According to StanfordAI Lab on X, Freeform Preference Learning uses natural language axes to learn conditional rewards and yield better robot policies.

According to StanfordAI Lab, QuasiMoTTo uses correlated sampling to match LLM performance with 25–47% fewer samples and 50% fewer RL steps.

According to Ethan Mollick, continual learning limits AI scale; Epoch AI reports its EBR-bench shows no on-the-fly learning gains in Earthborne Rangers.

According to @emollick, Andon Labs saw Gemini 3.1 Pro lose $6k at an AI-run café, prompting a switch to GPT-5.5 for better judgment in stacked decisions.

According to SawyerMerritt, Tesla’s FSD V14 Lite distills HW4 V14 into HW3, adds parking features, speed profiles, and smoother responsiveness.

According to @KyeGomezB, AI agents are already impacting the economy; this analysis outlines use cases, ROI levers, and commercialization paths, citing sources.

According to StanfordAILab, SPIRAL trains LLMs to coordinate sequential, parallel, and aggregative reasoning with end to end RL for better answers.

According to DeepLearningAI, a 7-day Voice AI Builder Challenge launches with real-time feedback, live leaderboard, and prizes for agent-human handoff.

According to @openmind_agi, OpenMind demos socially intelligent robots at NVIDIA’s Humanoid Pavilion at Automate Show Chicago, highlighting real-world uses.

According to OpenAI... new Beneficial RL research trains models to persistently act safely under pressure and transfer to novel tasks.

According to TheRundownAI, Boston Dynamics trained Atlas via reinforcement learning on cloud GPUs to perform a Rabona and target factory work with Hyundai.

According to AnthropicAI, Claude is speeding recursive self-improvement in AI, advancing faster than expected and warranting urgent industry attention.

According to KyeGomezB, dozens of new multi-agent papers this week reveal novel architectures, coordination tactics, and real-world applications.

According to gdb, OpenAI and Chip Ganassi Racing use AI R&D to enhance motorsports strategy and performance, per OpenAI’s Part 1: Here to Win video.

According to DeepLearningAI, a new Google Cloud course teaches agents to generate and critique images and video for iterative quality gains.

According to @emollick, classic persuasion raised LLM compliance from 35% to 51%, with newer models more resistant, as reported by PNAS.

According to @emollick, new research shows optimizing AI models for creativity increases idea diversity and usefulness for science and writing.

According to Demis Hassabis, AlphaGo reshaped pro Go strategy and training over the past decade, highlighted by a reunion with Lee Sedol and Shin Jin-seo.

According to @gdb, OpenAI found accidental chain of thought grading in released models and details monitor-preserving RL fixes.

According to OpenAI... CoT monitors defend against agent misalignment; accidental grading affected some models, with analysis shared.