# core automation: side quests

> the pursuits around the company — the founders' earlier work, the competition and the internship, and the neo-lab ecosystem.

- 2015–2018 — the quant years (tworek, ended): tworek's pre-openai career — senior quantitative researcher at independent view in amsterdam, building fully systematic futures signals as the fund's most senior researcher. the trading frame survives: 'being a good trader means understanding your algorithm very well… in the end all you see is the p&l.' (sources: [jerry tworek — linkedin](https://www.linkedin.com/in/jerry-tworek-b5b9aa56), [jerry tworek speaker bio (warsaw.ai)](https://warsaw.ai/speaker/jerry-tworek/), [jerry tworek deep-dive (agi house)](https://blog.agihouse.org/posts/jerry-tworek-deep-dive))
- 2026– — [million integrals](https://millionintegrals.com/) (tworek, continuing): millionintegrals.com — the self-authored public record of the openai work: copilot, codex, humaneval, plugins and code interpreter, o1, o3, gpt-5, the chatgpt agent. part cv, part deposition. (sources: [million integrals — tworeks homepage](https://millionintegrals.com/))
- 2018– — the shampoo→psgd line (anil, continuing): the optimizer line anil carries in: shampoo, co-created at google brain with vineet gupta and tomer koren, then the psgd framework that 'can recover shampoo… a much better framework' — implemented inside the internal coreauto codebase. the qr-factorization kernel competition (~60× over cusolver, per anil on training data) sits in the same lineage. (sources: [continual learning is the next bottleneck — rohan anil (the information bottleneck)](https://www.the-information-bottleneck.com/p/continual-learning-is-the-next-bottleneck-2fc), [the psgd optimizer clip (digg tech)](https://digg.com/tech/51hgscvd), [building the automated agi lab — timestamped transcript (teahose)](https://www.teahose.com/podcast/Training%20Data/Building%20the%20Automated%20AGI%20Lab_%20Core%20Automation's%20Jerry%20Tworek%20and%20Rohan%20Anil), [rohan anil — google scholar](https://scholar.google.com/citations?user=m2qHgbwAAAAJ&hl=en))
- 2024–2026 — project mariner (gulati, handed-off): the agents-and-computer-use line — anmol gulati as tech lead on project mariner in gemini, kanav garg on the same project ('one of the most stimulating 0->1 projects'). the multimodal-agent experience the lab's work-automation thesis needs. (sources: [anmol gulati — linkedin](https://www.linkedin.com/in/anmol01gulati), [kanav garg — linkedin](https://www.linkedin.com/in/kanav-garg), [anmol gulati — google scholar](https://scholar.google.com/citations?user=S2Pk9ooAAAAJ&hl=en))
- 2021–2026 — model behavior and openai labs (jang, handed-off): jang's openai arc — product lead on dall·e 2, gpt-4, the chatgpt api, tts, and embeddings; founding lead of model behavior and steward of the model spec; then openai labs, prototyping interfaces 'beyond chat or even agents,' reportedly adjacent to the jony ive effort. (sources: [joanne jang — linkedin](https://www.linkedin.com/in/jangjoanne), [joanne jang profile (prometheus)](https://prometheusroot.com/people/joanne-jang/), [joanne jang speaker bio (womenpm)](https://womenpm.org/speakers/speaker-4-keynote-speakers/))
- 2025– — [reservoir samples](https://joannejang.com/) (jang, continuing): reservoir samples — jang's own essays on model personality and human-ai relationships, running alongside the lab work. (sources: [joanne jang profile (prometheus)](https://prometheusroot.com/people/joanne-jang/), [joannejang.com](https://joannejang.com/))
- 2025– — the agi-storytelling project (villagra, continuing): the stated reason villagra left openai — 'art, music and storytelling to help people understand the transition to artificial general intelligence.' how it coexists with an executive-officer seat at a research lab is unreported. (sources: [openai chief people officer julia villagra leaving (business insider)](https://www.businessinsider.com/openai-chief-people-officer-julia-villagra-leaving-ai-talent-hr-2025-8), [julia villagra — linkedin](https://www.linkedin.com/in/juliavillagra))
- 2021– — [gpu mode](https://www.marksaroufim.com/) (saroufim, continuing): saroufim's other company — the discord community for gpu kernel engineering that became a business running public kernel competitions, plus 'popcorn,' training an llm to write kernels. the qr-factorization contest anil cites lives in this lineage. saroufim remains a pytorch maintainer. (sources: [marksaroufim.com](https://www.marksaroufim.com/), [from pytorch to gpu mode — saroufim interview](https://share.snipd.com/episode/c8577ba1-ca13-42ac-841f-5aceef0ca3af), [mark saroufim — linkedin](https://www.linkedin.com/in/marksaroufim), [building the automated agi lab — timestamped transcript (teahose)](https://www.teahose.com/podcast/Training%20Data/Building%20the%20Automated%20AGI%20Lab_%20Core%20Automation's%20Jerry%20Tworek%20and%20Rohan%20Anil))
