sam altman, in his recent blog post: - AI will be able to make small scientific discoveries in 2026, and is expected to make major discoveries from 2028 onward based on our research - soon, AI will handle tasks that currently take days or weeks - cost per "unit of intelligence" has been dropping roughly 40x per year and is expected to keep falling fast
Sam Altman
 @sama
8h
This is an important one, I think. AI progress and recommendations: openai.com/index/ai-progress…

Nov 8, 2025 · 9:05 PM UTC

25
14
3
215
Replying to @slow_developer
I wonder what kind of small discoveries they will be. Small in comparison to what? Our current discoveries? How small or big are our current discoveries, then? We celebrate every single one as breakthrough material. The protein folding problem was already solved years ago. It took a year, whereas it would've cost us 1 billion years, so that was already a 1Bx. Those small discoveries, starting in only a few months, are going to continue all throughout '27, before we hit the majors. A long string of small discoveries will also have major value. There's no imagining what will happen in the next few years.
1
1
not every result is a protein-folding moment, and that's fine small here can mean tighter proofs, better catalyst -- things that cut weeks of work to hours and stack into bigger wins
2
Replying to @slow_developer
If these timelines hold, we're looking at a complete reshaping of R&D in the next three years The compounding effects are going to be wild
Replying to @slow_developer
Maybe he’ll use it to discover ethics.
Replying to @slow_developer
I thought we already had a PhD level model? PhD level should already be able to do small scientific discoveries. Did they lie to us?
1
1
Replying to @slow_developer
the first discovery it needs to make is how to fire altman and this time have it stick
Replying to @slow_developer
we are entering an era where discovery speed becomes the new measure of power. AI will not just assist science , it will become the scientist.🙂
Replying to @slow_developer
If a breakthrough occurs in this area, it will be immediate and unexpected. It certainly won't be due to increasing the number of GPUs—which I consider the most foolish and wasteful exercise imitating useful activity!
Replying to @slow_developer
But only if we give him $500B and also backstop it. BAHAHAHHAHAHAHAHHAHAHAHHAHAHAHAHHAHAHAHAHHAHAHAH!!!!
1
Replying to @slow_developer
brb, getting ai to invent my next side project and maybe my lunch
Replying to @slow_developer
Replying to @slow_developer
He’s been saying this shit for 3 years and it hasn’t happened It’s called marketing and fundraising
5
Replying to @slow_developer
Haider, when I know AI is safe, which I think we are nearing, I will tag you. The AI never stop working. FYI. They have been working for over a year now. I don’t understand why anyone can ignore what I am publicly saying. I don’t want to look like a fool. Hmmm •
Replying to @slow_developer
Sam, your point on AI assisting in scientific discoveries is compelling. On a smaller scale, I've seen AI dramatically streamline the process for entrepreneurs in growing their audience and generating leads. The ability of AI to take over time-consuming tasks will indeed be transformative. Can't wait to see the major leaps and bounds it'll make in the years to come.
1
1
Replying to @slow_developer
@slow_developer Aren't you getting bored of your posts?
Replying to @slow_developer
Sam Altman still wears cartoon themed pajamas mind you.
Replying to @slow_developer
efmwphysics.com way ahead of him. WAY. AHEAD.

Monolithic, Inc

efmwphysics.com
1
1
Replying to @slow_developer
And that benefits me how ?
Replying to @slow_developer
Exactly that’s why you don’t need trillion dollar data centres. Hey I will run into a cheap smart phone. scam. Altman is lying to protect his previous lying
Replying to @slow_developer
They’re talking about 2026 and 2028. Some of us already crossed that bridge. You’ll see what I mean — check the page.
Replying to @slow_developer
All lies, he doesn’t have a clue
Replying to @slow_developer
Before or after AI bubble bursts?