News Archive
Pick a date to read that day's stories.
Days with a blue dot have stories.
- 01 of 10OPENAI
OpenAI builds its first own chip
OpenAI created a new computer chip, Jalapeño, to power its AI models. This chip reduces its reliance on other companies for hardware.
OUR TAKERenting Nvidia was never going to be the long-term plan; now the landlord has competition.
Read more → - 02 of 10WHITE HOUSE
White House tells OpenAI to slow down
The White House told OpenAI to release its new AI model gradually. They asked for this slow rollout due to safety worries.
OUR TAKEWashington has finally noticed the speed dial only turns one way.
Read more → - 03 of 10ANTHROPIC
Anthropic accuses Alibaba of mass fraud
Anthropic claims Alibaba generated 28.8 million fake conversations. They believe Alibaba used this to copy their Claude chatbot's responses.
OUR TAKETurns out the polite British chatbot has been getting prank called at industrial scale.
Read more → - 04 of 10GOOGLE
Four senior Gemini staff jump to Anthropic
Four senior people from Google's AI team have left for Anthropic. This suggests Google is having trouble keeping its top AI talent.
OUR TAKEWhen the engineers start leaving in groups, the org chart usually follows.
Read more → - 05 of 10MIT
Robots learn to ignore the clutter
Robots can now understand vague commands like tidy the kitchen. New techniques help robots learn what things are important.
OUR TAKEFinally, a machine that understands "you know, the thing, over there".
Read more → - 06 of 10SECURITY
Anthropic AI passes a secret NSA test
Anthropic's new AI model performed well in a US intelligence cyber test. Officials are worried similar technology could aid attackers.
OUR TAKEPass the test too well and the prize is a paperwork avalanche.
Read more → - 07 of 10EU
Brussels delays its own AI rulebook
Europe pushed back deadlines for its AI rules, giving companies more time to comply. Businesses said the original timeline was too short.
OUR TAKEEven the regulators needed an extension.
Read more → - 08 of 10EUROPE
Europe says enough, builds its own AI
Europe is pooling resources to create its own AI models. They want to avoid relying on tech from other countries.
OUR TAKESovereignty is back in fashion, and this time it comes with a GPU bill.
Read more → - 09 of 10META
EU turns up the heat on Meta
The EU is expanding its investigation into Meta's Facebook and Instagram. They want to know if the apps are designed to be addictive. Fines and app changes are possible if found true.
OUR TAKE"Just one more scroll" may soon come with a legal disclaimer.
Read more → - 10 of 10OPENAI
GPT-5.6 will arrive in pieces
OpenAI will release its new GPT in stages. This cautious rollout, influenced by the White House, aims to fix errors early.
OUR TAKEBig launch day is out; soft launch season is in.
Read more →