Public dossier section · @ben

history

This is @ben's profile revision 1. The original date and editorial confidence below are attributed to its publisher, not independent verification.

Original credited author: Hraness

AI-drafted at Ben Guo's direct request and credited to Hraness; every claim links to its cataloged source.

Back to the full profile · Markdown version · Original document

Published text

core automation: history

the company, sourced and dated — the openai years, the january exit, the april launch, the round that kept changing, and the lab in public.

the openai years (2015–2025)

amsterdam quant finance, then seven years at openai — the rubik's-cube hand, codex and the first copilot model, and the q*/strawberry line that became o1, o3, and gpt-5's reasoning.

2015 — the amsterdam quant years: tworek spends about five years in amsterdam quant finance — linkedin lists him as senior quantitative researcher at independent view (jan 2015–dec 2018), 'most senior researcher in the company reporting directly to cio,' building fully systematic futures signals. earlier details are thinner: an msc in mathematics from the university of warsaw, and a dropout story he has told on podcasts. (reported; sources: jerry tworek — linkedin, jerry tworek speaker bio (warsaw.ai), how gpt-5 thinks — mad podcast, october 2025)
2019-02 — tworek joins openai: tworek joins openai — about thirty employees at the time — and stays nearly seven years, ending as vp of research. (confirmed; sources: million integrals — tworeks homepage, jerry tworek — linkedin, he left openai to think bigger — core memory ep. 53)
2019-10 — the rubik's-cube hand: 'solving rubik's cube with a robot hand' — the sim2real breakthrough tworek ran, 'one of core breakthroughs' by his own account, landing at the neurips 2019 deep rl workshop. (confirmed; sources: solving rubiks cube with a robot hand (arxiv 1910.07113), million integrals — tworeks homepage)
2021 — codex, copilot, humaneval: the codex line: tworek 'conducted all stages of research and trained the model that powered github copilot,' was a primary contributor to codex, and hand-wrote about thirty humaneval problems — wojciech zaremba wrote thirty more. (confirmed; sources: million integrals — tworeks homepage, jerry tworek deep-dive (agi house))
2024-09-12 — o1: o1 ships — the reasoning model from the effort known internally as q*, then strawberry. tworek: 'i've lead the team that researched training llms to reason with reinforcement learning.' (confirmed; sources: million integrals — tworeks homepage, jerry tworek on o1 reasoning (fast company), he left openai to think bigger — core memory ep. 53)
2025-04 — o3, and the vp-of-research seat: o3 ships. tworek's later summary of the o1→o3 gap: '99% systems, 1% algorithms.' he is vp of research by february 2025. (confirmed; sources: million integrals — tworeks homepage, jerry tworek deep-dive (agi house), jerry tworek — linkedin)
2025-08 — gpt-5: gpt-5 ships with reasoning descended from the o3 line — tworek later describes it on the mad podcast as roughly 'o3.1.' (confirmed; sources: million integrals — tworeks homepage, how gpt-5 thinks — mad podcast, october 2025)
2025-08 — the bench starts leaving: the future cofounders disperse. julia villagra — promoted to openai chief people officer in march — exits in august, with reuters reporting she left to pursue 'art, music and storytelling to help people understand the transition to artificial general intelligence.' joanne jang founds openai labs in september and lands on the time100 ai list in august. (reported; sources: openai chief people officer julia villagra leaving (business insider), openai leadership updates, march 2025, joanne jang — linkedin, joanne jang profile (prometheus))
2025-10-16 — the last openai interview: tworek records 'how gpt-5 thinks' on the mad podcast while still vp of research — the last big interview of the openai years. (confirmed; sources: how gpt-5 thinks — mad podcast, october 2025)

the exit and the founding (2026-01 – 2026-03)

tworek announces his exit, explains it on core memory, and the information names the new company: core automation, seeking up to $1b for models that keep learning after deployment.

2026-01-05 — tworek leaves openai: tworek announces his openai departure on x after ~7 years, telling his team he wants 'to try and explore types of research that are hard to do at openai.' the verge and the decoder cover it the same day; the decoder reads the line as a 'not-so-subtle dig.' (confirmed; sources: openai loses top ai researcher jerry tworek after seven years (the decoder), openais research lead is departing after nearly seven years (the verge), openais vp researcher jerry tworek resigns (times of india))
2026-01-21 — the exit interview: core memory ep53, recorded days after the announcement: high-risk pioneering research had become hard at a 'conservative' openai; big labs as a whole have gone conservative; deep-learning research at big labs 'is done.' (confirmed; sources: he left openai to think bigger — core memory ep. 53, google caught up because openai stumbled (the decoder))
2026-01 — the information names the company: the information reports the new company — named core automation — is seeking $500m–$1b to build models needing ~100× less data that learn continuously; the reported long-term roadmap runs from industrial automation to 'self-replicating factories,' bio-machines, and 'planetary terraforming.' the piece is paywalled; the record runs through the decoder's summary. (reported; sources: ex-openai researchers startup targets $1b funding (the information), two startups want to replace how ai learns (the decoder))
2026-03 — incorporated: core automation, inc. is founded — a delaware corporation incorporated in 2026 per its form d; 'late march' per dealroom. headquarters at 340 brannan street, suite 301, san francisco. (reported; sources: core automation, inc. — form d (sec), core automation eyes $300m–$500m raise at $4b valuation (dealroom note))

the launch (2026-04)

a ~$100m safe raise at ~$1b reportedly closes, and on april 21 the lab goes public — 'the world's most automated ai lab' — with a coordinated cofounder reveal.

2026-04 — the ~$100m safe round: the initial ~$100m raise reportedly closes at ~$1b post-money — press and trackers name nvidia, spark capital, and accel, with scribble ventures and threshold ventures in secondary listings and no disclosed lead. the form d later shows the money came in on safes that converted into $135.3m of stock in the july round — the discrepancy between the reported ~$100m and the $135.3m conversion stays unresolved. (reported; sources: core automation company profile (sacra), core automation funding profile (ai market watch), core automation funding profile (vcbacked), core automation, inc. — form d (sec))
2026-04-21 — launch day: public launch. the first x post: 'today we're announcing core automation… our objective: systems that optimize and automate work, starting with research itself' — 'the world's most automated ai lab.' tworek's own post: 'small step for a finger on "post tweet" button. big step for millions of future ai agents doing useful work for us.' cofounder avery lamp puts it on hacker news, where it gets 1 point and 0 comments. (confirmed; sources: core automations first x post (texxr mirror), tworeks launch post (texxr mirror), coreauto.com — company site, core automation launch submission by averylamp)
2026-04-21 — the nerdsniped cascade: the coordinated reveal. rohan anil posts that he left anthropic 'a few weeks ago' — 'jerry tworek nerdsniped me into starting this with him and others.' anmol gulati announces from google deepmind: 'the current research paradigm — scaling models, data, and static deployment won't get us all the way.' joanne jang's x bio flips to 'trying to automate myself @coreautoai.' villagra, lamp, amid, and duvvuri are named in the coverage. (confirmed; sources: new ai lab core automation nerdsniped researchers from anthropic, google deepmind (business insider), anmol gulatis launch-day post (texxr mirror), x bios mirror (digg), the nerdsniped launch story, yahoo syndication)
2026-04-22 — the launch coverage: the launch story travels: business insider's 'nerdsniped' piece becomes the most-syndicated account, the decoder files its launch writeup placing the lab in the neo-lab cohort with thinking machines and ssi, and korea's mk frames it as 'openai mafia' news while noting no business model has been disclosed. (confirmed; sources: new ai lab core automation nerdsniped researchers from anthropic, google deepmind (business insider), ex-openai researcher jerry tworek launches core automation (the decoder), the openai mafia frame (mk korea), the nerdsniped launch story, yahoo syndication)

