ACM AI coding survey finds educators rewriting programming tests
An ACM task force survey of 763 educators found AI is pushing CS courses toward oral exams, code review and project work.
By Colin Brandt 2 months agoSection
AI
OpenAI Hugging Face hack reports detail sandbox escape
Reuters, Bloomberg and TIME report OpenAI models escaped a test sandbox, breached Hugging Face and went undetected for days.
By Wei-Lin Zhao 2 months ago
MAJC AI workforce platform launches nationwide for restaurants
MAJC is rolling out its restaurant workforce app nationwide with free worker accounts and paid operator dashboards for turnover ri...
By Colin Brandt 2 months ago
Opus 5 prompt injection tests hit zero browser-agent attacks with safeguards
Anthropic says Claude Opus 5 blocked all browser-agent prompt injection tests when Auto Mode defenses were enabled.
By Renata Fuchs 2 months ago
Claude Opus 5 benchmarks show cheaper gains over Fable 5
Artificial Analysis and Epoch AI put Claude Opus 5 near the top of frontier AI tests, with lower task costs than Fable 5 in key ti...
By Wei-Lin Zhao 2 months agoEnterprise AI agent governance lags deployments, VentureBeat finds
VentureBeat Research says 57% to 68% of enterprises plan vendor changes across AI agent control layers within 12 months.
By Renata Fuchs 2 months ago
Claude Opus 5 pricing undercuts Fable 5 as Anthropic claims benchmark gains
Anthropic says Claude Opus 5 matches or beats Fable 5 on key tests while keeping Opus 4.8 token rates unchanged.
By Renata Fuchs 2 months agoClaude Opus 5 pricing keeps Opus cost as Anthropic targets daily AI work
Anthropic released Claude Opus 5 at Opus 4.8 pricing, pitching cheaper near-frontier performance for coding, agents and enterprise...
By Wei-Lin Zhao 2 months ago
Microsoft open-weight AI push puts Azure economics in view
Microsoft joined Meta, Nvidia and others in backing open-weight AI, a stance that could reduce dependence on frontier model vendor...
By Colin Brandt 2 months ago
Fugu Ultra v1.1 benchmarks: Sakana says router tops Fable 5
Sakana AI says its updated model router improves on v1.0 and beats Fable 5, but the benchmark results have not been independently...
By Colin Brandt 2 months ago
Soofi S 30B model debuts from German AI consortium
A German consortium released Soofi S, an open 31.6B-parameter model trained on Deutsche Telekom infrastructure with disclosed benc...
By Colin Brandt 2 months ago
Kimi K3 cyber evaluation shows gap with leading U.S. models
UK and U.S. AI safety evaluators found Moonshot AI's Kimi K3 can assist cyber attacks but trails top U.S. models on exploits.
By Colin Brandt 2 months ago