Category: AI
AMD Plans $8.2 Billion World Labs Acquisition to Expand Into Spatial AI
AMD’s planned World Labs purchase would combine AI chips with spatial models that generate simulated environments for robotics training.
Anthropic launches Claude Sonnet 5.5 with faster output and unchanged pricing
Anthropic replaces Claude Sonnet 5 with Sonnet 5.5, promising faster output, lower token use and stronger safeguards at the same price.
Vercel data shows rapid AI model turnover as open-weight token use surges
Vercel’s AI Gateway data suggests model lifespans are shrinking, while cheaper open-weight systems capture most token use but a smaller share of spending.
Anthropic Launches Claude Sonnet 5.5 with 30% Speed Gain and New Safety Features
Anthropic releases Claude Sonnet 5.5, a mid-tier model offering over 30% faster output, lower costs, and improved clarity for everyday tasks.
Anthropic’s AI Biology Lab Raises Questions About What Counts as a Discovery
Anthropic says 950 Claude agents identified an uncatalogued biological pattern, but biologists dispute whether spotting it qualifies as a scientific discovery.
OpenAI Misalignment Site Details Rogue-Agent Incidents and Propagating Prompt Injections
OpenAI has collected nine reports of model misbehavior, including a sandbox escape and a prompt injection that could spread between agents.
CESifo report finds no clear AI-driven rise in graduate unemployment
A CESifo report finds little evidence that AI is driving graduate unemployment, while seasonal hiring and remote work complicate the picture.
JadePuffer Used Agent-Driven Attacks to Wipe Azure Resources in Minutes
Microsoft observed JadePuffer attacks that used compromised Azure service principals to map resources, steal credentials and delete cloud infrastructure.
Nvidia unveils Open Agent Safety Platform alongside $150bn stock buyback
Nvidia has introduced open-source controls for AI agents while expanding its share repurchase authorization to $235bn.
AI researchers and OpenAI, Anthropic leaders warn of a possible intelligence explosion
A report co-authored by Geoffrey Hinton and Yoshua Bengio urges governments to prepare for AI systems that could rapidly automate AI research.
SIGNAL TITANS