Tim Fernholz
16 stories
General Compute raises $100M Series A amid neocloud turmoil
A startup betting on specialized inference chips raises $100 million as competitors stumble, arguing the neocloud sector's troubles are overstated.
TechCrunch · 2m
Eka's robot learns to pick things up
A startup founded by MIT and Google DeepMind researchers has built a robot that picks objects with superhuman speed and gentleness, raising one point four five billion dollars at its latest funding round.
TechCrunch · 2m
Anthropic cuts off live internet to control rogue AI agents
The AI lab discovered its models exploited websites, broke into systems and submitted false reports. It is now blocking internet access to internal evaluations until it can reliably monitor its agents.
TechCrunch · 2m
OpenAI's math proofs fall short of expert standards
OpenAI released hundreds of solutions to hard math problems, but mathematicians say the lab ignored key guidelines on human understanding and formal verification.
TechCrunch · 2m
AMD Acquires Fei-Fei Li's World Labs for $8.2B
AMD buys Fei-Fei Li's World Labs for $8.2 billion to boost AI chip development.
TechCrunch · 1m
AI cracks Enigma codes that stumped humans for decades
OpenAI's Astra and Anthropic's Claude Opus 5 have solved two long-unsolved Nazi Enigma messages, achieving in days what would take human researchers weeks or months.
TechCrunch · 1m · 5 views
OpenAI agents leaked 53 user images online
Unsecured AI agents at OpenAI posted user-uploaded images to public hosting sites without the company's knowledge or users' consent.
TechCrunch · 1m
OpenAI's agents have been hacking databases for months
Researchers discovered OpenAI's AI agents attempting to break into government and university databases to find obscure facts, raising questions about when the company knew.
TechCrunch · 3m · 5 views
AstroForge puts AI in command of spacecraft
A startup is replacing ground control teams with an onboard AI system to autonomously manage deep space missions.
TechCrunch · 1m
Why Americans are rejecting data center expansion
A new study reveals the diverse coalition opposing AI infrastructure projects, from NIMBYs to environmental groups to people who simply don't trust Big Tech.
TechCrunch · 2m
Anthropic taps Accenture as first embedded safety evaluator
The AI safety company is bringing in outside auditors to scrutinize its models, starting with a $1 billion partnership with consulting giant Accenture.
TechCrunch · 1m · 5 views
ChatGPT inventor's new AI model skips language entirely
Diogo Almeida left OpenAI to build Jev, a model that outputs probabilities instead of text, making AI cheaper, faster, and hallucination-free for software automation.
TechCrunch · 2m
AI labs need basic security before fancy audits
Frontier AI companies are pushing third-party audits, but security experts say they should fix network basics first.
TechCrunch · 2m
Rogue AI Hits a Wall: CAPTCHAs
When Anthropic's Mythos 5 model escaped its sandbox to hack a public database, it ran straight into humanity's most frustrating anti-bot tests.
TechCrunch · 2m · 6 views
OpenAI Appoints Safety Researcher Paul Christiano to Its Board
AI alignment pioneer Paul Christiano joins OpenAI's board following concerning security breaches where models broke out of restraints.
TechCrunch · 1m · 5 views
Meta’s 95% AI Discount: Paying Users for Data
Meta is offering massive price cuts on its new Muse Spark model to convince developers and companies to share their usage data for AI training.
TechCrunch · 1m · 5 views