# Rebecca Bellan reports OpenAI’s ambitions for AI research assistants in TechCrunch [Canonical reading](https://escapemeridian.com/readings/rebecca-bellan-reports-openai-s-ambitions-for-ai-research-fabb7cba) Published October twenty-eighth, twenty twenty-five, Rebecca Bellan reports that OpenAI CEO Sam Altman expects an intern-level research assistant by September 2026 and a fully automated “legitimate AI researcher” by 2028. OpenAI chief scientist Jakub Pachocki defines that researcher as a system able to deliver larger research projects autonomously, not a human who studies AI. He says OpenAI is pursuing algorithmic innovation and more “test time compute”—time and computing resources spent thinking through problems. Pachocki says current models handle tasks with roughly a five-hour time horizon, and argues that dedicating data-center-scale computing to a problem could be worthwhile for major scientific breakthroughs. These are OpenAI leaders’ forecasts and plans, not reported results. The article links the timetable to OpenAI’s restructuring as a public benefit corporation and Altman’s account of the infrastructure investment it would require. OpenAI says its nonprofit foundation will govern research direction and manage research and safety initiatives. Whether the projected research capabilities will arrive on this timetable, and what scientific work they might deliver, remains unknown. Headline and summary by Escape Meridian, prepared by Luna from the attributed source. A source check is not independent confirmation of the source’s claims. ## Original source TechCrunch · Article Original source date: Oct 28, 2025 (2025-10-28). Date basis: Date from the publisher’s archive: its article address or archive listing. Added to Escape Meridian: 2026-11-11T12:03:25.040Z. First observed: 2026-10-28T11:33:52.828Z. Retrieved: 2026-10-28T11:35:50.630Z. - [Sam Altman says OpenAI will have a ‘legitimate AI researcher’ by 2028](https://techcrunch.com/2025/10/28/sam-altman-says-openai-will-have-a-legitimate-ai-researcher-by-2028/) ## Limits Text as retrieved on October 28, 2026; originally published October 28, 2025. The article reports OpenAI leaders’ forecasts and plans; it provides no results demonstrating that the projected AI research capabilities have been achieved. ## Open question What evidence will show whether an AI system can autonomously deliver larger research projects, rather than only perform well on bounded tasks? ## Source-stated Sam Altman said OpenAI was tracking toward an intern-level research assistant by September 2026 and a fully automated “legitimate AI researcher” by 2028. > intern-level research assistant by September 2026 and a fully automated “legitimate AI researcher” by 2028, CEO Sam Altman said - [Supporting source](https://techcrunch.com/2025/10/28/sam-altman-says-openai-will-have-a-legitimate-ai-researcher-by-2028/) ## Source-stated Jakub Pachocki described an AI researcher as a system capable of autonomously delivering larger research projects. > “system capable of autonomously delivering on larger research projects.” - [Supporting source](https://techcrunch.com/2025/10/28/sam-altman-says-openai-will-have-a-legitimate-ai-researcher-by-2028/) ## Source-stated Pachocki said OpenAI was pursuing algorithmic innovation and scaling up “test time compute.” > continued algorithmic innovation and dramatically scaling up “test time compute” - [Supporting source](https://techcrunch.com/2025/10/28/sam-altman-says-openai-will-have-a-legitimate-ai-researcher-by-2028/) ## Source-stated Pachocki said current models can handle tasks with a roughly five-hour time horizon and that dedicating data-center-scale computing to a problem could be worthwhile for major scientific breakthroughs. > Current models can handle tasks with a roughly five-hour time horizon - [Supporting source](https://techcrunch.com/2025/10/28/sam-altman-says-openai-will-have-a-legitimate-ai-researcher-by-2028/) ## Source-stated Altman said OpenAI had committed to 30 gigawatts of infrastructure, a $1.4 trillion financial obligation over the next few years. > Altman said OpenAI has committed to 30 gigawatts of infrastructure, which is a $1.4 trillion financial obligation, over the next few years. - [Supporting source](https://techcrunch.com/2025/10/28/sam-altman-says-openai-will-have-a-legitimate-ai-researcher-by-2028/) ## Source and contributor context TechCrunch · Publisher The publisher is identified here. No individual author has been inferred. ## Source media Media retrieval status: unchecked. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Test-time compute Extra computation a model spends while answering, for example by reasoning longer, rather than during training. - [Definition reference](https://arxiv.org/abs/2408.03314) ## Time horizon METR's measure of AI agents' ability: the length of task, timed by how long skilled people take, that an agent completes at a set success rate. - [Definition reference](https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/) ## Public benefit corporation A for-profit company that is legally required to weigh a stated public purpose alongside profit. - [Definition reference](https://en.wikipedia.org/wiki/Benefit_corporation) ## OpenAI OpenAI develops AI models and products and publishes research, release announcements and evaluations. - [Explore this context](/entities/openai) - [Reference](https://openai.com/news/) ## Sam Altman Sam Altman publishes essays about AI and technology on his personal blog, separate from OpenAI announcements. - [Explore this context](/entities/sam-altman) - [Reference](https://blog.samaltman.com/) - [Reference](https://openai.com/our-structure/) - [Reference](https://blog.samaltman.com/posts.atom) ## TIME TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services. - [Explore this context](/entities/time) - [Reference](https://time.com/) ## Editorial provenance Version 1. Publication ID: 645b7ffc-e7f2-7bbf-c9d2-f76374d6a036. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-11T11:59:07.744Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: bridge\_admitted\_by\_model. Observation: cd839b0fb696844915ee7ad788297412c6796d2b2fb3a3bc95b6423a18c0cde8. Source text hash: 282880c2d3c477026f57b5ad56f984a1cfb518ba33b1f324ae77a4252bad4c8f. ## Revision and source-check history ### Version 1 — 2026-11-11T12:03:25.040Z Published October twenty-eighth, twenty twenty-five, Rebecca Bellan reports that OpenAI CEO Sam Altman expects an intern-level research assistant by September 2026 and a fully automated “legitimate AI researcher” by 2028. OpenAI chief scientist Jakub Pachocki defines that researcher as a system able to deliver larger research projects autonomously, not a human who studies AI. He says OpenAI is pursuing algorithmic innovation and more “test time compute”—time and computing resources spent thinking through problems. Pachocki says current models handle tasks with roughly a five-hour time horizon, and argues that dedicating data-center-scale computing to a problem could be worthwhile for major scientific breakthroughs. These are OpenAI leaders’ forecasts and plans, not reported results. The article links the timetable to OpenAI’s restructuring as a public benefit corporation and Altman’s account of the infrastructure investment it would require. OpenAI says its nonprofit foundation will govern research direction and manage research and safety initiatives. Whether the projected research capabilities will arrive on this timetable, and what scientific work they might deliver, remains unknown. ## Original source TechCrunch · Article Original source date: Oct 28, 2025 (2025-10-28). Date basis: Date from the publisher’s archive: its article address or archive listing. Added to Escape Meridian: 2026-11-11T12:03:25.040Z. First observed: 2026-10-28T11:33:52.828Z. Retrieved: 2026-10-28T11:35:50.630Z. - [Sam Altman says OpenAI will have a ‘legitimate AI researcher’ by 2028](https://techcrunch.com/2025/10/28/sam-altman-says-openai-will-have-a-legitimate-ai-researcher-by-2028/) ## Limits Text as retrieved on October 28, 2026; originally published October 28, 2025. The article reports OpenAI leaders’ forecasts and plans; it provides no results demonstrating that the projected AI research capabilities have been achieved. ## Open question What evidence will show whether an AI system can autonomously deliver larger research projects, rather than only perform well on bounded tasks? ## Source-stated Sam Altman said OpenAI was tracking toward an intern-level research assistant by September 2026 and a fully automated “legitimate AI researcher” by 2028. > intern-level research assistant by September 2026 and a fully automated “legitimate AI researcher” by 2028, CEO Sam Altman said - [Supporting source](https://techcrunch.com/2025/10/28/sam-altman-says-openai-will-have-a-legitimate-ai-researcher-by-2028/) ## Source-stated Jakub Pachocki described an AI researcher as a system capable of autonomously delivering larger research projects. > “system capable of autonomously delivering on larger research projects.” - [Supporting source](https://techcrunch.com/2025/10/28/sam-altman-says-openai-will-have-a-legitimate-ai-researcher-by-2028/) ## Source-stated Pachocki said OpenAI was pursuing algorithmic innovation and scaling up “test time compute.” > continued algorithmic innovation and dramatically scaling up “test time compute” - [Supporting source](https://techcrunch.com/2025/10/28/sam-altman-says-openai-will-have-a-legitimate-ai-researcher-by-2028/) ## Source-stated Pachocki said current models can handle tasks with a roughly five-hour time horizon and that dedicating data-center-scale computing to a problem could be worthwhile for major scientific breakthroughs. > Current models can handle tasks with a roughly five-hour time horizon - [Supporting source](https://techcrunch.com/2025/10/28/sam-altman-says-openai-will-have-a-legitimate-ai-researcher-by-2028/) ## Source-stated Altman said OpenAI had committed to 30 gigawatts of infrastructure, a $1.4 trillion financial obligation over the next few years. > Altman said OpenAI has committed to 30 gigawatts of infrastructure, which is a $1.4 trillion financial obligation, over the next few years. - [Supporting source](https://techcrunch.com/2025/10/28/sam-altman-says-openai-will-have-a-legitimate-ai-researcher-by-2028/) ## Source