Articles & Guides

A new Anthropic paper describes an Automated Alignment Researcher that searches literature, proposes training methods, and iterates on alignment experiments. It improved results across 10 misalignment benchmarks without hurting overall performance.

Reported deals involving Hugging Face, Poolside, and OpenRouter suggest that open-weight AI is becoming an infrastructure play rather than a simple model-sharing movement. Chipmakers and software platforms are competing for control of the developer and inference layers.

A federal judge has ruled that the Trump administration unlawfully retaliated against Anthropic after the AI company refused to loosen limits on lethal autonomous warfare and mass surveillance. The court vacated the government’s directives and its supply-chain risk designation.

The US Environmental Protection Agency is proposing to remove a federal public notice and comment requirement for some “minor” pollution sources. Environmental advocates warn that data centers and related power facilities could move forward with less community oversight.

Meta is updating its AI glasses so recording stops when users cover the capture LED during a recording. The change closes an obvious bypass, but it does not resolve broader concerns about consent, data handling, and privacy claims.

LinkedIn built a multi-agent review platform instead of simply placing a general-purpose AI reviewer in front of its repositories. The system combines independent reviewers, repository-specific rules, filtering, and production-grade operations to improve signal quality and measure whether developers actually adopt its findings.

Cloudflare has unveiled Kitesurf, an experimental browser engine designed for agent-driven tasks such as HTML extraction, screenshots, and PDF rendering. It is not a Chromium replacement yet, but points toward browser infrastructure optimized for software agents rather than human users.

A new paper argues that scaling world models requires more than larger collections of web video. Game engines could provide executable environments, grounded rewards, and long-horizon trajectories for post-training spatial models.

GameWAM combines a visual world model with an action policy for closed-loop gameplay and GUI control. It predicts future observations, generates executable keyboard-mouse trajectories, and studies a new source of bias in generative actions.

A California federal judge ruled that the Trump administration’s decision to label Anthropic a supply-chain risk was unlawful retaliation. The company’s separate lawsuit against the Pentagon in Washington, D.C., is still pending.

Sandhya Devanathan is leaving Meta after more than a decade to join OpenAI, where she will oversee growth, enterprise adoption, partnerships and regulatory engagement across Southeast Asia and Australia.

Tencent Hunyuan has introduced Hy4 preview, an open model with 770B total parameters, 4.9B activated parameters, and a 1M-token context window. The release also highlights integrations with game engines and MCP-oriented tool workflows.

Unisound’s reported first-half results show that enterprise AI services remain its main revenue engine, while token-based model access is emerging as a second growth channel. The more important question is whether delivery-heavy projects can become repeatable and profitable.

Anthropic has introduced a research preview of its Model Hardware Standard, or MHS, a proposed common interface for AI agents to discover and operate robots, microscopes, cameras, and laboratory instruments. Rather than giving Claude one permanent body, the approach lets it use many connected devices as temporary extensions.

Amap has released ABot-Recon, a streaming 3D reconstruction model that uses only the latest 12 frames as local context while reconstructing scenes spanning tens of thousands of frames. Its design targets drift, slowing inference, and growing memory usage in long video sequences.

A federal judge has ruled that the Pentagon’s decision to designate Anthropic a supply chain risk was unconstitutional retaliation. The case grew out of the company’s refusal to remove two restrictions on military use of its AI.

VGI-Bench moves video-generation evaluation beyond visual quality to the logic of evolving visual processes. Its results suggest that current systems show pockets of reasoning ability but are still far from reliable.

WarpSAC argues that off-policy reinforcement learning should not rely on one fixed set of stabilizers across training setups. It introduces two SAC variants tailored to data-limited CPU training and data-abundant GPU simulation.

JIT-Agent treats an agent’s memory, planning, action protocol, and tool orchestration as a generative runtime harness. It aims to customize, repair, and evolve that harness for each task instead of relying on manually engineered workflows.

Video-IFBench shifts video evaluation beyond answer accuracy to the ability to satisfy visual, audio, semantic, formatting, and conditional constraints. Its evaluation of more than 20 recent MLLMs shows that complex video instruction following remains difficult.

AnTrap evaluates Android GUI agents by injecting realistic runtime anomalies into otherwise solvable task trajectories. Its results show broad performance degradation and a clear gap between traps that training can fix and failures rooted in deeper reasoning limits.