCorner new home | https://stvp.stanford.edu/ecorner/ | live 2026 | official archive | Confirms migration route and current transcript host.
• ★★ | Facebook S-1/A | https://www.sec.gov/Archives/edgar/data/1326801/000119312512232582/d287954ds1a.htm | 2012-05-15 | SEC filing | Board since April 2005, career/vehicle roles, combined beneficial ownership, fund entities, and IPO sale schedule.
• ★★ | Walmart 2012 proxy | https://www.sec.gov/Archives/edgar/data/104169/000119312512163664/d264672ddef14a.htm | 2012-04-16 | SEC proxy | Kosmix ownership, proceeds, recusal, Facebook advertising relationship, and independence analysis.
• ★★ | Blackstone director biography | https://www.blackstone.com/people/james-w-breyer/ | live 2026 | official corporate biography | Current board role and institutional biography; title/dates require current qualification.
• ★ | Breyer Capital “Jim’s List” | https://breyercapital.com/resources/ | live 2026 | recommendation list | Intellectual tastes and quantum/AI reading leads, not investment evidence.

Priority direct writings and current firm doctrine

• ★★★ | Some Thoughts on AI Investing | https://medium.com/@jimbreyer/some-thoughts-on-ai-investing-765c839fde6 | 2019-10-29 | essay | Four direct filters and vertical-AI focus. Note: notes/w-ai-investing.md.
• ★★★ | Healthcare AI Thesis — Fall 2022 Update | https://medium.com/@jimbreyer/breyer-capitals-healthcare-ai-investment-thesis-learnings-and-predictions-december-2021-update-45e2e81b0c86 | page date 2021-12-01; updated title/body 2022 | essay | Portfolio breadth, healthcare workflow thesis, team and university sourcing. Note: notes/w-healthcare-ai-2022.md.
• ★★★ | Healthcare AI Thesis — Fall 2023 Update | https://medium.com/@jimbreyer/breyer-capitals-healthcare-ai-investment-thesis-learnings-and-predictions-fall-2023-update-a5f131a4e6b2 | 2023-09-06 | essay | Five focus areas and collaborator attribution. Note: notes/w-healthcare-ai-2023.md.
• ★★★ | Principles for the Future of Healthcare and Life Sciences | https://breyercapital.com/healthcare-thesis/ | live 2026 | firm thesis | Current Breyer/Cheatham co-led doctrine; not solely Jim-authored. Note: notes/w-healthcare-life-sciences-thesis.md.
• ★★★ | Investing in the Future of Fintech | https://medium.com/@jimbreyer/breyer-capital-investing-in-the-future-of-fintech-e33235cb7f35 | 2021-05-03 | essay | Global inclusion thesis, founder referrals, and IDG/portfolio/co-investor sourcing. Note: notes/w-fintech.md.
• ★★ | Introducing Paige.AI | https://medium.com/@jimbreyer/introducing-paige-ai-df06c459b5f3 | 2018-02-05 | investment announcement | Contemporaneous data/domain/team thesis. Note: notes/w-paige.md.
• ★★ | Introducing Subtle Medical | https://medium.com/@jimbreyer/introducing-subtle-medical-771d241646f7 | 2018-06-01 | investment announcement | Workflow-control-point and interdisciplinary-team rationale. Note: notes/w-subtle-medical.md.
• ★★ | Announcing Iterative Scopes | https://medium.com/@jimbreyer/announcing-iterative-scopes-5df61cad4de2 | 2021-08-04 | investment announcement | GI workflow, data, and trials thesis. Note: notes/w-iterative-scopes.md.
• ★★ | Why We Invested in Harbor | https://medium.com/@jimbreyer/why-we-invested-in-harbor-dd279821f63b | 2021-07-27 | investment announcement/update | Known operator, sequential scope, and undisclosed portfolio acquisition boundary. Note: notes/w-harbor.md.
• ★ | Introducing Breyer Capital Austin | https://medium.com/@jimbreyer | 2020-08-04 | profile-indexed post | Institutional geography lead; not priority decision doctrine and no standalone stable URL was retrievable.
• ★ | Breyer Capital Austin — End of Year Update | https://medium.com/@jimbreyer | 2020-12-17 | profile-indexed post | Team/geography lead; metadata only in this run.