- 2020–2024 — mosaicml composer and adept (lamp, handed-off): lamp's earlier stack — early engineer at mosaicml, where his resume credits the composer training framework and the yahp hyperparameter system; mts at adept ai labs; a stealth cofounder in between; google deepmind after. (sources: [avery lamp — resume](https://averylamp.me/Resume.pdf), [avery lamp — linkedin](https://www.linkedin.com/in/averylamp))
- 2025– — the fysk seat (lamp, continuing): lamp's community seat — 'longtime engineer and lp at fysk,' the founders you should know community whose showcase hosted tworek on september 16. (sources: [your showcase primer: core automation (fysk)](https://newsletter.foundersysk.com/p/your-showcase-primer-core-automation))
- 2026-05 – — [the blog corpus](https://www.coreauto.com/blog) (company, continuing): the company blog — five posts in 2026: saroufim on boring authoring and ai-written systems code, the one layer deeper launch and results, amid on neural architecture discovery. the lab's public instrumentation. (sources: [the core automation blog](https://www.coreauto.com/blog))
- 2026-08 – 2026-09 — [one layer deeper](https://www.coreauto.com/blog) (company, ended): the public research competition co-run with tilde research: repeated modular squaring — learn the computation, then run longer chains than seen in training. hard track: one h100, one hour of training, certify depth t=1…64. results september 9: 15,602 submissions, only 18 hard uploads from 5 accounts certified. the first public artifact of the research program. (sources: [the core automation blog](https://www.coreauto.com/blog), [tilde research — linkedin](https://linkedin.com/company/tilderesearch), [sean mcleish — personal site](https://mcleish7.github.io/))
- 2026-09 – 2027-03 — [the agm program](https://www.coreauto.com/AGM) (company, active): the associate general manager program — four interns in sf, january–march 2027, $10k/month plus profit share, each running real online businesses on the company's agent infrastructure with a budget. 'by watching where people intervene, we want to understand what agents still need to learn.' a hiring instrument and a data-generation scheme in one. (sources: [associate general manager program (coreauto.com)](https://www.coreauto.com/AGM))
- 2026-07 — [the pacing letter](https://www.pacingthefrontier.com/) (company, ended): the 'pacing the frontier' open letter — 1,386 frontier-ai employees requesting '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 automation lab signing the pacing letter. (sources: [pacing the frontier — open letter](https://www.pacingthefrontier.com/), [the pacing the frontier hn thread](https://news.ycombinator.com/item?id=49089240))
- 2026– — [tilde research](https://linkedin.com/company/tilderesearch) (community, continuing): the collaborator lab — a ~12-person sf 'moonshot' interpretability/architectures outfit co-founded by dhruv pai and mason wang; co-ran one layer deeper and published 'wall attention' mit-licensed in june 2026. (sources: [tilde research — linkedin](https://linkedin.com/company/tilderesearch), [tilde research open-sources wall attention (ai insiders)](https://aiinsiders.net/article/tilde-research-open-sources-an-attention-variant-built-for), [sean mcleish — personal site](https://mcleish7.github.io/))
- 2026-08 — [the alumni ventures feeder](https://www.sec.gov/Archives/edgar/data/2146979/000214697926000001/primary_doc.xml) (community, ended): 'avsf – core automation 2026, llc' — the alumni ventures feeder spv with its own form d: $2,109,250 sold to 100 investors, first sale august 3. the only investor provable by filing. (sources: [avsf – core automation 2026, llc — form d (sec)](https://www.sec.gov/Archives/edgar/data/2146979/000214697926000001/primary_doc.xml))
- 2025– — the neo-lab cohort (community, continuing): the cohort core automation gets grouped with — thinking machines (murati), ssi (sutskever), periodic labs, ami labs (lecun), adaption labs (hooker), flapping airplanes ($180m, data-efficiency), inherent, prime intellect — competing for the same frontier-talent pool and the same neolab capital. (sources: [ex-openai researcher jerry tworek launches core automation (the decoder)](https://the-decoder.com/ex-openai-researcher-jerry-tworek-launches-core-automation-to-build-the-most-automated-ai-lab-in-the-world/), [core automation poaches top anthropic and deepmind talent (ai2work)](https://ai2.work/blog/core-automation-poaches-top-anthropic-and-deepmind-talent-for-ai-lab), [two startups want to replace how ai learns (the decoder)](https://the-decoder.com/two-startups-want-to-replace-how-ai-learns-one-just-raised-180m-another-is-seeking-up-to-1b/), [the pacing the frontier hn thread](https://news.ycombinator.com/item?id=49089240))

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/side-quests


Published by @ben in profile revision 1.
Original document: https://hraness.com/core-automation/side-quests
Original credited author: Hraness
AI-drafted at Ben Guo's direct request and credited to Hraness; every claim links to its cataloged source.

