3,698 items tracked: 33 reported by two or more outlets, the rest by a single source. Most-corroborated first.
The latest example of workers under surveillance comes from Co-op, a British consumer cooperative that runs thousands of businesses This was originally published in TechScape, a newsletter about how technology shapes our…
It's more than just an enlarged Apple Watch.
Character.AI, a controversial generative AI startup that lets users create their own chatbot characters, has filed for Chapter 11 bankruptcy
OpenAI is hoping that Dots will make automating online tasks a lot easier.
96-chip configuration promises seven times its predecessor's token throughput, with production deployments due in 2027
The full TechCrunch Disrupt Stage lineup revealed, featuring Max Hodak, Mark Wahlberg, Benchmark partners, and more.
Ubitium reported $14m
How should humans relate to generative AI ?
American Express is introducing a new framework designed to ensure AI agents only make purchases their users have authorized, potentially preventing costly mistakes and disputes as AI-powered shopping grows.
Suki growth reported $90m
The company’s new releases point to an AI market becoming more personal, with agents and models completing tasks, but enterprises still aren’t ready for this new phase of AI.
Financial watchdog fears AI spending bust could rattle markets, with mounting debt and cyberattacks adding to the headache
Clueso reported $10m
David Africa & Nick Levine TLDR We ran several persona experiments on Talkie, a 13B model trained only on English text from before 1931.
NVIDIA RTX Spark is a new beginning for the PC — a reinvention combining the capabilities of a gaming PC, a creative studio and a personal AI supercomputer into a single superchip.
This is today’s edition of The Download, our weekday newsletter that provides a daily dose of what’s going on in the world of technology.
FLÆ Robotics seed reported €1m
Anthropic says it has cut the extraction of Claude's internal reasoning tokens by 90 to 99 percent.
In a world of AI Overviews, Wikipedia usage declines.
(Soft prerequisites: A high-level model of AI bargaining and Taboo “equilibrium”: Less confused frames for research on AI bargaining .) Section II.1 of CLR’s safe Pareto improvements (SPI) agenda briefly reviews the chal…
Backed by Silicon Valley titans, Leading the Future has raised huge sums to promote pro-AI candidates on both side of the aisle.
Causa Prima reported $20m
Molmo 2’s open release helped MIT researcher Kai Van Brunt develop Molmo2Fish, an experimental model that tracks salmon in sonar footage and corrects some tracking errors using everyday language.
Ministers promised faster decisions, but campaigners say the technology is making mistakes while asylum processing times are getting worse
Companies in France are deploying agentic computing across industries spanning space, financial services, the public sector and scientific research, demonstrating how the country’s ecosystem has moved from ambition to ap…
Release: web-edit 0.1a1 Includes performance optimizations written by Claude Opus 5.5 after I fed it this Python 3.15 profiling report generated by GPT-6 Astra.
Intel Corp.
Committee chair rejects criticism that letter was ‘pre-determined’ by the government Follow our Australia politics live blog for latest updates Get our new political email , free app or daily news podcast Anthony Albanes…
For most languages of the world, language model pre-training operates in a data-constrained regime where models must repeat their training data many times, degrading generalization.
We present SWE-MISBEHAVIOR, the first taxonomy and evaluation framework for diagnosing boundary-violating misbehaviors in coding agents.
Off-policy data in reinforcement learning with verifiable rewards (RLVR) can arise from asynchronous generation, rollout-training mismatch, and response reuse.
A robot needs a world model that predicts how its actions will change the scene and a policy that chooses those actions.
Reconstructing complete, high-quality meshes of large-scale scenes from unposed multi-view images requires faithful geometric grounding while plausibly completing surfaces that are unobserved or only sparsely captured.
Large language models have evolved from single-turn chatbots into interactive agents capable of long-horizon reasoning.
Sparse-to-dense motion completion recovers dense mesh motion from sparse vertex displacements.
Reliable video reasoning requires models to produce correct answers and expose supporting visual evidence.
Vision-language-action (VLA) policies excel at generalist robotic manipulation but generate actions without explicitly modeling their environmental consequences.
We introduce Fish Audio Drama 3, an open-source speech generation model that turns text into directed performances, built on a 35B-parameter mixture-of-experts backbone that activates 3B parameters per token.
Visual agents must turn video observations into timely decisions to track progress and choose what to do next.
Rubric-based reward models (RMs) evaluate responses against explicit criteria.
Learning humanoid loco-manipulation from human demonstrations rests on observing how locomotion, posture, and object interaction evolve together as tasks unfold.
Release: llm-meta-ai 0.2 Meta AI release Muse Wonder (and Muse Wonder Full), their new "most capable models".
Meta Platforms Inc.
Enterprise AI scalability is running into a hard budget wall for some customers.
Good empirical safety research is heavily bottlenecked on thinking more clearly.