Priority talks, interviews, and transcript routes

• ★★★ | FRONTLINE Dot Con interview | https://www.pbs.org/wgbh/frontline/wgbh/pages/frontline/shows/dotcon/interviews/breyer.html | interview 2001-05 | full web transcript | Bubble psychology, over-capitalization, staged risk, TCV, and failed AOL forecast. Note: notes/t-pbs-dot-con.md.
• ★★★ | Stanford eCorner: Investing in New Ideas | https://stvp.stanford.edu/wp-content/uploads/sites/3/2024/09/investing-in-new-ideas-transcript.pdf | 2005-10-26 | official transcript PDF | Contemporaneous Facebook/prepared-mind segment. Note: notes/t-stanford-investing-new-ideas.md.
• ★★ | Stanford eCorner: Listening Skills in an Entrepreneur | https://stvp.stanford.edu/wp-content/uploads/sites/3/2024/09/listening-skills-in-an-entrepreneur-transcript.pdf | 2005-10-26 | official transcript PDF | Founder learning and experimentation heuristic. Note: notes/t-stanford-listening.md.
• ★★ | Stanford eCorner: Defensibility Through Regulatory Constraints | https://stvp.stanford.edu/wp-content/uploads/sites/3/2024/09/defensibility-through-regulatory-constraints-transcript.pdf | 2006-12-01 | mixed-speaker transcript PDF | Breyer question plus Larsen answer; Prosper counterevidence. Note: notes/t-stanford-regulatory-defensibility.md.
• ★★ | Stanford eCorner: Managing the Burn Rate | https://stvp.stanford.edu/wp-content/uploads/sites/3/2024/09/managing-the-burn-rate-transcript.pdf | 2006-12-01 | mixed-speaker transcript PDF | Milestone burn and cash-buffer separation. Note: notes/t-stanford-burn-rate.md.
• ★★★ | HBS Alumni: One-on-One with Jim Breyer | https://www.alumni.hbs.edu/stories/Pages/story-bulletin.aspx?num=634 | 2008-03-01 | edited Q&A | Early-stage/team-over-strategy, fund mix, China, cycles, and carried-interest bias. Note: notes/t-hbs-one-on-one.md.
• ★★ | TechCrunch: Finding the Next Mark Zuckerberg | https://techcrunch.com/2017/11/08/famed-vc-jim-breyer-on-finding-the-next-mark-zuckerberg-and-much-more/ | 2017-11-08 | edited Q&A | Post-Accel university/AI sourcing. Note: notes/t-techcrunch-2017.md.
• ★★★ | Circle Money Movement: The Power of Courage | https://www.circle.com/the-money-movement/ep-86-the-power-of-courage | 2023-06-15 | official transcript/video | Circle, Facebook, recruiting, board practice, China, and current interdisciplinary thesis. Note: notes/t-circle-power-of-courage.md.
• ★★ | Forbes video: Digital Money and the AI Revolution | https://www.forbes.com/video/112022b0-a4e4-4fb0-a026-b49739ae30aa/ | 2025-09-29 | video plus public clips | Realized-return self-report and long-hold doctrine; partial transcript. Note: notes/t-forbes-2025.md.
• ★ | From Harvard to the Facebook | https://ecorner.stanford.edu/videos/from-harvard-to-the-facebook/ | 2005-10-26 | 57-minute podcast | Breyer interviews Zuckerberg; founder answers dominate, so it is supporting biography rather than priority Breyer doctrine.
• ★ | Career Advice for Students about Entrepreneurship | https://ecorner.stanford.edu/videos/career-advice-for-students-about-entrepreneurship/ | 2006-12-01 | mixed-speaker video | Archive lead; speaker-level transcript boundary was too weak for priority synthesis.
• ★ | The Emotional Roller Coaster | https://ecorner.stanford.edu/wp-content/uploads/sites/2/2006/12/1623.pdf | 2006-12-01 | mixed-speaker transcript PDF | Chris Larsen supplies nearly all substantive content; retained as Prosper context, not Breyer evidence.