## Structured source and record data

````json
{"anchors":[{"id":"founders","originalUrl":"https://hraness.com/core-automation/side-quests#founders","title":"the founders' earlier work"},{"id":"company","originalUrl":"https://hraness.com/core-automation/side-quests#company","title":"the company's quests"},{"id":"ecosystem","originalUrl":"https://hraness.com/core-automation/side-quests#ecosystem","title":"the ecosystem"}],"provenance":{"draftingDisclosure":"AI-drafted at Ben Guo's direct request and credited to Hraness; every claim links to its cataloged source.","originalAuthor":"Hraness","originalRevision":"fa4b91d9929ce3af02109de3247374c0623a12995049b6972ada711f9503226e","originalUrl":"https://hraness.com/core-automation/side-quests","source":"hraness"},"records":[{"date":{"precision":"approximate","text":"2015–2018"},"id":"record-e03aada0f4407d087a0c","kind":"core-automation-quests","label":"the quant years","original":{"dates":"2015–2018","detail":"tworek's pre-openai career — senior quantitative researcher at independent view in amsterdam, building fully systematic futures signals as the fund's most senior researcher. the trading frame survives: 'being a good trader means understanding your algorithm very well… in the end all you see is the p&l.'","id":"qst-independent-view","owner":"tworek","sourceIds":["src-hp-tworek-linkedin","src-hp-warsawai","src-hp-agihouse-tworek"],"status":"ended","title":"the quant years","url":null},"originalUrl":"https://hraness.com/core-automation/side-quests","sourceIds":["source-ce9607ef27681cfe7a5b","source-c4b5a5f5065dd90c38cd","source-3a613855a780c7129b16"]},{"date":{"precision":"approximate","text":"2026–"},"id":"record-ecdcb564c24934ff4c4e","kind":"core-automation-quests","label":"million integrals","original":{"dates":"2026–","detail":"millionintegrals.com — the self-authored public record of the openai work: copilot, codex, humaneval, plugins and code interpreter, o1, o3, gpt-5, the chatgpt agent. part cv, part deposition.","id":"qst-millionintegrals","owner":"tworek","sourceIds":["src-hp-millionintegrals"],"status":"continuing","title":"million integrals","url":"https://millionintegrals.com/"},"originalUrl":"https://hraness.com/core-automation/side-quests","sourceIds":["source-5d2027ae004a5ad35639"]},{"date":{"precision":"approximate","text":"2018–"},"id":"record-aafc6d895b03a45bfe4e","kind":"core-automation-quests","label":"the shampoo→psgd line","original":{"dates":"2018–","detail":"the optimizer line anil carries in: shampoo, co-created at google brain with vineet gupta and tomer koren, then the psgd framework that 'can recover shampoo… a much better framework' — implemented inside the internal coreauto codebase. the qr-factorization kernel competition (~60× over cusolver, per anil on training data) sits in the same lineage.","id":"qst-optimizer-stack","owner":"anil","sourceIds":["src-ci-infobottleneck-anil","src-ci-digg-psgd","src-hp-teahose-transcript","src-hp-scholar-anil"],"status":"continuing","title":"the shampoo→psgd line","url":null},"originalUrl":"https://hraness.com/core-automation/side-quests","sourceIds":["source-0efe063653f2c8ace9bc","source-12ce6122f568bec5868e","source-62df074325d19b25863c","source-6d2d6841cb1d34ccb861"]},{"date":{"precision":"approximate","text":"2024–2026"},"id":"record-12cb8403bce94a99dbe9","kind":"core-automation-quests","label":"project mariner","original":{"dates":"2024–2026","detail":"the agents-and-computer-use line — anmol gulati as tech lead on project mariner in gemini, kanav garg on the same project ('one of the most stimulating 0->1 projects'). the multimodal-agent experience the lab's work-automation thesis needs.","id":"qst-mariner","owner":"gulati","sourceIds":["src-hp-gulati-linkedin","src-hp-garg-linkedin","src-hp-gulati-scholar"],"status":"handed-off","title":"project mariner","url":null},"originalUrl":"https://hraness.com/core-automation/side-quests","sourceIds":["source-bb5ba9b889a717a5b1ab","source-8c5ff91f71eae2e3b4ec","source-22a20288eb982f2c7b41"]},{"date":{"precision":"approximate","text":"2021–2026"},"id":"record-810a0b5a1619e40a259e","kind":"core-automation-quests","label":"model behavior and openai labs","original":{"dates":"2021–2026","detail":"jang's openai arc — product lead on dall·e 2, gpt-4, the chatgpt api, tts, and embeddings; founding lead of model behavior and steward of the model spec; then openai labs, prototyping interfaces 'beyond chat or even agents,' reportedly adjacent to the jony ive