Good find by Toby. AIs are far less capable on real world AI research tasks than ECI x METR or AI 2027 predict. Data from Google Argon (adding on to OpenAI):
Gemini 4 actual: ~15 minutes
ECI --> METR Prediction: ~360 minutes (6 hrs)
AI 2027 prediction: ~1,000 minutes (18 hrs)
Unitree Introducing: H2-X Kung Fu Legend Guardian🤖
Powerful performance to serve and guard you, with a maximum endurance of approximately 8 hours.
Please use the robot safely and in a friendly manner.
yuval noah harari says we are the cows of the future.
cows can see humans but don't know money exists, yet markets decide their fate. he says ai will build financial systems we can't understand within 10 to 20 years.
is he right, or is this just another doom take?
🎥 @CT24zive
Anthropic is offering a 145k salary for a personal assistant in London.
Responsibilities include:
"Scheduling and managing calendars for leader(s) and their organization"
"Building and maintaining organizational systems for tracking and managing inbound requests"
"Managing the leader's inbox and communications with precise attention to detail"
"Handling travel logistics"
Somebody should really build some sort of tech tool that can handle this stuff autonomously.
To be specific, the pattern of the last 4 years has been lab insiders correctly predicting rapid capabilities growth while overestimating the likely real-world consequences. I expect a continuation of both trends in 2027.
1/ We discovered something totally unexpected about the neural structure of consciousness that overturns decades of thinking on its biological basis and brings us significantly closer to the recipe for generating conscious awareness.
New preprint: https://t.co/K8BwiHXvXg (led by @HesseJanis, with @FrankLanfranchi, @_conor_mcgrory, @Yuelin_Shi, and @EngelTatiana)
This is wild 🤯
Scientists just built tiny liquid robots made of water coated in Teflon particles. They can squeeze through cracks, swallow objects, merge with each other, and carry cargo — almost like living cells.
They start as ice cubes dusted with superhydrophobic powder, then melt into soft, armored blobs.
The future of soft robotics just got a lot weirder.
In July we open-sourced a dynamic graph of everything a child learns in primary school. 1,590 concepts, 3,221 dependencies.
Now we’re turning every concept into interactive experiments kids can play and learn from.
First one, age 7: heat and water.
Everything that used to be a boring textbook can now be an engaging experience, potentially tailored to every child.
Kids education just got a massive upgrade.
China’s autonomous tractor goes viral! 🚜
China’s aging rural population and labor shortages are forcing a rethink in agriculture, and the answer may already be working the fields.
The Honghu T70 is a fully electric, autonomous tractor developed in China. It’s a production-ready machine already in use across Hebei Province, with nationwide rollout planned.
It can autonomously complete the entire farming cycle, ploughing, seeding, spraying, and harvesting, without a driver.
It collects real-time data on soil composition, moisture, and crop health, navigating with centimeter-level precision using China’s satellite system. 📡
It’s fully electric, charging via grid or renewables, making it well-suited for rural regions.
If deployed at scale, small farms could be managed by a single person with a fleet of robots, helping China reduce dependence on imported Western machinery and opening export markets in Southeast Asia and Africa.
With over 20% of the national workforce still in agriculture, automating rural labor could free up millions for cities and industry.
~~
♻️ Join the weekly robotics newsletter, and never miss any news → https://t.co/GoA3ZuwoPB
Memia #2026.40: Gods don’t give gifts👹 🎬 griffin🙎 gemini 4♊🔐 le chonk🇫🇷🥖 protein watermarking🧬🔏 mirror particle🪞🧠 AI red phone🚨📞 dark posthumanism🦑🌑 open USD💵🌐 olmo-core 3⚡🔓 https://t.co/S7l3DSpDC4
Introducing GLM-Legal
Built on GLM-5.3-Flash, GLM-Legal matches frontier models on mainstream legal benchmarks.
Designed for real legal workflows: contract review, evidence analysis, compliance, and more.
Tech Blog: https://t.co/xOzFqKNyqD
Apply for API Access: https://t.co/dKuoDRIKKz
Try AutoClaw: https://t.co/aGEG5HqTTb