
Sep 8, 2026 · 25 min
OpenAI bets computer-using agents will replace the chatbot box
Greg Brockman on Astra and the Future of OpenAI
The conversation traces how stronger reasoning, computer use, memory, and health tools could shift AI from answering questions to pursuing supervised goals.
- 1OpenAI sees Astra as a step from conversational AI toward agents that operate computers on a user’s behalf.
- 2Brockman argues that capability thresholds matter because useful applications emerge only when models become reliably good at specific tasks.
- 3OpenAI’s health vision spans personal assistance, clinician tools, and hospital systems, with memory potentially connecting fragmented medical histories.
Don't miss
Brockman describes how improved memory might connect scattered symptoms over time, helping reveal patterns that fragmented specialist care can miss.
The brief
Greg Brockman opens with OpenAI’s claimed Navier–Stokes counterexample, presenting advances in reasoning as a possible engine for broader mathematical and scientific discovery.
Astra marks a shift from chat to computer use: people can set goals while AI handles more of the low-level operation, with human oversight still in the loop.
The discussion frames Codex, ChatGPT, desktop products, and local hardware as pieces of a more unified system rather than isolated tools.
OpenAI’s challenge is no longer only building capabilities; it must show users what models can do through concrete workflows instead of another text box.
The sharpest example comes in health, where memory could connect symptoms across fragmented conversations while products serve consumers, clinicians, and hospitals.
Brockman describes Operator as an imperfect but necessary predecessor to Astra, showing how real-world deployment helped OpenAI cross a computer-use threshold.
What was said on this episode
15 statements · 13 positive · 2 negative
AI models can generate knowledge and help solve otherwise inaccessible problems.
“we can actually generate new knowledge that we can learn from these models to have them help us solve problems that are otherwise outside of reach or would take us very long to solve.”
Listen at 1:21
Advanced AI models will make new medicines and disease cures more feasible.
“we're going to be able to develop, all of that starts to become much more real when you have models that are at this level of assistance and capability.”
Listen at 2:32
Astra can operate across applications beyond previous computer-use capabilities.
“we've reached a new threshold of computer use. So this model is able to really work with different kinds of applications in a way that was not previously possible.”
Listen at 3:33
ChatGPT has more than one billion weekly users.
“we're now well over a billion users every week”
Listen at 6:37
AI systems are moving toward operating computers according to users’ goals.
“we're moving towards real amplification, real giving you time back, real... having computers that are able to operate according to your goals, to your desires.”
Listen at 7:39
ChatGPT should proactively recommend ways to enable additional task capabilities.
“it should also be able to proactively say to you, hey, actually, if you ask me this other way, or if you added this connector, or if you authorize me to do this, or if you hook up your credentials in this way, I can go and do this other thing for you.”
Listen at 10:03
Three hundred million people use ChatGPT weekly for health queries.
“If you look at how many people use chat for health, 300 million people every single week with health queries”
Listen at 10:59
Higher AI capability thresholds unlock fundamentally new applications.
“as you hit new thresholds of capability my experience has always been that fundamentally new applications become unlocked”
Listen at 15:13
AI image-generation applications represent a large market opportunity.
“there's a huge market there.”
Listen at 15:54
AI healthcare applications will substantially improve many people’s quality of life.
“It's going to be absolutely transformative to many people's quality of life, really uplift so many people.”
Listen at 18:37
Many people may have undiagnosed conditions because care remains fragmented.
“it's really hard to say how many people have. similar kinds of conditions and just never find out”
Listen at 20:25
A broadly capable AI could assist across life domains and provide medical expertise.
“one of the powers and potentials of an AI that really deeply is able to help you across all parts of your life. And also is a deep domain expert in all areas of medicine.”
Listen at 20:41
OpenAI’s Operator was below the capability threshold for effective computer use.
“I think operator was just kind of below threshold in terms of the model capabilities.”
Listen at 21:32
Astra has reached the threshold for effective computer-use assistance.
“i think we're there with astra”
Listen at 22:42
Users will discover many additional applications for Astra.
“there's just so much more that people are going to uncover in terms of applications where this can go.”
Listen at 22:44
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.