effort.","id":"qst-openai-labs","owner":"jang","sourceIds":["src-hp-jang-linkedin","src-hp-prometheus-jang","src-hp-womenpm-jang"],"status":"handed-off","title":"model behavior and openai labs","url":null},"originalUrl":"https://hraness.com/core-automation/side-quests","sourceIds":["source-2f2f8fb90e168d9c6c4e","source-383b986f707f88869167","source-2ea21ec97b1476cb6335"]},{"date":{"precision":"approximate","text":"2025–"},"id":"record-60c1d0fadbc0bdf8167b","kind":"core-automation-quests","label":"reservoir samples","original":{"dates":"2025–","detail":"reservoir samples — jang's own essays on model personality and human-ai relationships, running alongside the lab work.","id":"qst-reservoir-samples","owner":"jang","sourceIds":["src-hp-prometheus-jang","src-hp-jang-site"],"status":"continuing","title":"reservoir samples","url":"https://joannejang.com/"},"originalUrl":"https://hraness.com/core-automation/side-quests","sourceIds":["source-383b986f707f88869167","source-4decf0ad8f33e642f723"]},{"date":{"precision":"approximate","text":"2025–"},"id":"record-b5bfd7abcd953c9a0380","kind":"core-automation-quests","label":"the agi-storytelling project","original":{"dates":"2025–","detail":"the stated reason villagra left openai — 'art, music and storytelling to help people understand the transition to artificial general intelligence.' how it coexists with an executive-officer seat at a research lab is unreported.","id":"qst-agi-storytelling","owner":"villagra","sourceIds":["src-hp-bi-villagra","src-hp-villagra-linkedin"],"status":"continuing","title":"the agi-storytelling project","url":null},"originalUrl":"https://hraness.com/core-automation/side-quests","sourceIds":["source-cfdb4e911a2810c4b2a9","source-0bd8396556423736da54"]},{"date":{"precision":"approximate","text":"2021–"},"id":"record-87ff9606afe899bdacbd","kind":"core-automation-quests","label":"gpu mode","original":{"dates":"2021–","detail":"saroufim's other company — the discord community for gpu kernel engineering that became a business running public kernel competitions, plus 'popcorn,' training an llm to write kernels. the qr-factorization contest anil cites lives in this lineage. saroufim remains a pytorch maintainer.","id":"qst-gpu-mode","owner":"saroufim","sourceIds":["src-ci-marksaroufim","src-hp-saroufim-podcast","src-hp-saroufim-linkedin","src-hp-teahose-transcript"],"status":"continuing","title":"gpu mode","url":"https://www.marksaroufim.com/"},"originalUrl":"https://hraness.com/core-automation/side-quests","sourceIds":["source-7c7ecac2354d57f641a2","source-d66d4e69d37c14da6ddc","source-411bd5c1ec5c1a39e633","source-62df074325d19b25863c"]},{"date":{"precision":"approximate","text":"2020–2024"},"id":"record-4fcbd04797a0cb2d60ce","kind":"core-automation-quests","label":"mosaicml composer and adept","original":{"dates":"2020–2024","detail":"lamp's earlier stack — early engineer at mosaicml, where his resume credits the composer training framework and the yahp hyperparameter system; mts at adept ai labs; a stealth cofounder in between; google deepmind after.","id":"qst-mosaicml","owner":"lamp","sourceIds":["src-hp-lamp-resume","src-hp-lamp-linkedin"],"status":"handed-off","title":"mosaicml composer and adept","url":null},"originalUrl":"https://hraness.com/core-automation/side-quests","sourceIds":["source-539e7cfe0dec1574170c","source-2eafe985d97fd135b095"]},{"date":{"precision":"approximate","text":"2025–"},"id":"record-92e7696a606f154b74e0","kind":"core-automation-quests","label":"the fysk seat","original":{"dates":"2025–","detail":"lamp's community seat — 'longtime engineer and lp at fysk,' the founders you should know community whose showcase hosted tworek on september 16.","id":"qst-fysk","owner":"lamp","sourceIds":["src-hp-fysk-primer"],"status":"continuing","title":"the fysk seat","url":null},"originalUrl":"https://hraness.com/core-automation/side-quests","sourceIds":["source-b29a173bcf2332d3ef60"]},{"date":{"precision":"approximate","text":"2026-05 –"},"id":"record-ec72254c782ae043ad8c","kind":"core-automation-quests","label":"the blog corpus","original":{"dates":"2026-05 –","detail":"the company blog — five posts in 2026: saroufim on boring authoring and ai-written systems code, the one layer deeper launch and results, amid on neural architecture discovery. the lab's public instrumentation.","id":"qst-blog","owner":"company","sourceIds":["src-hp-coreauto-blog-data"],"status":"continuing","title":"the blog corpus","url":"https://www.coreauto.com/blog"},"originalUrl":"https://hraness.com/core-automation/side-quests#company","sourceIds":["source-8d49001907b593381195"]},{"date":{"precision":"approximate","text":"2026-08 – 2026-09"},"id":"record-3ab8715c26a482ffb9aa","kind":"core-automation-quests","label":"one layer deeper","original":{"dates":"2026-08 – 2026-09","detail":"the public research competition co-run with tilde research: repeated modular squaring — learn the computation, then run longer chains than seen in training. hard track: one h100, one hour of training, certify depth t=1…64. results september 9: 15,602 submissions, only 18 hard uploads from 5 accounts certified. the first public artifact of the research program.","id":"qst-one-layer-deeper","owner":"company","sourceIds":["src-hp-coreauto-blog-data","src-hp-tilde-linkedin","src-hp-mcleish-site"],"status":"ended","title":"one layer deeper","url":"https://www.coreauto.com/blog"},"originalUrl":"https://hraness.com/core-automation/side-quests#company","sourceIds":["source-8d49001907b593381195","source-1173d6130c2380c00ff1","source-06d5104fa5e2332667b6"]},{"date":{"precision":"approximate","text":"2026-09 – 2027-03"},"id":"record-2d149e465f41b9dc1d6c","kind":"core-automation-quests","label":"the agm program","original":{"dates":"2026-09 – 2027-03","detail":"the associate general manager program — four interns in sf, january–march 2027, $10k/month plus profit share, each running real online businesses on the company's agent infrastructure with a budget. 'by watching where people intervene, we want to understand what agents still need to learn.' a hiring instrument and a data-generation scheme in one.","id":"qst-agm","owner":"company","sourceIds":["src-hp-coreauto-agm"],"status":"active","title":"the agm program","url":"https://www.coreauto.com/AGM"},"originalUrl":"https://hraness.com/core-automation/side-quests#company","sourceIds":["source-7291abf59197f90c7071"]},{"date":{"precision":"month","text":"2026-07"},"id":"record-c449d78a659e96b2fdb5","kind":"core-automation-quests","label":"the pacing letter","original":{"dates":"2026-07","detail":"the 'pacing the frontier' open letter — 1,386 frontier-ai employees requesting '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 automation lab signing the pacing letter.","id":"qst-pacing-letter","owner":"company","sourceIds":["src-ci-pacing-frontier","src-ci-hn-pacing"],"status":"ended","title":"the pacing letter","url":"https://www.pacingthefrontier.com/"},"originalUrl":"https://hraness.com/core-automation/side-quests#company","sourceIds":["source-909a0a9616b8d7414ce2","source-e244266c685f0c07fd75"]},{"date":{"precision":"approximate","text":"2026–"},"id":"record-ca3d3863a8e506e41597","kind":"core-automation-quests","label":"tilde research","original":{"dates":"2026–","detail":"the collaborator lab — a ~12-person sf 'moonshot' interpretability/architectures outfit co-founded by dhruv pai and mason wang; co-ran one layer deeper and published 'wall attention' mit-licensed in june 2026.","id":"qst-tilde","owner":"community","sourceIds":["src-hp-tilde-linkedin","src-hp-aiinsiders-tilde","src-hp-mcleish-site"],"status":"continuing","title":"tilde research","url":"https://linkedin.com/company/tilderesearch"},"originalUrl":"https://hraness.com/core-automation/side-quests","sourceIds":["source-1173d6130c2380c00ff1","source-8cdc1ae200c4f11c0f7d","source-06d5104fa5e2332667b6"]},{"date":{"precision":"month","text":"2026-08"},"id":"record-64870f4e4cdb85423015","kind":"core-automation-quests","label":"the alumni ventures feeder","original":{"dates":"2026-08","detail":"'avsf – core automation 2026, llc' — the alumni ventures feeder spv with its own form d: $2,109,250 sold to 100 investors, first sale august 3. the only investor provable by filing.","id":"qst-avsf","owner":"community","sourceIds":["src-hp-sec-formd-avsf"],"status":"ended","title":"the alumni ventures feeder","url":"https://www.sec.gov/Archives/edgar/data/2146979/000214697926000001/primary_doc.xml"},"originalUrl":"https://hraness.com/core-automation/side-quests","sourceIds":["source-2b86bb5a146e13b4d444"]},{"date":{"precision":"approximate","text":"2025–"},"id":"record-4a61b0c79224d470a001","kind":"core-automation-quests","label":"the neo-lab cohort","original":{"dates":"2025–","detail":"the cohort core automation gets grouped with — thinking machines (murati), ssi (sutskever), periodic labs, ami labs (lecun), adaption labs (hooker), flapping airplanes ($180m, data-efficiency), inherent, prime intellect — competing for the same frontier-talent pool and the same neolab