Meta CEO Mark Zuckerberg brought Alexandr Wang into the company last year to lead an AI reset, and the focus has now shifted to commercializing Wang's work, with the first major product, Muse Spark, r…
Pearl Research Labs claims to have built a live GPU-secured blockchain without proving the work is useful AI computation, facing a 112 MW usefulness test. A preprint estimates Pearl's network runs at …
UC Berkeley doctoral student Yichuan Wang and collaborators from UC Berkeley, Princeton, EPFL and Databricks published PixelRAG, a research system that bypasses text parsing by rendering web pages as …
Ataraxy Labs released Weave, a Git merge driver that merges code by language structure instead of adjacent lines, aiming to resolve merge conflicts created by AI agents. The tool is available via Home…
Moonshot AI's Kimi K2.7 Code ranked second behind Claude Fable 5 max and ahead of GPT 5.5 xhigh in ErdosBench's mathematical research test, according to a public smoke test by ulam.ai founder Przemek …
Depthfirst's autonomous security agent discovered 21 zero-day vulnerabilities in FFmpeg, demonstrating that AI-driven security tools can find exploitable bugs in heavily scrutinized codebases. The fin…
Dapple, co-founded by Tricia Martinez Saab and Salam Al Mosawi, raised a $30 million seed round from Raptor Group and Ion Pacific to sell private AI cloud as owned infrastructure. The funding came fiv…
AnimateDiff Turbo scored 5.3 in a head-to-head comparison against Luma Ray 3.2 Image to Video, which scored 14.4, demonstrating that Luma Ray 3.2 better executes prompt-specific scene generation from …
Databricks co-founder and CTO Matei Zaharia and Kasey Uhlenhuth open-sourced Omnigent, an Apache 2.0 alpha project described as a "meta harness" for AI agents. The tool sits above existing agent harne…
Rio de Janeiro's municipal IT company, IplanRIO, has released Rio 3.5 Open 397B, a large open-weight AI model post-trained from Alibaba's Qwen 3.5 397B base model. The release was spotted by SemiAnaly…
Meta CEO Mark Zuckerberg's latest AI restructuring has sparked a messy and unpopular all-hands meeting, where an expletive-laced outburst interrupted a livestreamed presentation. The reorganization, w…
OpenRouter's Fusion model outperformed DeepSeek-V4-Pro in a head-to-head coding test, delivering a complete, compilable Go rate limit parser with correct edge cases while DeepSeek V4 Pro made basic er…
Malware authors are embedding nuclear and biological weapons language into malicious code to evade AI-powered security scanners, as flagged by Citizen Lab researcher John Scott Railton. The tactic tri…
Anthropic disabled its Claude Fable 5 and Claude Mythos 5 models after a U.S. export control directive, but CEO Dario Amodei disputed the shutdown's technical basis, saying the government's evidence w…
Z.ai released GLM-5.2 on Saturday to all GLM Coding Plan subscribers, including Lite, Pro, Max, and Team tiers, integrating the flagship model into its coding agent subscription channel rather than li…
Kilo Code published a model test Saturday comparing Claude Fable 5 and ChatGPT 5.5, finding Claude Fable 5 strongest at planning, but the model is already offline due to a U.S. export control directiv…
White House AI adviser David Sacks intervened in a dispute over export controls on Anthropic's frontier AI model, which the company itself had flagged as requiring unusual safeguards. The intervention…
Anthropic CEO Dario Amodei shut off the company's Fable 5 and Mythos 5 models for all customers on June 12 after the US government issued an export control directive barring access by foreign national…
Apollo Atomics, a Y Combinator-backed nuclear startup based in Cambridge, has unveiled a Compact Steam Generator designed to shrink proven pressurized water reactor architecture for factory production…
Happy Horse outperformed AnimateDiff Turbo on prompt fidelity, scoring 17.3 to 4.2 in a test of video generation accuracy. The result highlights the importance of following instructions over surface p…