Ground Truth.
AI, checked against the source.
← 2026-06-262026-06-27later →

The government cleared one Anthropic model and kept the other locked up

2026-06-27

Washington partially reopened access to Anthropic's Mythos 5 for about a hundred organizations, but its more powerful sibling Fable 5 stays blocked - and Anthropic is still suing.

anthropic · policy · frontier-models · regulation · geopolitics

An AI's hallucinations turned out to be a map with blank spots

2026-06-27

Researchers showed that when a world-model AI imagines impossible futures, it's usually in places it barely saw in training - and that you can predict and fix those blind spots cheaply.

world-models · hallucination · robotics · research · training

This AI predicts how objects move by tracking shapes, not pixels

2026-06-27

PhysiFormer forecasts physical motion as real 3D meshes in space - and recovers rigidity and momentum without anyone hand-coding the laws of physics.

world-models · physics · diffusion · robotics · research

The trick that makes AI type faster just hit the top of Hacker News

2026-06-27

A small model guesses ahead and a big model checks the work in parallel - and this week two efforts pushing that idea, DeepSeek's DSpark and JetSpec, lit up the front page while the community argued over whether it's truly 'lossless.'

inference · efficiency · deepseek · open-source · speculative-decoding

Image generators can't plan. This one bolts on a brain that can.

2026-06-27

Qwen-Image-Agent wraps planning, reasoning, and memory around a text-to-image model so it can break a hard request into steps - and the local-AI crowd immediately asked whether it runs on a gaming GPU.

image-generation · agents · qwen · multimodal · research

A wave of new methods trains AI without a human answer key

2026-06-27

Several research groups landed on the same idea at once - improve a model by learning from its own attempts instead of expensive human labels - and the field is debating whether it really removes the labeling burden or just hides it.

reinforcement-learning · distillation · training · rlhf · research

AI is learning a 'dark art' that even expert engineers struggle with

2026-06-27

Designing radio-frequency chips is so reliant on hard-won physical intuition that engineers call it a dark art - and now AI is starting to do it, a sign the automation frontier is moving into deep specialist craft.

hardware · chip-design · engineering · automation · industry

AI video has a consistency problem. This model targets it.

2026-06-27

DomainShuttle goes after the tug-of-war in subject-driven text-to-video: keeping a specific character or object recognizable across frames while still letting the scene move freely.

video-generation · text-to-video · creative-tools · research · open-source

← 2026-06-262026-06-27later →