capital.","id":"qst-neolabs","owner":"community","sourceIds":["src-hp-decoder-launch","src-ci-ai2work","src-hp-decoder-two-startups","src-ci-hn-pacing"],"status":"continuing","title":"the neo-lab cohort","url":null},"originalUrl":"https://hraness.com/core-automation/side-quests","sourceIds":["source-1fd215b19b213629a925","source-77621659e1c8914522a0","source-6d0a82da75d2e94a501b","source-e244266c685f0c07fd75"]}],"sources":[{"author":null,"id":"source-77621659e1c8914522a0","note":"ceres model claim via the information; neolab comparison table; \"up to $1b\" phrasing","originalId":"src-ci-ai2work","published":{"precision":"day","text":"2026-04-22"},"publisher":"ai2.work","title":"core automation poaches top anthropic and deepmind talent (ai2work)","type":"reference","url":"https://ai2.work/blog/core-automation-poaches-top-anthropic-and-deepmind-talent-for-ai-lab"},{"author":null,"id":"source-12ce6122f568bec5868e","note":"psgd optimizer clip; \"recovers shampoo\"; internal `coreauto` codebase","originalId":"src-ci-digg-psgd","published":{"precision":"year","text":"2026"},"publisher":"digg","title":"the psgd optimizer clip (digg tech)","type":"reporting","url":"https://digg.com/tech/51hgscvd"},{"author":null,"id":"source-e244266c685f0c07fd75","note":"\"pacing the frontier\" discussion; 0xdeafbead cites core automation signatories vs. regulatory-capture theory","originalId":"src-ci-hn-pacing","published":{"precision":"day","text":"2026-07-28"},"publisher":"hacker news","title":"the pacing the frontier hn thread","type":"community","url":"https://news.ycombinator.com/item?id=49089240"},{"author":null,"id":"source-0efe063653f2c8ace9bc","note":"anil interview; pretrain/rl split as \"organizational convenience\"; kernel-competition + hiring segments; dolores park origin","originalId":"src-ci-infobottleneck-anil","published":{"precision":"day","text":"2026-09-10"},"publisher":"the-information-bottleneck.com","title":"continual learning is the next bottleneck — rohan anil (the information bottleneck)","type":"interview","url":"https://www.the-information-bottleneck.com/p/continual-learning-is-the-next-bottleneck-2fc"},{"author":"mark saroufim","id":"source-7c7ecac2354d57f641a2","note":"saroufim: core automation cofounder, pytorch maintainer, gpu mode cofounder","originalId":"src-ci-marksaroufim","published":{"precision":"month","text":"2026-09"},"publisher":"marksaroufim.com","title":"marksaroufim.com","type":"primary","url":"https://www.marksaroufim.com/"},{"author":null,"id":"source-909a0a9616b8d7414ce2","note":"1,386 frontier-ai employees; requests usg-backed international pacing tools for automated ai development","originalId":"src-ci-pacing-frontier","published":{"precision":"month","text":"2026-07"},"publisher":"pacingthefrontier.com","title":"pacing the frontier — open letter","type":"reporting","url":"https://www.pacingthefrontier.com/"},{"author":null,"id":"source-3a613855a780c7129b16","note":"deep-dive on tworek keynote/interview — evals, o1→o3, continual learning, founding","originalId":"src-hp-agihouse-tworek","published":{"precision":"day","text":"2026-07-29"},"publisher":"blog.agihouse.org","title":"jerry tworek deep-dive (agi house)","type":"interview","url":"https://blog.agihouse.org/posts/jerry-tworek-deep-dive"},{"author":null,"id":"source-8cdc1ae200c4f11c0f7d","note":"tilde's wall attention open-sourced","originalId":"src-hp-aiinsiders-tilde","published":{"precision":"day","text":"2026-06-02"},"publisher":"aiinsiders.net","title":"tilde research open-sources wall attention (ai insiders)","type":"reporting","url":"https://aiinsiders.net/article/tilde-research-open-sources-an-attention-variant-built-for"},{"author":null,"id":"source-cfdb4e911a2810c4b2a9","note":"villagra's openai exit; reuters' \"art, music and storytelling…agi\" line","originalId":"src-hp-bi-villagra","published":{"precision":"month","text":"2025-08"},"publisher":"business insider","title":"openai chief people officer julia villagra leaving (business insider)","type":"reporting","url":"https://www.businessinsider.com/openai-chief-people-officer-julia-villagra-leaving-ai-talent-hr-2025-8"},{"author":null,"id":"source-7291abf59197f90c7071","note":"associate general manager internship page — jan–mar 2027, $10k/mo, apply by oct 6, 2026","originalId":"src-hp-coreauto-agm","published":{"precision":"day","text":"2026-09-16"},"publisher":"coreauto.com","title":"associate general manager program (coreauto.com)","type":"primary","url":"https://www.coreauto.com/AGM"},{"author":"core