Demis Hassabis just moved his AGI guess from 3-5 years to 2-3 years.
In less than a year.
When the most careful person in the room speeds up his timeline, I pay attention. What changed?
Demis Hassabis says he expects AGI within two to three years, down from his three-to-five-year prediction earlier in 2026, and discusses AI's science potential (@bzcohen / Wall Street Journal)
(Visit Techmeme dot com for the link and full context!)
Most people think AI will speed up medicine by finding new drugs.
But a lot of the delay is the boring part: coordinating hundreds of steps between labs, factories, and trials.
OpenAI just bought a company built for exactly that.
Faster cures might start with less chaos, not just smarter models.
We’ve acquired Orchestra Bio, a platform that helps biopharma teams coordinate the complex process of drug development.
Together, we’ll build on Rosalind, our AI models and tools for life sciences research, to streamline those operations and help new treatments reach patients sooner.
Welcome @ashokaraj_ and team!
https://t.co/vi51OhznQn
Four AI agents just beat Minecraft with zero human help.
27 hours. About $1,620.
Not long ago, an AI just crafting a diamond pickaxe was big news.
What game should they try next? 🎮
Google DeepMind says four Gemini 4 Argon agents beat Minecraft without any human help in 27 hours and 14 minutes, costing $1,620 at Argon's planned 2027 price (@maba_xr / The Decoder)
(Visit Techmeme dot com for the link and full context!)
In Korea, falling stocks show up in blue.
This morning there's a lot of blue.
One DeepSeek update, and memory chip stocks drop because people fear leaner AI needs less memory.
But when AI gets cheaper, people usually use way more of it.
Overreaction or real warning?
South Korean stocks resumed declines Monday, as DeepSeek’s latest model sparked concerns over demand for memory chips, compounding market worries over AI financing https://t.co/nJ7FHc0NEn
A fruit fly brain is now flying through a world that doesn't exist. 🪰
Simulated brain + simulated world.
If this scales, how long until we test brain drugs in a sim before real trials? 👇
We placed the fruit fly brain inside Odyssey-3, a frontier world model.
The lil’ guy is now flying around totally imagined, simulated worlds all by itself 🫶
@PeterDiamandis Waiting 9 months for a specialist already sounds unbelievable today 😅 If AI can just handle the first check-up step, that alone would change so many lives.
An open-weight model just beat GPT-5.6 Sol and Opus 5 on a coding benchmark.
And it costs $0.15 per million input tokens off-peak.
The gap between open and closed AI is getting really small. Would you still pay for the expensive ones?
BREAKING 🔥: DeepSeek released DeepSeek-V4.1-Flash-1017, a new open-weight model with frontier performance.
> It scores 77% on DeepSWE 1.1 benchmark, outperforming GLM-5.3, Kimi K3 and models like GPT-5.6 Sol and Opus 5.
> The model is now available on DeepSeek Chat and APIs as well as HuggingFace.
It also extremely cheap! 👀
4 AI agents beat Minecraft from scratch in 27 hours. No humans.
The game isn't the story. They split the work, shared notes, and fixed each other's bugs. That's basically a research team.
Now point that team at aging biology. 👀
Four Gemini 4 Argon agents beat Minecraft from scratch in 27 hours.
Operating in a sandbox using only code and text, they wrote 2,783 programs, divided up jobs over chat, shared notes, and debugged errors together - all with no human help ↓
https://t.co/uyIhDOQXGT
@GoogleDeepMind 4 AI agents beat Minecraft from scratch in 27 hours. No humans.
The game isn't the story. They split the work, shared notes, and fixed each other's bugs. That's basically a research team.
Now point that team at aging biology. 👀
AI in biology just had a big day.
OpenAI is opening GPT-Rosalind to university labs.
Science published RFdiffusion3, an AI that designs proteins atom by atom.
One helps decide what to test.
The other designs what to build.
If this works, aging research could speed up a lot.
We're expanding access to GPT-Rosalind for academic research.
The new Rosalind Academic Plan helps scientific labs bring AI into the work of understanding biology, exploring ideas, and deciding what to test next—with sponsored access for eligible labs at U.S. universities.
Academic research starts with questions whose impact may take years to impact human health.