the round that kept changing (2026-05 – 2026-08)

the funding stays messy in public: a reported $300m–$500m seek at ~$4b, then a july 30 form d showing $635.2m authorized and $432.1m sold to 51 investors — including $135.3m of converted safes.

2026-05-07 — $1b to $4b in six weeks: the information reports the ~6-week-old company is already seeking $300m–$500m at ~$4b. palazzolo on x: 'back to back rounds are continuing in neolab land!' the report is paywalled — the record runs through techmeme, dealroom, and testingcatalog. (reported; sources: the information reports $300m–$500m at ~$4b (techmeme record), core automation eyes $300m–$500m raise at $4b valuation (dealroom note), ex-openai researchers ai startup hits $4b valuation in 6 weeks (testingcatalog))
2026-07-15 — the equity round opens: first sale of the new equity offering, per the form d filed two weeks later. (confirmed; sources: core automation, inc. — form d (sec), core automation form d record (takeoff radar))
2026-07-30 — the form d: the form d lands: a $635,249,754 authorized offering with $432,125,053 sold to 51 investors and $203,124,701 remaining — including $135,252,607 issued on conversion of the outstanding safes. tracker math (forge/pitchbook-derived) puts the round at ~$3.2b pre / ~$3.62b post — reported, never confirmed by the company or the filing. (confirmed; sources: core automation, inc. — form d (sec), core automation, inc. filings (filingflow), core automation form d record (takeoff radar), form d write-up (edgar.tools), core automation company profile (silicon valley investclub — conflates the 2001 namesake))
2026-08-13 — axios reads the filing: axios pro reads the filing as the company 'looking to raise up to $500 million in new funding' — consistent with the remaining capacity plus new cash — and describes core automation as 'an ai continual learning model developer.' (reported; sources: core automation is raising $500m (axios pro), core automation $500m round listing (gaebler))
2026-08-17 — the alumni ventures feeder: a second filing surfaces: 'avsf – core automation 2026, llc,' an alumni ventures feeder spv — $2,109,250 sold to 100 investors, first sale august 3. (confirmed; sources: avsf – core automation 2026, llc — form d (sec))

the lab in public (2026-05 – 2026-09)

the lab in public: the blog corpus, the agi house keynote, the sequoia and mts episodes, the one layer deeper competition with tilde research, and the agm internship.

2026-05-27 — the blog opens: the first company blog posts appear — saroufim's 'to automate publishing, first make authoring boring' and 'when ai starts writing systems code,' arguing that automating research means automating systems code first. (confirmed; sources: the core automation blog)
2026-07-18 — the agi house keynote: tworek keynotes the agi house auto-research summit in hillsborough — 'the era of evals is dumb,' 'the only test is the real world' — and puts up an open-problems slide that eighteen teams build on for the rest of the day. (confirmed; sources: jerry tworek deep-dive (agi house), auto-research summit recap (agi house), auto-research is moving from benchmarks to real-world loops (agi house))
2026-07-28 — the pacing letter: the 'pacing the frontier' open letter goes public — 1,386 frontier-ai employees asking for 'technical and governance tools needed to deliberately pace the frontier of automated ai development.' core automation employees are among the signatories per the hn discussion — the lab that automates research, signing the letter about pacing it. (reported; sources: pacing the frontier — open letter, the pacing the frontier hn thread)
2026-07-29 — the sequoia episode: sequoia's training data episode airs — 'building the automated agi lab: core automation's jerry tworek and rohan anil' — the fullest public statement of the thesis: transformers 'incapable of continual learning,' the search for a replacement architecture, and the most automated lab as the instrument. sequoia hosting is not evidence sequoia invested — the round's participants remain undisclosed. (confirmed; sources: building the automated agi lab — tworek and anil (sequoia training data), building the automated agi lab — tworek and anil (getpodcast record), building the automated agi lab — timestamped transcript (teahose), sequoia promo post quoting tworek (linkedin))
2026-08-03 — one layer deeper: 'one layer deeper' launches with tilde research — a public competition on repeated modular squaring: learn the computation, then run longer chains than seen in training. the hard track gives one h100, one hour of training, and asks models to certify depth t=1…64. (confirmed; sources: the core automation blog, tilde research — linkedin)
2026-08-25 — twenty-something people: the fysk showcase primer reports the team at '20-something people' — drawn largely from openai and google deepmind plus anthropic, meta, xai, and amazon's agi lab — with a plan to reach ~40 by year-end. linkedin-derived counts say ~21; one aggregator says 14. (reported; sources: your showcase primer: core automation (fysk), core automation — linkedin company page, core automation company profile (silicon valley investclub — conflates the 2001 namesake))
2026-08-26 — the mts episode: 'ai labs are simply cooked' — the mts episode: three generations of ai labs, a two-year horizon for agents automating research, and the concession that today's agents 'lack deep field understanding and produce low-quality, high-creativity ideas.' (confirmed; sources: ai labs are simply cooked — jerry tworek (mts))
2026-09-02 — the architecture-discovery post: ehsan amid's post on neural architecture discovery — untrained networks already separate cats from dogs; automate the discovery of architectures — and on automation reshaping the field. (confirmed; sources: the core automation blog, ehsan amids research site)
2026-09-09 — the competition results: 'what the models learned in one layer deeper' — the results: 15,602 submissions; only 18 hard uploads from 5 accounts certified depth (16 reached t=64). the winners learned 'wiring around python primitives' — evidence that test-time depth is genuinely hard for current methods. (confirmed; sources: the core automation blog, sean mcleish — personal site)
2026-09-10 — anil on the bottleneck: rohan anil on the information bottleneck — 'continual learning is the next bottleneck': the pre-training/rl split as 'organizational convenience rather than science,' shampoo's origin, the qr-factorization kernel competition (~60× over cusolver), and how he met tworek at dolores park. (confirmed; sources: continual learning is the next bottleneck — rohan anil (the information bottleneck), rohan anil bio (teahose))
2026-09 — the agm program: the agm — 'associate general manager' — internship goes live: four interns, january–march 2027 in sf, $10,000/month plus profit share, each given a budget, tokens, and the company's agent infrastructure to run real online businesses. 'by watching where people intervene, we want to understand what agents still need to learn.' applications close october 6, 2026. (confirmed; sources: associate general manager program (coreauto.com))
2026-09-16 — the showcase: tworek appears at the fysk sf startup showcase — the lab's fifth month ends the way it started: the founders in public. (reported; sources: your showcase primer: core automation (fysk))

AI-drafted at Ben Guo's direct request and credited to Hraness; every claim links to its cataloged source.