automation","id":"source-8d49001907b593381195","note":"company blog corpus: five 2026 posts — saroufim on authoring and systems code, the one layer deeper launch and results, amid on neural architecture discovery","originalId":"src-hp-coreauto-blog-data","published":{"precision":"day","text":"2026-05-27"},"publisher":"coreauto.com","title":"the core automation blog","type":"primary","url":"https://www.coreauto.com/blog"},{"author":null,"id":"source-1fd215b19b213629a925","note":"launch coverage; \"deep learning research 'is done'\"; neo-lab framing","originalId":"src-hp-decoder-launch","published":{"precision":"day","text":"2026-04-22"},"publisher":"the-decoder.com","title":"ex-openai researcher jerry tworek launches core automation (the decoder)","type":"reporting","url":"https://the-decoder.com/ex-openai-researcher-jerry-tworek-launches-core-automation-to-build-the-most-automated-ai-lab-in-the-world/"},{"author":null,"id":"source-6d0a82da75d2e94a501b","note":"pairs core automation w/ flapping airplanes; 100× less data goal; \"self-replicating factories…planetary terraforming\"","originalId":"src-hp-decoder-two-startups","published":{"precision":"day","text":"2026-01-29"},"publisher":"the-decoder.com","title":"two startups want to replace how ai learns (the decoder)","type":"reporting","url":"https://the-decoder.com/two-startups-want-to-replace-how-ai-learns-one-just-raised-180m-another-is-seeking-up-to-1b/"},{"author":null,"id":"source-b29a173bcf2332d3ef60","note":"\"20-something people,\" ~40 by year-end; lamp = fysk lp; sept 16 showcase","originalId":"src-hp-fysk-primer","published":{"precision":"day","text":"2026-08-25"},"publisher":"newsletter.foundersysk.com","title":"your showcase primer: core automation (fysk)","type":"reporting","url":"https://newsletter.foundersysk.com/p/your-showcase-primer-core-automation"},{"author":null,"id":"source-8c5ff91f71eae2e3b4ec","note":"co-founder; ex-deepmind (project mariner post, dec 2024)","originalId":"src-hp-garg-linkedin","published":{"precision":"day","text":"2026-09-16"},"publisher":"linkedin","title":"kanav garg — linkedin","type":"reference","url":"https://www.linkedin.com/in/kanav-garg"},{"author":null,"id":"source-bb5ba9b889a717a5b1ab","note":"deepmind roles: mariner/agents tl; gemini 3.0/2.5/2.0 credits; google brain speech","originalId":"src-hp-gulati-linkedin","published":{"precision":"day","text":"2026-09-16"},"publisher":"linkedin","title":"anmol gulati — linkedin","type":"reference","url":"https://www.linkedin.com/in/anmol01gulati"},{"author":null,"id":"source-22a20288eb982f2c7b41","note":"gulati publications","originalId":"src-hp-gulati-scholar","published":{"precision":"day","text":"2026-09-16"},"publisher":"scholar.google.com","title":"anmol gulati — google scholar","type":"repo","url":"https://scholar.google.com/citations?user=S2Pk9ooAAAAJ&hl=en"},{"author":null,"id":"source-2f2f8fb90e168d9c6c4e","note":"openai gm dec 2021–apr 2026; labs/model-behavior history; time100 post","originalId":"src-hp-jang-linkedin","published":{"precision":"day","text":"2026-09-16"},"publisher":"linkedin","title":"joanne jang — linkedin","type":"reference","url":"https://www.linkedin.com/in/jangjoanne"},{"author":"joanne jang","id":"source-4decf0ad8f33e642f723","note":"personal site — stanford cs/applied math","originalId":"src-hp-jang-site","published":{"precision":"day","text":"2026-09-16"},"publisher":"joannejang.com","title":"joannejang.com","type":"primary","url":"https://joannejang.com/"},{"author":null,"id":"source-2eafe985d97fd135b095","note":"deepmind swe-research; adept mts; stealth co-founder","originalId":"src-hp-lamp-linkedin","published":{"precision":"day","text":"2026-09-16"},"publisher":"linkedin","title":"avery lamp — linkedin","type":"reference","url":"https://www.linkedin.com/in/averylamp"},{"author":"avery lamp","id":"source-539e7cfe0dec1574170c","note":"mosaicml composer/yahp; google intern 2018","originalId":"src-hp-lamp-resume","published":{"precision":"day","text":"2026-09-16"},"publisher":"averylamp.me","title":"avery lamp — resume","type":"primary","url":"https://averylamp.me/Resume.pdf"},{"author":null,"id":"source-06d5104fa5e2332667b6","note":"sean mcleish (co-author, sep 9 results post) — tilde research intern","originalId":"src-hp-mcleish-site","published":{"precision":"day","text":"2026-09-16"},"publisher":"mcleish7.github.io","title":"sean mcleish — personal site","type":"reference","url":"https://mcleish7.github.io/"},{"author":"jerry tworek","id":"source-5d2027ae004a5ad35639","note":"tworek's