Filings and transaction controls

• ★★★ | Circle IPO 424B4 | https://www.sec.gov/Archives/edgar/data/1876042/000119312525135795/d737521d424b4.htm | 2025-06-04 | SEC prospectus | Breyer affiliates' pre-offer shares, 3,355,468-share sale at $31, and retained shares; gross liquidity is not return.
• ★★ | Circle Schedule 13G | https://www.sec.gov/Archives/edgar/data/1537061/000095017025108279/xslSCHEDULE_13G_X01/primary_doc.xml | 2025-06-30 event | SEC ownership filing | James Breyer and trust reporting-person control after IPO.
• ★★ | Circle 2025 annual report | https://www.sec.gov/Archives/edgar/data/1876042/000187604226000116/a2025annualreport0331.pdf | 2026 | annual report | Current revenue/reserve economics and risk context; not venture-return evidence.
• ★★★ | Etsy S-1/A | https://www.sec.gov/Archives/edgar/data/1370637/000119312515114143/d806992ds1a.htm | 2015-03-31 | SEC filing | Separates Accel and Breyer/Breyer Capital holdings, sales, governance, and company metrics.
• ★★ | Brightcove S-1/A | https://www.sec.gov/Archives/edgar/data/1313275/000119312512040155/d200370ds1a.htm | 2012-02-06 | SEC filing | Breyer board role, company capital history, ownership, and IPO evidence.
• ★★★ | Prosper SEC cease-and-desist order | https://www.sec.gov/files/litigation/admin/2008/33-8984.pdf | 2008-11-24 | SEC order | Finds Prosper—not Breyer—sold unregistered securities from 2006 through 2008.
• ★★★ | Prosper S-1/A | https://www.sec.gov/Archives/edgar/data/1416265/000110465908074769/a08-29602_1s1a.htm | 2008-12-08 | SEC filing | State settlement, class action, quiet-period and operating/ownership facts.
• ★★ | Prosper 2011 post-effective amendment | https://www.sec.gov/Archives/edgar/data/1416265/000141626511000170/prosperposam63d2d11.htm | 2011 | SEC filing | Separates Breyer trust from Accel entities and disclaims excess beneficial ownership.
• ★★ | Spotify F-1 | https://www.sec.gov/Archives/edgar/data/1639920/000119312518092759/d494294df1.htm | 2018-03-20 | SEC filing | Direct-listing company history and ownership control; does not identify Breyer economics.
• ★★ | Foundry Networks merger proxy | https://www.sec.gov/Archives/edgar/data/1042797/000119312508215811/dprem14a.htm | 2008 | SEC merger filing | Company consideration and process control; not a fund multiple.
• ★★ | Facebook Form 4 | https://www.sec.gov/Archives/edgar/data/1194696/000120919112050884/xslF345X03/doc4.xml | 2012-10-25 | SEC ownership filing | Post-IPO share movement and entity attribution.

Deal history, performance reporting, and attribution controls

• ★★★ | Forbes: The Comeback Kid | https://www.forbes.com/2011/04/06/midas-list-11-jim-breyer-venture-capital-comeback-kid.html | 2011-04-06 | long-form profile | Accel VIII losses, IX fundraising, reported fund/deal returns, Efrusy/Chien sourcing, personal side-investment criticism, and China vehicle history.
• ★★★ | Forbes: Facebook Made Him a Billionaire…Circle | https://www.forbes.com/sites/kerryadolan/2025/09/30/facebook-made-him-a-billionaire-now-this-venture-capitalist-is-back-with-another-grand-slam/ | 2025-09-30 | long-form profile | Current Circle/Facebook/Redback/Foundry claims, family-office succession, healthcare shift, and self-reported economics.
• ★★ | Forbes: Departure from Facebook, Walmart and Dell | https://www.forbes.com/sites/luisakroll/2013/05/01/exclusive-billionaire-venture-capitalist-jim-breyer-explains-his-departure-from-facebook-wal-mart-and-dell/ | 2013-05-01 | interview/profile | Operating transition and early-stage focus; access partial due paywall.
• ★★ | Prosper 2015 Series D announcement | https://www.prosper.com/blog/prosper-marketplace-announces-165-million-series-d-financing-led-by-credit-suisse-next-investors | 2015-04-08 | company release | Breyer as returning investor after remediation; issuer-controlled outcome framing.
• ★★ | Breyer Capital portfolio: Facebook/Etsy/Legendary/Marvel/Circle claims | https://breyercapital.com/portfolio/ | live 2026 | official portfolio descriptions | Lead/board/round/acquisition claims requiring filing or independent controls.
• ★★ | S&P Global to acquire Kensho | https://press.spglobal.com/2018-03-06-S-P-Global-to-Acquire-Kensho-Bolsters-Core-Capabilities-in-Artificial-Intelligence-Natural-Language-Processing-and-Data-Analytics | 2018-03-06 | acquirer release | $550 million company consideration and founder continuity; not investor proceeds.
• ★ | Breyer Capital official featured portfolio | https://breyercapital.com/ | live 2026 | official marketing | Selected winners; useful survivorship-bias control when compared with the full 110-name page.