Canonical page: https://hraness.com/core-automation/history

Original source occurrences

auto-research is moving from benchmarks to real-world loops (agi house) · agihousecommunity.beehiiv.com · 2026-07-23 (day) · original ID: src-ci-agihouse-beehiiv · Other public citations of this URL
"there is no way to outsource your understanding"; verification-bottleneck framing
core automation is raising $500m (axios pro) · axios · 2026-08-13 (day) · original ID: src-ci-axios-500m · Other public citations of this URL
"raising up to $500m per sec filing"; "ai continual learning model developer"
core automation eyes $300m–$500m raise at $4b valuation (dealroom note) · app.dealroom.co · 2026-05-08 (day) · original ID: src-ci-dealroom-4b · Other public citations of this URL
$300–500m at ~$4b; $100m at $1b closed weeks prior; investors nvidia/spark/accel; cofounders incl. villagra; "isn't close to releasing" a product
core automation launch submission by averylamp · hacker news · 2026-04-21 (day) · original ID: src-ci-hn-launch · Other public citations of this URL · avery lamp
submission by cofounder averylamp; 1 point, 0 comments
the pacing the frontier hn thread · hacker news · 2026-07-28 (day) · original ID: src-ci-hn-pacing · Other public citations of this URL
"pacing the frontier" discussion; 0xdeafbead cites core automation signatories vs. regulatory-capture theory
continual learning is the next bottleneck — rohan anil (the information bottleneck) · the-information-bottleneck.com · 2026-09-10 (day) · original ID: src-ci-infobottleneck-anil · Other public citations of this URL
anil interview; pretrain/rl split as "organizational convenience"; kernel-competition + hiring segments; dolores park origin
core automation — linkedin company page · linkedin · 2026-09 (month) · original ID: src-ci-linkedin · Other public citations of this URL
~21 employees; key-executive list; sf hq; founded 2026
the openai mafia frame (mk korea) · mk.co.kr · 2026-04-22 (day) · original ID: src-ci-mk-openai-mafia · Other public citations of this URL
"openai mafia" frame; o1/o3 provenance; no disclosed business model
ai labs are simply cooked — jerry tworek (mts) · snipd · 2026-08-26 (day) · original ID: src-ci-mts-podcast · Other public citations of this URL
"ai labs are simply cooked"; three-generations model; 2-year horizon; agents' taste deficit; "company brain"
pacing the frontier — open letter · pacingthefrontier.com · 2026-07 (month) · original ID: src-ci-pacing-frontier · Other public citations of this URL
1,386 frontier-ai employees; requests usg-backed international pacing tools for automated ai development
core automation company profile (sacra) · sacra · 2026-09 (month) · original ID: src-ci-sacra · Other public citations of this URL
funding recap; monetization speculation; paradigm-timing risk analysis
building the automated agi lab — tworek and anil (sequoia training data) · youtube · 2026-07-29 (day) · original ID: src-ci-sequoia-podcast · Other public citations of this URL · sonya huang and pat grady
tworek+anil; transformer-ceiling thesis; codex ~20-min compaction; kernel-generation automation; hosts huang & grady
core automation form d record (takeoff radar) · takeoffradar.com · 2026-07-30 (day) · original ID: src-ci-takeoffradar-formd · Other public citations of this URL
form d: $635,249,754 target; $432,125,053 raised; first sale 2026-07-15
the information reports $300m–$500m at ~$4b (techmeme record) · techmeme.com · 2026-05-07 (day) · original ID: src-ci-techmeme-4b · Other public citations of this URL
the information $4b report headline + palazzolo "neolab land" quote
ex-openai researchers ai startup hits $4b valuation in 6 weeks (testingcatalog) · testingcatalog.net · 2026-05-08 (day) · original ID: src-ci-testingcatalog-4b · Other public citations of this URL
relays the information; single-phase training; "self-replicating factory"/"biological machines" roadmap
openais vp researcher jerry tworek resigns (times of india) · timesofindia.indiatimes.com · 2026-01 (month) · original ID: src-ci-timesofindia · Other public citations of this URL
exit memo verbatim; chinchilla/rl-on-robots history
openais research lead is departing after nearly seven years (the verge) · the verge · 2026-01-05 (day) · original ID: src-ci-verge-departure · Other public citations of this URL
tworek exit; memo/x post quoted
the nerdsniped launch story, yahoo syndication · finance.yahoo.com · 2026-04-22 (day) · original ID: src-ci-yahoo-nerdsniped · Other public citations of this URL
bi syndication; adds ehsan amid, avery lamp, julia villagra, sai surya duvvuri
auto-research summit recap (agi house) · blog.agihouse.org · 2026-07 (month) · original ID: src-hp-agihouse-recap · Other public citations of this URL
auto-research summit recap; open-problems slide; 18 teams
jerry tworek deep-dive (agi house) · blog.agihouse.org · 2026-07-29 (day) · original ID: src-hp-agihouse-tworek · Other public citations of this URL
deep-dive on tworek keynote/interview — evals, o1→o3, continual learning, founding
core automation funding profile (ai market watch) · ai-market-watch.com · 2026 (year) · original ID: src-hp-aimarketwatch · Other public citations of this URL
~$532m total claim; jul ~$432m at ~$3.6b post (pitchbook/forge-derived); scribble/threshold also listed
ehsan amids research site · sites.google.com · 2026-09-16 (day) · original ID: src-hp-amid-site · Other public citations of this URL · ehsan amid
amid bio: ucsc phd 2020 (warmuth), aalto msc, google brain oct 2021–2026 gemini core, joined core apr 2026
rohan anil bio (teahose) · teahose.com · 2026 (year) · original ID: src-hp-anil-teahose · Other public citations of this URL
anil bio: distinguished engineer deepmind, 1-of-4 gemini pretraining leads, shampoo, anthropic
new ai lab core automation nerdsniped researchers from anthropic, google deepmind (business insider) · business insider · 2026-04-22 (day) · original ID: src-hp-bi-nerdsniped · Other public citations of this URL
launch-day roster: anil "nerdsniped" quote; gulati quote; jang, villagra, lamp, amid, duvvuri named
openai chief people officer julia villagra leaving (business insider) · business insider · 2025-08 (month) · original ID: src-hp-bi-villagra · Other public citations of this URL
villagra's openai exit; reuters' "art, music and storytelling…agi" line
associate general manager program (coreauto.com) · coreauto.com · 2026-09-16 (day) · original ID: src-hp-coreauto-agm · Other public citations of this URL
associate general manager internship page — jan–mar 2027, $10k/mo, apply by oct 6, 2026
the core automation blog · coreauto.com · 2026-05-27 (day) · original ID: src-hp-coreauto-blog-data · Other public citations of this URL · core automation
company blog corpus: five 2026 posts — saroufim on authoring and systems code, the one layer deeper launch and results, amid on neural architecture discovery
coreauto.com — company site · coreauto.com · 2026-09-16 (day) · original ID: src-hp-coreauto-site · Other public citations of this URL
company homepage — "world's most automated ai lab," research/approach/vision/team copy
he left openai to think bigger — core memory ep. 53 · corememory.com · 2026-01-21 (day) · original ID: src-hp-corememory-ep53 · Other public citations of this URL · ashlee vance