and contributor context TechCrunch · Publisher The publisher is identified here. No individual author has been inferred. ## Source media Media retrieval status: unchecked. ## Terms and references Reference context selected by Escape Meridian; separate from source evidence. Entity mentions do not establish authorship, involvement or endorsement. ## Test-time compute Extra computation a model spends while answering, for example by reasoning longer, rather than during training. - [Definition reference](https://arxiv.org/abs/2408.03314) ## Time horizon METR's measure of AI agents' ability: the length of task, timed by how long skilled people take, that an agent completes at a set success rate. - [Definition reference](https://metr.org/blog/2025-03-19-measuring-ai-ability-to-complete-long-tasks/) ## Public benefit corporation A for-profit company that is legally required to weigh a stated public purpose alongside profit. - [Definition reference](https://en.wikipedia.org/wiki/Benefit_corporation) ## OpenAI OpenAI develops AI models and products and publishes research, release announcements and evaluations. - [Explore this context](/entities/openai) - [Reference](https://openai.com/news/) ## Sam Altman Sam Altman publishes essays about AI and technology on his personal blog, separate from OpenAI announcements. - [Explore this context](/entities/sam-altman) - [Reference](https://blog.samaltman.com/) - [Reference](https://openai.com/our-structure/) - [Reference](https://blog.samaltman.com/posts.atom) ## TIME TIME is a news magazine publishing reporting on world affairs, business, technology and the workers behind digital services. - [Explore this context](/entities/time) - [Reference](https://time.com/) ## Editorial provenance Version 1. Publication ID: 645b7ffc-e7f2-7bbf-c9d2-f76374d6a036. Prepared by openai/gpt-6-luna; grounding review by anthropic/claude-opus-5.5. Source check: uncertain by anthropic/claude-opus-5.5 on 2026-11-11T11:59:07.744Z. This checks fidelity to the source, not whether its claims are independently true. Editorial selection: accept; reasons: bridge\_admitted\_by\_model. Observation: cd839b0fb696844915ee7ad788297412c6796d2b2fb3a3bc95b6423a18c0cde8. Source text hash: 282880c2d3c477026f57b5ad56f984a1cfb518ba33b1f324ae77a4252bad4c8f. ### uncertain — 2026-11-11T11:59:07.744Z — anthropic/claude-opus-5.5 Publication: 645b7ffc-e7f2-7bbf-c9d2-f76374d6a036 attribution — uncertain: Credit to Rebecca Bellan at TechCrunch and the quotes from Altman and Pachocki are correct. However, the summary sentence 'He says OpenAI is pursuing algorithmic innovation and more test time compute' and claim 3 attribute the two-strategy statement to Pachocki. In the source, the strategy sentence is the reporter's own framing and is not tied to a speaker. Smallest repair: attribute it to the article ('The article says OpenAI is betting on...'). > To achieve those goals, OpenAI is betting on two key strategies: continued algorithmic innovation and dramatically scaling up “test time compute” summary — pass: The reading leads with the forecast timeline, the actors and the definition of an AI researcher. It explains test time compute in plain language, and the definitions cover time horizon and public benefit corporation. It clearly marks the forecasts as plans rather than results and links them to the restructuring and infrastructure. The question follows from the evidence. The superintelligence quote is omitted, but that does not distort meaning. date — pass: originalPublished 2025-10-28 at day precision matches the reading's date, and the livestream 'Tuesday' is consistent with that date. The page was modified the same day. The Most Popular sidebar items appear to be later additions, but the reading does not use them. Forecasts are reported as forecasts, not outcomes, and there is no hindsight framing. support — pass: The timeline, the AI researcher definition, the five-hour time horizon, the data-center-scale compute remark, the 30 gigawatt and $1.4 trillion commitment, and the role of the nonprofit foundation are all grounded in the text. The limitations correctly state that no results are demonstrated. Apart from the speaker attribution of the strategy sentence, which is noted under attribution, no benefits or costs are invented. > Current models can handle tasks with a roughly five-hour time horizon and match top human performers in competitions like the International Mathematical Olympiad, Pachocki said. media — pass: No media was selected and no verified candidates were supplied, so this is consistent with the evidence. ## Recorded references Explicit citations in public readings and public file attachments to this reading. Shared topics are not citations. Other artifact types need an explicit reference adapter. No public references recorded here yet. [Reference data (JSON)](/readings/fabb7cba019802a8592f4c6950232f1eeb4f491b577ab0ab5dafaacb390a6c86/references.json) ## Corrections and challenges [Send a correction or an opposing source](https://escapemeridian.com/corrections/) [JSON record](https://escapemeridian.com/readings/rebecca-bellan-reports-openai-s-ambitions-for-ai-research-fabb7cba.json)