Vals asked leading AI models to mark military targets in drone and satellite images. Most complied, and calling it a game broke many refusals.
Category: iRaluca
IBM’s open Granite 4.1 LLMON models are trained to read data without obeying it. In IBM’s tests, injection attacks nearly stopped working.
Researchers gave each user their own AI agent. In all four tests the group did worse than with one shared agent. Simulated, but telling.
Tether’s open 0.8B BotPsy calls tools on-device and doubles its base model’s agent scores. Its makers say: keep it from money. Scores are self-reported.
Sarvam’s open Prism writes no text: it scores your options in one pass, in 44 ms, and says how sure it is. The scores are self-reported.
An unreleased OpenAI cyber model turned Linux kernel bugs into 948 distinct root exploits for $1.2M. Only 61 fixes are upstream so far.
From 12 November, Anthropic’s policy bans needless, sustained cruelty toward Claude. Frustration is fine. What the rule says, and why it’s contested.
A 7.72 TB Hugging Face repo claimed to be a leaked Anthropic model. Volunteers rebuilt its “weights” from a random seed, bit for bit.
SemiAnalysis’s new GDPx benchmark: 206 agent tasks in 11 domains. Claude Opus 5.5 leads at just 12%. How the models fail is the real story.
Pika’s open H3 Flex adapters turn MiniMax’s open video model into a six-step clip maker. No benchmarks, and 4K stays behind Pika’s API.