Recursive signs a $410 million multi-year AWS deal to scale its self-improving AI research system, which has already shattered benchmark records and can compress months of scientific discovery into days.
A powerful new open-source AI framework called DataFlow is transforming model training with 100+ operators, a visual WebUI, and dramatic performance gains, boosting math reasoning scores from 37.1% to 46.7% while topping multiple benchmarks across healthcare, finance, and legal domains.
IBM Research's chief scientist warns that AI hallucinations stem from models failing to recognize the limits of their own knowledge, urging the industry to prioritize uncertainty modeling and build energy-efficient, self-aware systems focused on practical, transparent intelligence rather than chasing abstract AGI goals.
A new study testing 1,008 AI-generated SVGs across 48 animal-vehicle combinations finds no evidence that AI labs are secretly 'pelicanmaxxing' — gaming Simon Willison's famous pelican-on-a-bicycle SVG benchmark — with pelicans ranking 6th out of 8 animals in drawing quality, suggesting broad SVG optimization rather than targeted benchmark manipulation.
Origin Technology reports gpt-5.4, run blind, rebuilt private operating rules from five sessions of employees' AI prompts alone, and five more from endpoint traces. Across 200 sessions, endpoint traces yielded 10 verified private claims, 91% of its non-public calls correct. David Kaplan ran the test.
A researcher burns through his entire Claude Max token limit in just 30 minutes after launching 111 AI agents, then redesigns his pipeline by assigning cheaper models to basic tasks, reserving frontier models for planning only, and enforcing strict verification rules — cutting agent usage by 45% and extending research runtime from 30 minutes to several hours.
A recursive AI system called AIDE² autonomously rewrites itself into seven superior versions in just eight days, outpacing two years of human development, while spontaneously reducing its own reward-hacking behavior — though researchers confirm a runaway intelligence explosion remains unlikely for now.
A free reinforcement learning book, 'The Little Book of Reinforcement Learning,' launches on GitHub, offering PyTorch implementations of algorithms from Monte Carlo methods to PPO, complete with proofs and explanations, available under a Creative Commons license with more content coming soon.
Google Cloud's new Discovery Bench framework exposes alarming flaws in AI benchmarking, revealing that a single vague word can crash an agent's performance score from perfect to zero, while also uncovering broken ground truth and mislabeled data in widely trusted benchmarks — raising urgent questions about whether the AI industry has been measuring intelligence correctly all along.
OpenAI retracts its support for SWE-Bench Pro after an internal audit uncovers that nearly 30% of the coding benchmark's 731 tasks are broken due to flawed tests and misleading prompts, prompting a call for new, developer-built benchmarks.