AI Research | Tech
Tech news, context and sources.
Latest articles
Anthropic reports early, largely preclinical drug-development work with Claude
Design and documentation improvements are not evidence that the resulting medicines work in patients.
Qwythos 9B v2 runs on an 8 GB GPU, with important limits
A quantized RTX 3070 test shows practical local use, not verified Claude equivalence or million-token context.
Pennsylvania enacts disclosure rules for deceptive AI advertising
Scope depends on significant AI changes and misleading impressions, with exceptions for some expression.
Patent-lawyer trial separates AI productivity from skill development
A three-month experiment improved drafting quality, but unaided learning gains differed by seniority.
Dismissed OpenAI safety researchers dispute the stated grounds
The dispute raises questions about clear rules for outside safety collaboration and confidentiality.
Forced-choice API test finds different party preferences across models
The results concern APIs required to select candidates, not ordinary consumer-chatbot answers.
Apollo questions whether AI gains extend beyond big tech
Aggregate company indicators and concentrated spending support an interpretation, not a causal finding.
You’ve reached the end.