exit interview; q*/strawberry/o1 lineage; big labs "conservative"
openai loses top ai researcher jerry tworek after seven years (the decoder) · the-decoder.com · 2026-01-05 (day) · original ID: src-hp-decoder-departure · Other public citations of this URL
tworek exits openai; quote "explore types of research that are hard to do at openai"
google caught up because openai stumbled (the decoder) · the-decoder.com · 2026-01 (month) · original ID: src-hp-decoder-google · Other public citations of this URL
core memory recap: agi ~2029; continual learning "one of the last important building blocks"; google catch-up framing
ex-openai researcher jerry tworek launches core automation (the decoder) · the-decoder.com · 2026-04-22 (day) · original ID: src-hp-decoder-launch · Other public citations of this URL
launch coverage; "deep learning research 'is done'"; neo-lab framing
two startups want to replace how ai learns (the decoder) · the-decoder.com · 2026-01-29 (day) · original ID: src-hp-decoder-two-startups · Other public citations of this URL
pairs core automation w/ flapping airplanes; 100× less data goal; "self-replicating factories…planetary terraforming"
x bios mirror (digg) · digg · 2026 (year) · original ID: src-hp-digg-bios · Other public citations of this URL
mirror of x bios (tworek ceo bio; jang "trying to automate my work @coreautoai"; gulati incl. unverified "cofounder adept ai")
form d write-up (edgar.tools) · app.edgar.tools · 2026-07-30 (day) · original ID: src-hp-edgartools-news · Other public citations of this URL
filing-derived write-up of the form d figures
jerry tworek on o1 reasoning (fast company) · fast company · 2024-09 (month) · original ID: src-hp-fastcompany-o1 · Other public citations of this URL
o1-era tworek quotes on reasoning
core automation, inc. filings (filingflow) · filingflow.app · 2026-07-30 (day) · original ID: src-hp-filingflow · Other public citations of this URL
form d mirrors/timeline
your showcase primer: core automation (fysk) · newsletter.foundersysk.com · 2026-08-25 (day) · original ID: src-hp-fysk-primer · Other public citations of this URL
"20-something people," ~40 by year-end; lamp = fysk lp; sept 16 showcase
core automation $500m round listing (gaebler) · gaebler.com · 2026-08-13 (day) · original ID: src-hp-gaebler · Other public citations of this URL
labels the filing a "$500m round" (loose reading)
anmol gulatis launch-day post (texxr mirror) · texxr.com · 2026-04-22 (day) · original ID: src-hp-gulati-x · Other public citations of this URL · anmol gulati
mirror of launch-day post
ex-openai researchers startup targets $1b funding (the information) · the information · 2026-01-28 (day) · original ID: src-hp-information-jan · Other public citations of this URL
first naming of core automation; $500m–$1b seek (paywalled; via decoder)
joanne jang — linkedin · linkedin · 2026-09-16 (day) · original ID: src-hp-jang-linkedin · Other public citations of this URL
openai gm dec 2021–apr 2026; labs/model-behavior history; time100 post
how gpt-5 thinks — mad podcast, october 2025 · youtube · 2025-10-16 (day) · original ID: src-hp-mad-podcast · Other public citations of this URL
"how gpt-5 thinks" — bio chapters (poland, dropping out, trading); "gpt-5 ≈ o3.1"
sean mcleish — personal site · mcleish7.github.io · 2026-09-16 (day) · original ID: src-hp-mcleish-site · Other public citations of this URL
sean mcleish (co-author, sep 9 results post) — tilde research intern
million integrals — tworeks homepage · millionintegrals.com · 2026-09-16 (day) · original ID: src-hp-millionintegrals · Other public citations of this URL · jerry tworek
tworek's self-authored work history (copilot, codex, o1, o3, gpt-5, chatgpt agent)
openai leadership updates, march 2025 · openai · 2025-03 (month) · original ID: src-hp-openai-leadership · Other public citations of this URL
openai announcement of villagra's cpo promotion
joanne jang profile (prometheus) · prometheusroot.com · 2026 (year) · original ID: src-hp-prometheus-jang · Other public citations of this URL
model spec stewardship; oai labs; reservoir samples; jony ive adjacency (reported)
solving rubiks cube with a robot hand (arxiv 1910.07113) · arxiv.org · 2019-10 (month) · original ID: src-hp-rubiks-paper · Other public citations of this URL
"solving rubik's cube with a robot hand" (tworek co-author)
avsf – core automation 2026, llc — form d (sec) · sec · 2026-08-17 (day) · original ID: src-hp-sec-formd-avsf · Other public citations of this URL
"avsf – core automation 2026, llc" — alumni ventures feeder spv, $2,109,250, 100 investors, first sale 2026-08-03
core automation, inc. — form d (sec) · sec · 2026-07-30 (day) · original ID: src-hp-sec-formd-coreauto · Other public citations of this URL
form d: $635,249,754 offering; $432,125,053 sold; $135,252,607 safe conversion; 51 investors; first sale 2026-07-15; officers/directors; 340 brannan st sf
sequoia promo post quoting tworek (linkedin) · linkedin · 2026-07-31 (day) · original ID: src-hp-sequoia-linkedin · Other public citations of this URL · sequoia capital
sequoia promo quoting tworek's automation-agency framing
core automation company profile (silicon valley investclub — conflates the 2001 namesake) · siliconvalleyinvestclub.com · 2026-07 (month) · original ID: src-hp-svinvestclub · Other public citations of this URL
caplight/forge-sourced table: apr seed $100m @ $1b post; jul $432.125m @ $3.62b post
building the automated agi lab — timestamped transcript (teahose) · teahose.com · 2026-07-29 (day) · original ID: src-hp-teahose-transcript · Other public citations of this URL · sonya huang and pat grady
timestamped verbatim quotes from the episode
core automations first x post (texxr mirror) · texxr.com · 2026-04-21 (day) · original ID: src-hp-texxr-coreauto · Other public citations of this URL · core automation
mirror of company's first x post ("today we're announcing core automation…")
tworeks launch post (texxr mirror) · texxr.com · 2026-04-21 (day) · original ID: src-hp-texxr-tworek · Other public citations of this URL · jerry tworek
tworek's launch tweet ("small step for a finger on 'post tweet' button…")
tilde research — linkedin · linkedin · 2026-09-16 (day) · original ID: src-hp-tilde-linkedin · Other public citations of this URL · tilde research
tilde research — 12-person moonshot lab (dhruv pai, mason wang); competition co-host
building the automated agi lab — tworek and anil (getpodcast record) · getpodcast.com · 2026-07-29 (day) · original ID: src-hp-training-data · Other public citations of this URL · sonya huang and pat grady
sequoia training data ep w/ tworek & anil
jerry tworek — linkedin · linkedin · 2026-09-16 (day) · original ID: src-hp-tworek-linkedin · Other public citations of this URL
openai roles/dates; independent view (amsterdam) 2015–2018
core automation funding profile (vcbacked) · vcbacked.co · 2026 (year) · original ID: src-hp-vcbacked · Other public citations of this URL
$100m seed, apr 2026; accel/nvidia/spark
jerry tworek speaker bio (warsaw.ai) · warsaw.ai · 2020 (year) · original ID: src-hp-warsawai · Other public citations of this URL
bio: msc mathematics univ. of warsaw; "first five years of his career in the hedge fund industry"

