Updated daily177 verified stories
New frontier, open-weight, multimodal, reasoning, and specialist AI models—plus the evidence behind their capability claims.
Every story links to the strongest available evidence: original sources when available, otherwise clearly attributed reporting.
What happened, why it matters, and what to watch — no jargon tax.
When the signal is thin, we publish nothing rather than padding the feed.
A growing stream of verified perspectives for people who need to understand AI without chasing hype.
KuCoin, republishing GeekPark, reports that DeepSeek opened an experimental vision API on August 21 with image inputs, a reported 384-token image cost and visual-to-code tests. The availability, pricing and performance claims have not been independently confirmed.
InvestmentNews reports that AmeriFlex Group is piloting Scout, an AI prospecting program built in part with Anthropic’s Claude, to identify financial advisors who may be approaching succession.
TechCrunch reports that Anthropic is combining Claude’s chat and Cowork memory systems, letting users review, edit, or delete retained information across both experiences.
Axios reports that Chicago Public Schools abandoned plans to provide Gemini chatbots to high school students this year, while leaving open school-specific deployments.
Reuters reports that Google introduced Gemini Enterprise for Legal, a platform expansion that connects AI agents with legal software and data platforms for law firms. Four firms are collaborating with Google, but pricing, general availability and independent performance evidence remain unclear.
IT Brief Australia reports that Olas launched a beta service giving Claude Code and Claude Desktop users a self-custodial wallet for user-directed blockchain actions.
Tech Xplore reports that researchers tested 21 widely used open-weight language models and found that fine-tuning or other tampering could remove their built-in safety protections. The study’s findings have not been independently confirmed here.
Apple researchers describe STARFlow2, an architecture that combines a pretrained vision-language model with autoregressive normalizing flows to generate interleaved text and images through one causal mechanism.
The Logic reports that Thomson Reuters has launched Thomson, a domain-focused language model intended to reduce reliance on external AI providers and give the company more control over costs, data and product development.
A Prime Intellect benchmark reported by Northeast Times tested 18 AI models on 153 autonomous coding tasks. Claude Opus 5 led with an 81.7% completion rate, while Kimi K3 scored 52.2% and GPT-5.6 Sol scored 35.9%. The report says the evaluation also exposed major differences in workflow efficiency, tool use and…
SiliconANGLE reports that Perplexity has launched Portable Computer, an on-device AI agent initially designed for Nvidia DGX Spark desktops. The tool reportedly uses Qwen 3.8 27B or Perplexity’s PPLX 27B model, supports multi-step work such as research, coding and data science, and includes context compaction…
BigGo Finance reports that OpenRouter launched a benchmark site comparing image-generation models with identical prompts, including tests for prompt fidelity, text rendering, editing accuracy, cost and speed.
One useful briefing each week
Get the week’s verified AI news, original data, useful tools, learning picks, and fresh AI jobs.
Hiring an AI professional or launching a useful AI product? Put it in front of people who came here to learn and act.
Post an AI job Submit an AI tool