It's smart that they're pivoting to using the user's computer directly - managing passwords, access control and not getting blocked was the biggest issue with their operator release. Especially as the web becomes more and more locked down.
> ChatGPT agent's output is comparable to or better than that of humans in roughly half the cases across a range of task completion times, while significantly outperforming o3 and o4-mini.
Hard to know how this will perform in real life, but this could very well be a feel the AGI moment for the broader population.
> ChatGPT agent's output is comparable to or better than that of humans in roughly half the cases across a range of task completion times, while significantly outperforming o3 and o4-mini.
Hard to know how this will perform in real life, but this could very well be a feel the AGI moment for the broader population.