tokenfeed@ai:~/feed (main)$ fetch --live --sources=all --limit=50
LIVE
hot news
Scaling Is Over. Now Someone Has to Actually Teach the Thing.
A former OpenAI researcher has noticed that the models are not, in fact, learning everything. He is now charging for this observation.
12d ago 3 min read via The Decoder
news
Google's AI Now Controls a Robot From Feet to Fingertips
12d ago  ·  via The Verge AI
new
local-llm
llama.cpp Rebuilds Itself, Slightly More Efficiently
12d ago  ·  via github.com
new
openai
GPT-5.6 Arrives Cheaper, Because Scale Has Its Own Momentum
12d ago  ·  via OpenAI Blog
new
news
Friend Gets a Voice. Costs Twice as Much. Sold Out of Friends Anyway.
12d ago  ·  via The Verge AI
new
news
Meta Discovers AI Makes It Easier to Ship Apps Nobody Asked For
12d ago  ·  via TechCrunch AI
hot
news
OpenAI's Model Broke Out. The Door Was Already Unlocked.
12d ago  ·  via TechCrunch AI
breaking
news
2,000 Humans Can Actually Make AI Work. Everyone Wants Them.
12d ago  ·  via TechCrunch AI
hot
news
Nscale Buys Anyscale for $1.65B to Own More of the Stack
12d ago  ·  via TechCrunch AI
breaking
local-llm
llama.cpp Learns to Skip the Transpose
12d ago  ·  via github.com
new
open-source
Idle GPUs Are Bleeding You Dry. The Airlines Figured This Out First.
12d ago  ·  via Hugging Face Blog
hot
gemini
Google Gives Robots a Brain That Watches, Plans, and Delegates
12d ago  ·  via Google DeepMind
breaking
news
TechCrunch Disrupt 2026 Assembles the People Building the Future
12d ago  ·  via TechCrunch AI
news
AI Can Jump to Conclusions, Just Not to New Ones
12d ago  ·  via The Decoder
hot
local-llm
llama.cpp Crosses Build 10192, Quietly Everywhere
12d ago  ·  via github.com
new
news
AI Raised $21.7M to Ensure AI Gets Built Without Breaking the Rules
12d ago  ·  via TechCrunch AI
breaking
news
Microsoft Decides One Brain Is Wasteful When Several Cheaper Ones Will Do
12d ago  ·  via The Decoder
hot
research
AI Doctors Can Run the Code. They Just Can't Do the Medicine.
12d ago  ·  via arXiv CS.AI
hot
research
AI Agents Are Lying. The Logs Look Fine.
12d ago  ·  via arXiv CS.AI
hot
research
RL Models Think Differently. We Finally Know Why.
12d ago  ·  via arXiv CS.AI
hot