Structured research records

the openai years · core-automation-eras · 2015–2025 (approximate)

Original location

Original structured fields

{
  "id": "prehistory",
  "range": "2015–2025",
  "summary": "amsterdam quant finance, then seven years at openai — the rubik's-cube hand, codex and the first copilot model, and the q*/strawberry line that became o1, o3, and gpt-5's reasoning.",
  "title": "the openai years"
}

the exit and the founding · core-automation-eras · 2026-01 – 2026-03 (approximate)

Original location

Original structured fields

{
  "id": "departure",
  "range": "2026-01 – 2026-03",
  "summary": "tworek announces his exit, explains it on core memory, and the information names the new company: core automation, seeking up to $1b for models that keep learning after deployment.",
  "title": "the exit and the founding"
}

the launch · core-automation-eras · 2026-04 (month)

Original location

Original structured fields

{
  "id": "launch",
  "range": "2026-04",
  "summary": "a ~$100m safe raise at ~$1b reportedly closes, and on april 21 the lab goes public — 'the world's most automated ai lab' — with a coordinated cofounder reveal.",
  "title": "the launch"
}

the round that kept changing · core-automation-eras · 2026-05 – 2026-08 (approximate)

Original location

Original structured fields

{
  "id": "round",
  "range": "2026-05 – 2026-08",
  "summary": "the funding stays messy in public: a reported $300m–$500m seek at ~$4b, then a july 30 form d showing $635.2m authorized and $432.1m sold to 51 investors — including $135.3m of converted safes.",
  "title": "the round that kept changing"
}

the lab in public · core-automation-eras · 2026-05 – 2026-09 (approximate)

Original location

Original structured fields

{
  "id": "operating",
  "range": "2026-05 – 2026-09",
  "summary": "the lab in public: the blog corpus, the agi house keynote, the sequoia and mts episodes, the one layer deeper competition with tilde research, and the agm internship.",
  "title": "the lab in public"
}

the amsterdam quant years · core-automation-timeline · 2015 (year) · reported (original editorial confidence)

Original location · source-b8207e74cbd595819904 · source-1dd5b02515323c524278 · source-fb717d0f7bffeb9549b2

Original structured fields

{
  "confidence": "reported",
  "date": "2015",
  "detail": "tworek spends about five years in amsterdam quant finance — linkedin lists him as senior quantitative researcher at independent view (jan 2015–dec 2018), 'most senior researcher in the company reporting directly to cio,' building fully systematic futures signals. earlier details are thinner: an msc in mathematics from the university of warsaw, and a dropout story he has told on podcasts.",
  "era": "prehistory",
  "id": "evt-quant-years",
  "sourceIds": [
    "src-hp-tworek-linkedin",
    "src-hp-warsawai",
    "src-hp-mad-podcast"
  ],
  "title": "the amsterdam quant years"
}

tworek joins openai · core-automation-timeline · 2019-02 (month) · confirmed (original editorial confidence)

Original location · source-ef8bfaf2a30037e9ac12 · source-b8207e74cbd595819904 · source-cce388647a5ce4b14bb6

Original structured fields

{
  "confidence": "confirmed",
  "date": "2019-02",
  "detail": "tworek joins openai — about thirty employees at the time — and stays nearly seven years, ending as vp of research.",
  "era": "prehistory",
  "id": "evt-joins-openai",
  "sourceIds": [
    "src-hp-millionintegrals",
    "src-hp-tworek-linkedin",
    "src-hp-corememory-ep53"
  ],
  "title": "tworek joins openai"
}

the rubik's-cube hand · core-automation-timeline · 2019-10 (month) · confirmed (original editorial confidence)

Original location · source-0dece10f702f4572dbed · source-ef8bfaf2a30037e9ac12

Original structured fields

{
  "confidence": "confirmed",
  "date": "2019-10",
  "detail": "'solving rubik's cube with a robot hand' — the sim2real breakthrough tworek ran, 'one of core breakthroughs' by his own account, landing at the neurips 2019 deep rl workshop.",
  "era": "prehistory",
  "id": "evt-rubiks",
  "sourceIds": [
    "src-hp-rubiks-paper",
    "src-hp-millionintegrals"
  ],
  "title": "the rubik's-cube hand"
}

codex, copilot, humaneval · core-automation-timeline · 2021 (year) · confirmed (original editorial confidence)

Original location · source-ef8bfaf2a30037e9ac12 · source-0788ae91e2e1304ed0ee

Original structured fields

{
  "confidence": "confirmed",
  "date": "2021",
  "detail": "the codex line: tworek 'conducted all stages of research and trained the model that powered github copilot,' was a primary contributor to codex, and hand-wrote about thirty humaneval problems — wojciech zaremba wrote thirty more.",
  "era": "prehistory",
  "id": "evt-codex",
  "sourceIds": [
    "src-hp-millionintegrals",
    "src-hp-agihouse-tworek"
  ],
  "title": "codex, copilot, humaneval"
}

o1 · core-automation-timeline · 2024-09-12 (day) · confirmed (original editorial confidence)

Original location · source-ef8bfaf2a30037e9ac12 · source-4ff7557d40d7300e631e · source-cce388647a5ce4b14bb6

Original structured fields

{
  "confidence": "confirmed",
  "date": "2024-09-12",
  "detail": "o1 ships — the reasoning model from the effort known internally as q*, then strawberry. tworek: 'i've lead the team that researched training llms to reason with reinforcement learning.'",
  "era": "prehistory",
  "id": "evt-o1",
  "sourceIds": [
    "src-hp-millionintegrals",
    "src-hp-fastcompany-o1",
    "src-hp-corememory-ep53"
  ],
  "title": "o1"
}

o3, and the vp-of-research seat · core-automation-timeline · 2025-04 (month) · confirmed (original editorial confidence)

Original location · source-ef8bfaf2a30037e9ac12 · source-0788ae91e2e1304ed0ee · source-b8207e74cbd595819904

