Sam Altman says OpenAI will have a 'legitimate AI researcher' by 2028 | TechCrunch
Briefly

Sam Altman says OpenAI will have a 'legitimate AI researcher' by 2028 | TechCrunch
""We believe that it is possible that deep learning systems are less than a decade away from superintelligence," Pachocki added."
""system capable of autonomously delivering on larger research projects.""
""test time compute""
""legitimate AI researcher""
OpenAI is tracking internal timelines that aim for an intern-level research assistant by September 2026 and a fully automated "legitimate AI researcher" by 2028. The organization finalized its transition to a public benefit corporation, removing nonprofit charter limits and enabling new capital-raising opportunities. The firm plans to pursue algorithmic innovation and dramatic scaling of "test time compute" so models can spend far longer thinking through problems. Current models have roughly a five-hour time horizon and can match top human performers in competitions like the International Mathematical Olympiad. For major scientific breakthroughs, OpenAI would consider dedicating entire data centers to single problems, linking these efforts to accelerating scientific discovery and potential superintelligence.
Read at TechCrunch
Unable to calculate read time
[
|
]