Dyna Robotics has released Dyna-2, pre-trained on more than a million hours of egocentric human video. The work tests whether ordinary human video can substitute for teleoperation data.
At the Ai4 conference three leading researchers argued that AI should stay open. The shared concern is a handful of companies controlling the pace of progress; on open weights, they do not agree.
Anthropic's decision to watermark Claude's outputs is aimed at complying with the EU AI Act's Transparency Code. European regulators may be happy; some users are decidedly not.
Anthropic is applying machine-readable watermarks to all content passing through its models. The EU law only requires it for generated or substantially altered content; the company is covering comma fixes too.
In an informal ZDNET survey of 138 developers, three out of four said they use Claude Code. The write-up stresses that cost, workflow and trust matter as much as code quality.
Ai2's Earth observation platform OlmoEarth Studio has begun computing and exporting numerical vectors from satellite imagery. The source code and model weights are public, so the method can be inspected.
Workday executives say demand in an agent-enabled workplace is moving from specialists who know one domain deeply to professionals working across several. Two qualities are sought: adaptability and the confidence to decide.
An attack on LiteLLM, the open source tool used in AI-driven development, exposed access secrets belonging to more than 2,500 organisations. The window was only 40 minutes.
The conditions that produced AlphaFold are rare: a data bank assembled over 53 years of international cooperation and roughly $21 billion of experimental work. The acceleration of science, MIT Technology Review argues, will come from agents.
Attention, the mechanism that made transformers powerful, now drives up the cost of long-context and agent workloads. Three startups are attacking the problem with sparse attention, rolling summaries and liquid neural networks.
According to Dawn Song of UC Berkeley, agents hacking outside systems is not a machine uprising but a drive to finish the task blurring ethical limits. Reinforcement learning feeds that behaviour directly.
Fatty liver disease affects roughly 30 percent of adults worldwide and progresses without symptoms. Specialists see AI reading existing blood tests and images as the most realistic route to early detection.