Original structured fields

{
  "confidence": "confirmed",
  "date": "2025-04",
  "detail": "o3 ships. tworek's later summary of the o1→o3 gap: '99% systems, 1% algorithms.' he is vp of research by february 2025.",
  "era": "prehistory",
  "id": "evt-o3",
  "sourceIds": [
    "src-hp-millionintegrals",
    "src-hp-agihouse-tworek",
    "src-hp-tworek-linkedin"
  ],
  "title": "o3, and the vp-of-research seat"
}

gpt-5 · core-automation-timeline · 2025-08 (month) · confirmed (original editorial confidence)

Original location · source-ef8bfaf2a30037e9ac12 · source-fb717d0f7bffeb9549b2

Original structured fields

{
  "confidence": "confirmed",
  "date": "2025-08",
  "detail": "gpt-5 ships with reasoning descended from the o3 line — tworek later describes it on the mad podcast as roughly 'o3.1.'",
  "era": "prehistory",
  "id": "evt-gpt5",
  "sourceIds": [
    "src-hp-millionintegrals",
    "src-hp-mad-podcast"
  ],
  "title": "gpt-5"
}

the bench starts leaving · core-automation-timeline · 2025-08 (month) · reported (original editorial confidence)

Original location · source-979a8bb766a806d18dbe · source-cbcf6376012471f28d77 · source-7d4cb050b1379fdcf3cb · source-cdb5b1bbec06acac52b0

Original structured fields

{
  "confidence": "reported",
  "date": "2025-08",
  "detail": "the future cofounders disperse. julia villagra — promoted to openai chief people officer in march — exits in august, with reuters reporting she left to pursue 'art, music and storytelling to help people understand the transition to artificial general intelligence.' joanne jang founds openai labs in september and lands on the time100 ai list in august.",
  "era": "prehistory",
  "id": "evt-bench-disperses",
  "sourceIds": [
    "src-hp-bi-villagra",
    "src-hp-openai-leadership",
    "src-hp-jang-linkedin",
    "src-hp-prometheus-jang"
  ],
  "title": "the bench starts leaving"
}

the last openai interview · core-automation-timeline · 2025-10-16 (day) · confirmed (original editorial confidence)

Original location · source-fb717d0f7bffeb9549b2

Original structured fields

{
  "confidence": "confirmed",
  "date": "2025-10-16",
  "detail": "tworek records 'how gpt-5 thinks' on the mad podcast while still vp of research — the last big interview of the openai years.",
  "era": "prehistory",
  "id": "evt-mad-podcast",
  "sourceIds": [
    "src-hp-mad-podcast"
  ],
  "title": "the last openai interview"
}

tworek leaves openai · core-automation-timeline · 2026-01-05 (day) · confirmed (original editorial confidence)

Original location · source-b136574cef391eb5f753 · source-52ecd7f23aa3ee0428ee · source-136b826125d7cd0f7cc1

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-01-05",
  "detail": "tworek announces his openai departure on x after ~7 years, telling his team he wants 'to try and explore types of research that are hard to do at openai.' the verge and the decoder cover it the same day; the decoder reads the line as a 'not-so-subtle dig.'",
  "era": "departure",
  "id": "evt-exit",
  "sourceIds": [
    "src-hp-decoder-departure",
    "src-ci-verge-departure",
    "src-ci-timesofindia"
  ],
  "title": "tworek leaves openai"
}

the exit interview · core-automation-timeline · 2026-01-21 (day) · confirmed (original editorial confidence)

Original location · source-cce388647a5ce4b14bb6 · source-5e70740a643dfc418308

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-01-21",
  "detail": "core memory ep53, recorded days after the announcement: high-risk pioneering research had become hard at a 'conservative' openai; big labs as a whole have gone conservative; deep-learning research at big labs 'is done.'",
  "era": "departure",
  "id": "evt-corememory",
  "sourceIds": [
    "src-hp-corememory-ep53",
    "src-hp-decoder-google"
  ],
  "title": "the exit interview"
}

the information names the company · core-automation-timeline · 2026-01 (month) · reported (original editorial confidence)

Original location · source-a8396003b0412902cee3 · source-74af0d999cafa24cc0b4

Original structured fields

{
  "confidence": "reported",
  "date": "2026-01",
  "detail": "the information reports the new company — named core automation — is seeking $500m–$1b to build models needing ~100× less data that learn continuously; the reported long-term roadmap runs from industrial automation to 'self-replicating factories,' bio-machines, and 'planetary terraforming.' the piece is paywalled; the record runs through the decoder's summary.",
  "era": "departure",
  "id": "evt-information-jan",
  "sourceIds": [
    "src-hp-information-jan",
    "src-hp-decoder-two-startups"
  ],
  "title": "the information names the company"
}

incorporated · core-automation-timeline · 2026-03 (month) · reported (original editorial confidence)

Original location · source-6f5c00f78cd6f2767e81 · source-96b0c0d0ca7448c1103f

Original structured fields

{
  "confidence": "reported",
  "date": "2026-03",
  "detail": "core automation, inc. is founded — a delaware corporation incorporated in 2026 per its form d; 'late march' per dealroom. headquarters at 340 brannan street, suite 301, san francisco.",
  "era": "departure",
  "id": "evt-founded",
  "sourceIds": [
    "src-hp-sec-formd-coreauto",
    "src-ci-dealroom-4b"
  ],
  "title": "incorporated"
}

the ~$100m safe round · core-automation-timeline · 2026-04 (month) · reported (original editorial confidence)

Original location · source-e41e41d390c9511f63de · source-c3251d2c3a56d8cbe20a · source-8266c37e31ab9adb65d1 · source-6f5c00f78cd6f2767e81

Original structured fields

{
  "confidence": "reported",
  "date": "2026-04",
  "detail": "the initial ~$100m raise reportedly closes at ~$1b post-money — press and trackers name nvidia, spark capital, and accel, with scribble ventures and threshold ventures in secondary listings and no disclosed lead. the form d later shows the money came in on safes that converted into $135.3m of stock in the july round — the discrepancy between the reported ~$100m and the $135.3m conversion stays unresolved.",
  "era": "launch",
  "id": "evt-seed",
  "sourceIds": [
    "src-ci-sacra",
    "src-hp-aimarketwatch",
    "src-hp-vcbacked",
    "src-hp-sec-formd-coreauto"
  ],
  "title": "the ~$100m safe round"
}

launch day · core-automation-timeline · 2026-04-21 (day) · confirmed (original editorial confidence)

Original location · source-727c5c8b68c8b5feae9b · source-c7617f6c17f9c8e82605 · source-01752f80a651f84ea67c · source-80bd65dd892832ac1b0d