self-authored work history (copilot, codex, o1, o3, gpt-5, chatgpt agent)","originalId":"src-hp-millionintegrals","published":{"precision":"day","text":"2026-09-16"},"publisher":"millionintegrals.com","title":"million integrals — tworeks homepage","type":"primary","url":"https://millionintegrals.com/"},{"author":null,"id":"source-383b986f707f88869167","note":"model spec stewardship; oai labs; reservoir samples; jony ive adjacency (reported)","originalId":"src-hp-prometheus-jang","published":{"precision":"year","text":"2026"},"publisher":"prometheusroot.com","title":"joanne jang profile (prometheus)","type":"reference","url":"https://prometheusroot.com/people/joanne-jang/"},{"author":null,"id":"source-411bd5c1ec5c1a39e633","note":"gpu mode/pytorch work list","originalId":"src-hp-saroufim-linkedin","published":{"precision":"day","text":"2026-09-16"},"publisher":"linkedin","title":"mark saroufim — linkedin","type":"reference","url":"https://www.linkedin.com/in/marksaroufim"},{"author":null,"id":"source-d66d4e69d37c14da6ddc","note":"\"from pytorch to gpu mode\" — microsoft→graphcore→pytorch→gpu mode path","originalId":"src-hp-saroufim-podcast","published":{"precision":"year","text":"2026"},"publisher":"snipd","title":"from pytorch to gpu mode — saroufim interview","type":"interview","url":"https://share.snipd.com/episode/c8577ba1-ca13-42ac-841f-5aceef0ca3af"},{"author":null,"id":"source-6d2d6841cb1d34ccb861","note":"anil publications (gemini, palm, etc.)","originalId":"src-hp-scholar-anil","published":{"precision":"day","text":"2026-09-16"},"publisher":"scholar.google.com","title":"rohan anil — google scholar","type":"repo","url":"https://scholar.google.com/citations?user=m2qHgbwAAAAJ&hl=en"},{"author":null,"id":"source-2b86bb5a146e13b4d444","note":"\"avsf – core automation 2026, llc\" — alumni ventures feeder spv, $2,109,250, 100 investors, first sale 2026-08-03","originalId":"src-hp-sec-formd-avsf","published":{"precision":"day","text":"2026-08-17"},"publisher":"sec","title":"avsf – core automation 2026, llc — form d (sec)","type":"primary","url":"https://www.sec.gov/Archives/edgar/data/2146979/000214697926000001/primary_doc.xml"},{"author":"sonya huang and pat grady","id":"source-62df074325d19b25863c","note":"timestamped verbatim quotes from the episode","originalId":"src-hp-teahose-transcript","published":{"precision":"day","text":"2026-07-29"},"publisher":"teahose.com","title":"building the automated agi lab — timestamped transcript (teahose)","type":"archive","url":"https://www.teahose.com/podcast/Training%20Data/Building%20the%20Automated%20AGI%20Lab_%20Core%20Automation's%20Jerry%20Tworek%20and%20Rohan%20Anil"},{"author":"tilde research","id":"source-1173d6130c2380c00ff1","note":"tilde research — 12-person moonshot lab (dhruv pai, mason wang); competition co-host","originalId":"src-hp-tilde-linkedin","published":{"precision":"day","text":"2026-09-16"},"publisher":"linkedin","title":"tilde research — linkedin","type":"reference","url":"https://linkedin.com/company/tilderesearch"},{"author":null,"id":"source-ce9607ef27681cfe7a5b","note":"openai roles/dates; independent view (amsterdam) 2015–2018","originalId":"src-hp-tworek-linkedin","published":{"precision":"day","text":"2026-09-16"},"publisher":"linkedin","title":"jerry tworek — linkedin","type":"reference","url":"https://www.linkedin.com/in/jerry-tworek-b5b9aa56"},{"author":null,"id":"source-0bd8396556423736da54","note":"openai cpo jun 2024–oct 2025 (per li text); hrt head of people; \"member of the xyz staff\"","originalId":"src-hp-villagra-linkedin","published":{"precision":"day","text":"2026-09-16"},"publisher":"linkedin","title":"julia villagra — linkedin","type":"reference","url":"https://www.linkedin.com/in/juliavillagra"},{"author":null,"id":"source-c4b5a5f5065dd90c38cd","note":"bio: msc mathematics univ. of warsaw; \"first five years of his career in the hedge fund industry\"","originalId":"src-hp-warsawai","published":{"precision":"year","text":"2020"},"publisher":"warsaw.ai","title":"jerry tworek speaker bio (warsaw.ai)","type":"reference","url":"https://warsaw.ai/speaker/jerry-tworek/"},{"author":null,"id":"source-2ea21ec97b1476cb6335","note":"bio: google assistant nlu, dropbox; ms cs + bs applied math stanford; joined openai 2021","originalId":"src-hp-womenpm-jang","published":{"precision":"unknown","text":"undated"},"publisher":"womenpm.org","title":"joanne jang speaker bio (womenpm)","type":"reference","url":"https://womenpm.org/speakers/speaker-4-keynote-speakers/"}]}
````