Skeptical evidence, failures, context, and completed peers

• ★★★ | Accel VIII/IX adverse and credit-attribution record | https://www.forbes.com/2011/04/06/midas-list-11-jim-breyer-venture-capital-comeback-kid.html | 2011-04-06 | journalism | LP losses, fund-size/fee conflict, missed Google, Efrusy credit, and off-record side-investment criticism.
• ★★★ | Prosper legal proceedings and registration failure | https://www.sec.gov/Archives/edgar/data/1416265/000110465908074769/a08-29602_1s1a.htm | 2008-12-08 | issuer filing | Named adverse case with board-era chronology, settlement accrual, and claimant allegations.
• ★★ | Walmart Kosmix/Facebook conflict review | https://www.sec.gov/Archives/edgar/data/104169/000119312512163664/d264672ddef14a.htm | 2012 | proxy | Documented recusal and limited economic interest; a control, not misconduct evidence.
• ★★ | Trae Stephens completed peer | ../099-trae-stephens/investmentphilosophy.md | updated 2026-08-19 | canon package | Current founder/mission, sponsor-led IC, vehicle, conflict, and ex-outlier comparator.
• ★★ | Stephens underlying 20VC process interview | https://www.thetwentyminutevc.com/trae-stephens | 2024-04-02 | direct interview | Explicit upside cases, sponsor-led process, calibration, and follow-on doctrine.
• ★★ | Martin Casado completed peer | ../100-martin-casado/investmentphilosophy.md | updated 2026-08-20 | canon package | Technical specialist, workflow/economics, coauthorship, vehicle, and denominator comparator.
• ★★ | Casado/Levine: Market Annealing | https://a16z.com/market-annealing-getting-to-10m-arr-in-very-early-markets/ | 2023 | primary coauthored essay | Evidence-gated GTM comparison and coauthorship attribution.
• ★★ | Georges Doriot completed peer | ../001-georges-doriot/investmentphilosophy.md | updated 2026-07-26 | canon package | Long-duration governance and one-outlier portfolio comparison; era mismatch is explicit.
• ★ | Venture Capital Greats: James Swartz oral history | https://digicoll.lib.berkeley.edu/record/103633/files/swartz_james_donated.pdf | archival transcript | Accel institutional history and Breyer-led China entry lead; not deeply processed in this run.
• ★ | The Power Law by Sebastian Mallaby | https://www.penguinrandomhouse.com/books/565570/the-power-law-by-sebastian-mallaby/ | 2022 | book metadata | Secondary cross-VC context; full text not accessed, so no content claim is derived from it.

Coverage assessment

The discovery battery covered “Jim/James W. Breyer” with Accel, Breyer Capital, Facebook, Circle, IDG/IDG-Accel China, writings, transcript, podcast, 13F/SEC, returns, failures, criticism, lawsuit, LP, underperformance, and decade terms 1990–2026. Direct routes attempted were the Breyer Capital site, Medium author index, Stanford eCorner/STVP archive, PBS, HBS, Circle, Forbes video/clip pages, SEC EDGAR issuer filings, acquirer releases, and the public Swartz oral-history catalog.

The priority written denominator is nine full items: nine full, zero partial, zero inaccessible, nine deeply analyzed. Two profile-indexed Austin posts are ★ metadata leads and intentionally outside required parity. The priority spoken denominator is nine items: eight full/full-segment and one partial Forbes corpus; all nine have notes and source-by-source analysis. Mixed-speaker Stanford segments are bounded to identifiable Breyer speech.

The official Breyer Capital page contains 110 unique company headings after case-insensitive exact deduplication and exclusion of “Contact Us” and “Featured Companies.” The bounded union adds exactly 12 materially researched names absent from that page—Brightcove, Foundry Networks, Kosmix, Redback Networks, RealNetworks, Paige, Harbor, 21st Century Oncology, Clinkle, Rilla, Twenty, and LegalZoom—for 122 listed scoped names. All 25 researched-ledger lineages resolve into that union; seven are sufficiently reconstructed and 115 remain unresolved, so the bounded reconstruction ratio is 7 / 122 = 5.738%, rounded 5.7%. Exact 122-name inventory

The official page expressly presents only “some” past and present investments and is vehicle-level, not a Jim-only career ledger. Additional undisclosed, stealth, passed, and written-off holdings make career-wide coverage fraction: not computable. No audited career/fund DPI, TVPI, MOIC, IRR, complete loss ledger, carry allocation, or personal cash-flow schedule was found.

Highest-value future retrieval is Accel LP reports by vintage; Accel IX/VIII partnership and allocation documents; Facebook 2005 IC/board records; full Circle financing/cap-table history; Foundry/Redback security-level cash flows; Breyer Capital IC and loss/write-off ledger; Prosper board/counsel minutes; and IDG-Accel fund attribution by partner and vehicle. No archive or person was contacted.

Source notes