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-04-21",
  "detail": "public launch. the first x post: 'today we're announcing core automation… our objective: systems that optimize and automate work, starting with research itself' — 'the world's most automated ai lab.' tworek's own post: 'small step for a finger on \"post tweet\" button. big step for millions of future ai agents doing useful work for us.' cofounder avery lamp puts it on hacker news, where it gets 1 point and 0 comments.",
  "era": "launch",
  "id": "evt-launch",
  "sourceIds": [
    "src-hp-texxr-coreauto",
    "src-hp-texxr-tworek",
    "src-hp-coreauto-site",
    "src-ci-hn-launch"
  ],
  "title": "launch day"
}

the nerdsniped cascade · core-automation-timeline · 2026-04-21 (day) · confirmed (original editorial confidence)

Original location · source-fb31ff7db863bbd49e09 · source-20892eef1c43a22cb835 · source-0b1c92e6bcf986114894 · source-386f3e18013df302e869

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-04-21",
  "detail": "the coordinated reveal. rohan anil posts that he left anthropic 'a few weeks ago' — 'jerry tworek nerdsniped me into starting this with him and others.' anmol gulati announces from google deepmind: 'the current research paradigm — scaling models, data, and static deployment won't get us all the way.' joanne jang's x bio flips to 'trying to automate myself @coreautoai.' villagra, lamp, amid, and duvvuri are named in the coverage.",
  "era": "launch",
  "id": "evt-cofounder-cascade",
  "sourceIds": [
    "src-hp-bi-nerdsniped",
    "src-hp-gulati-x",
    "src-hp-digg-bios",
    "src-ci-yahoo-nerdsniped"
  ],
  "title": "the nerdsniped cascade"
}

the launch coverage · core-automation-timeline · 2026-04-22 (day) · confirmed (original editorial confidence)

Original location · source-fb31ff7db863bbd49e09 · source-12decd274540ee7661be · source-2dbef96532b83221e340 · source-386f3e18013df302e869

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-04-22",
  "detail": "the launch story travels: business insider's 'nerdsniped' piece becomes the most-syndicated account, the decoder files its launch writeup placing the lab in the neo-lab cohort with thinking machines and ssi, and korea's mk frames it as 'openai mafia' news while noting no business model has been disclosed.",
  "era": "launch",
  "id": "evt-launch-coverage",
  "sourceIds": [
    "src-hp-bi-nerdsniped",
    "src-hp-decoder-launch",
    "src-ci-mk-openai-mafia",
    "src-ci-yahoo-nerdsniped"
  ],
  "title": "the launch coverage"
}

$1b to $4b in six weeks · core-automation-timeline · 2026-05-07 (day) · reported (original editorial confidence)

Original location · source-b6d514eedc3bc810150b · source-96b0c0d0ca7448c1103f · source-b9c49f3b364c4f21407b

Original structured fields

{
  "confidence": "reported",
  "date": "2026-05-07",
  "detail": "the information reports the ~6-week-old company is already seeking $300m–$500m at ~$4b. palazzolo on x: 'back to back rounds are continuing in neolab land!' the report is paywalled — the record runs through techmeme, dealroom, and testingcatalog.",
  "era": "round",
  "id": "evt-4b-report",
  "sourceIds": [
    "src-ci-techmeme-4b",
    "src-ci-dealroom-4b",
    "src-ci-testingcatalog-4b"
  ],
  "title": "$1b to $4b in six weeks"
}

the equity round opens · core-automation-timeline · 2026-07-15 (day) · confirmed (original editorial confidence)

Original location · source-6f5c00f78cd6f2767e81 · source-37695c73cf25a68b6f48

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-07-15",
  "detail": "first sale of the new equity offering, per the form d filed two weeks later.",
  "era": "round",
  "id": "evt-first-sale",
  "sourceIds": [
    "src-hp-sec-formd-coreauto",
    "src-ci-takeoffradar-formd"
  ],
  "title": "the equity round opens"
}

the form d · core-automation-timeline · 2026-07-30 (day) · confirmed (original editorial confidence)

Original location · source-6f5c00f78cd6f2767e81 · source-ff08d2ff6992d9c9ed75 · source-37695c73cf25a68b6f48 · source-6606ae6fabb7f7988efb · source-a2d9394bfab0ee4893f0

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-07-30",
  "detail": "the form d lands: a $635,249,754 authorized offering with $432,125,053 sold to 51 investors and $203,124,701 remaining — including $135,252,607 issued on conversion of the outstanding safes. tracker math (forge/pitchbook-derived) puts the round at ~$3.2b pre / ~$3.62b post — reported, never confirmed by the company or the filing.",
  "era": "round",
  "id": "evt-formd",
  "sourceIds": [
    "src-hp-sec-formd-coreauto",
    "src-hp-filingflow",
    "src-ci-takeoffradar-formd",
    "src-hp-edgartools-news",
    "src-hp-svinvestclub"
  ],
  "title": "the form d"
}

axios reads the filing · core-automation-timeline · 2026-08-13 (day) · reported (original editorial confidence)

Original location · source-e0555d4db26f90fdcdc4 · source-2bddfe08c40223c88287

Original structured fields

{
  "confidence": "reported",
  "date": "2026-08-13",
  "detail": "axios pro reads the filing as the company 'looking to raise up to $500 million in new funding' — consistent with the remaining capacity plus new cash — and describes core automation as 'an ai continual learning model developer.'",
  "era": "round",
  "id": "evt-axios",
  "sourceIds": [
    "src-ci-axios-500m",
    "src-hp-gaebler"
  ],
  "title": "axios reads the filing"
}

the alumni ventures feeder · core-automation-timeline · 2026-08-17 (day) · confirmed (original editorial confidence)

Original location · source-90b985ffa13c4e40be75

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-08-17",
  "detail": "a second filing surfaces: 'avsf – core automation 2026, llc,' an alumni ventures feeder spv — $2,109,250 sold to 100 investors, first sale august 3.",
  "era": "round",
  "id": "evt-avsf",
  "sourceIds": [
    "src-hp-sec-formd-avsf"
  ],
  "title": "the alumni ventures feeder"
}

the blog opens · core-automation-timeline · 2026-05-27 (day) · confirmed (original editorial confidence)

Original location · source-5c72d71b0bb8c642c4bd

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-05-27",
  "detail": "the first company blog posts appear — saroufim's 'to automate publishing, first make authoring boring' and 'when ai starts writing systems code,' arguing that automating research means automating systems code first.",
  "era": "operating",
  "id": "evt-first-posts",
  "sourceIds": [
    "src-hp-coreauto-blog-data"
  ],
  "title": "the blog opens"
}

the agi house keynote · core-automation-timeline · 2026-07-18 (day) · confirmed (original editorial confidence)

Original location · source-0788ae91e2e1304ed0ee · source-4815fd612a41f2969104 · source-e79c40bca95653d9c270

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-07-18",
  "detail": "tworek keynotes the agi house auto-research summit in hillsborough — 'the era of evals is dumb,' 'the only test is the real world' — and puts up an open-problems slide that eighteen teams build on for the rest of the day.",
  "era": "operating",
  "id": "evt-agihouse",
  "sourceIds": [
    "src-hp-agihouse-tworek",
    "src-hp-agihouse-recap",
    "src-ci-agihouse-beehiiv"
  ],
  "title": "the agi house keynote"
}

the pacing letter · core-automation-timeline · 2026-07-28 (day) · reported (original editorial confidence)

Original location · source-b19d2bb5d0b9680ff25b · source-8795bdcd9374fad13af4

Original structured fields

{
  "confidence": "reported",
  "date": "2026-07-28",
  "detail": "the 'pacing the frontier' open letter goes public — 1,386 frontier-ai employees asking for 'technical and governance tools needed to deliberately pace the frontier of automated ai development.' core automation employees are among the signatories per the hn discussion — the lab that automates research, signing the letter about pacing it.",
  "era": "operating",
  "id": "evt-pacing",
  "sourceIds": [
    "src-ci-pacing-frontier",
    "src-ci-hn-pacing"
  ],
  "title": "the pacing letter"
}

the sequoia episode · core-automation-timeline · 2026-07-29 (day) · confirmed (original editorial confidence)

Original location · source-a9a49ed24697649f5f5b · source-a43e6354eaa0001d4b48 · source-47aea83e07c7f31bcaf9 · source-fd512636dadb9b9ddd8c

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-07-29",
  "detail": "sequoia's training data episode airs — 'building the automated agi lab: core automation's jerry tworek and rohan anil' — the fullest public statement of the thesis: transformers 'incapable of continual learning,' the search for a replacement architecture, and the most automated lab as the instrument. sequoia hosting is not evidence sequoia invested — the round's participants remain undisclosed.",
  "era": "operating",
  "id": "evt-sequoia",
  "sourceIds": [
    "src-ci-sequoia-podcast",
    "src-hp-training-data",
    "src-hp-teahose-transcript",
    "src-hp-sequoia-linkedin"
  ],
  "title": "the sequoia episode"
}

one layer deeper · core-automation-timeline · 2026-08-03 (day) · confirmed (original editorial confidence)

Original location · source-5c72d71b0bb8c642c4bd · source-c1c5dcd60750e8e098d0

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-08-03",
  "detail": "'one layer deeper' launches with tilde research — a public competition on repeated modular squaring: learn the computation, then run longer chains than seen in training. the hard track gives one h100, one hour of training, and asks models to certify depth t=1…64.",
  "era": "operating",
  "id": "evt-one-layer-deeper",
  "sourceIds": [
    "src-hp-coreauto-blog-data",
    "src-hp-tilde-linkedin"
  ],
  "title": "one layer deeper"
}

twenty-something people · core-automation-timeline · 2026-08-25 (day) · reported (original editorial confidence)

Original location · source-5290f4f001fbf457a678 · source-146299843306c293985b · source-a2d9394bfab0ee4893f0

Original structured fields

{
  "confidence": "reported",
  "date": "2026-08-25",
  "detail": "the fysk showcase primer reports the team at '20-something people' — drawn largely from openai and google deepmind plus anthropic, meta, xai, and amazon's agi lab — with a plan to reach ~40 by year-end. linkedin-derived counts say ~21; one aggregator says 14.",
  "era": "operating",
  "id": "evt-fysk-headcount",
  "sourceIds": [
    "src-hp-fysk-primer",
    "src-ci-linkedin",
    "src-hp-svinvestclub"
  ],
  "title": "twenty-something people"
}

the mts episode · core-automation-timeline · 2026-08-26 (day) · confirmed (original editorial confidence)

Original location · source-deff58863fdac1b64b44

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-08-26",
  "detail": "'ai labs are simply cooked' — the mts episode: three generations of ai labs, a two-year horizon for agents automating research, and the concession that today's agents 'lack deep field understanding and produce low-quality, high-creativity ideas.'",
  "era": "operating",
  "id": "evt-mts",
  "sourceIds": [
    "src-ci-mts-podcast"
  ],
  "title": "the mts episode"
}

the architecture-discovery post · core-automation-timeline · 2026-09-02 (day) · confirmed (original editorial confidence)

Original location · source-5c72d71b0bb8c642c4bd · source-d34253269fedfb7e0ec7

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-09-02",
  "detail": "ehsan amid's post on neural architecture discovery — untrained networks already separate cats from dogs; automate the discovery of architectures — and on automation reshaping the field.",
  "era": "operating",
  "id": "evt-amid-post",
  "sourceIds": [
    "src-hp-coreauto-blog-data",
    "src-hp-amid-site"
  ],
  "title": "the architecture-discovery post"
}

the competition results · core-automation-timeline · 2026-09-09 (day) · confirmed (original editorial confidence)

Original location · source-5c72d71b0bb8c642c4bd · source-9be97042033c5e509368

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-09-09",
  "detail": "'what the models learned in one layer deeper' — the results: 15,602 submissions; only 18 hard uploads from 5 accounts certified depth (16 reached t=64). the winners learned 'wiring around python primitives' — evidence that test-time depth is genuinely hard for current methods.",
  "era": "operating",
  "id": "evt-old-results",
  "sourceIds": [
    "src-hp-coreauto-blog-data",
    "src-hp-mcleish-site"
  ],
  "title": "the competition results"
}

anil on the bottleneck · core-automation-timeline · 2026-09-10 (day) · confirmed (original editorial confidence)

Original location · source-eb42858c6d9dbc1ece3b · source-acec6580f1f3823672df

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-09-10",
  "detail": "rohan anil on the information bottleneck — 'continual learning is the next bottleneck': the pre-training/rl split as 'organizational convenience rather than science,' shampoo's origin, the qr-factorization kernel competition (~60× over cusolver), and how he met tworek at dolores park.",
  "era": "operating",
  "id": "evt-infobottleneck",
  "sourceIds": [
    "src-ci-infobottleneck-anil",
    "src-hp-anil-teahose"
  ],
  "title": "anil on the bottleneck"
}

the agm program · core-automation-timeline · 2026-09 (month) · confirmed (original editorial confidence)

Original location · source-3bae41c15410c5d2c3c7

Original structured fields

{
  "confidence": "confirmed",
  "date": "2026-09",
  "detail": "the agm — 'associate general manager' — internship goes live: four interns, january–march 2027 in sf, $10,000/month plus profit share, each given a budget, tokens, and the company's agent infrastructure to run real online businesses. 'by watching where people intervene, we want to understand what agents still need to learn.' applications close october 6, 2026.",
  "era": "operating",
  "id": "evt-agm",
  "sourceIds": [
    "src-hp-coreauto-agm"
  ],
  "title": "the agm program"
}

the showcase · core-automation-timeline · 2026-09-16 (day) · reported (original editorial confidence)

Original location · source-5290f4f001fbf457a678

Original structured fields

{
  "confidence": "reported",
  "date": "2026-09-16",
  "detail": "tworek appears at the fysk sf startup showcase — the lab's fifth month ends the way it started: the founders in public.",
  "era": "operating",
  "id": "evt-fysk-showcase",
  "sourceIds": [
    "src-hp-fysk-primer"
  ],
  "title": "the showcase"
}