Google TPUs are custom chips built for AI models, and the latest generation processes 121 exaflops.
A setup connecting llama.cpp with VSCode to run Gemma-4-26B-A4B-IT-Q8_0 has been shared, along with its performance results.
Evidence has emerged that Anthropic lowered Claude Code's default prompt cache TTL from 1 hour to 5 minutes.
Nokia's CEO warned that Europe lacks AI data centre infrastructure and faces an investment gap.
It summarizes GPT-5.5's scope of use, safety evaluations, and new guardrails.
Nordcraft 2.0, which combines an AI agent with a visual builder, has been released.
Tencent Hunyuan has open-sourced the 295B MoE Hy3 preview weights.
OpenAI explained how to start your first project and task using the Codex desktop app.
This piece rounds up 10 ways to use Codex for real work tasks, from daily briefs to workflow audits.
Covers Codex's sleep prevention, detail level, tone, and avatar settings.
Introduces Codex's sidebar, projects, settings, chat window, and the Steer feature.
Codex connects to external tools with plugins and follows task procedures with skills.
Codex automatically performs repetitive tasks using schedules and trigger conditions.
Codex is an AI agent that performs real work across files and tools on your behalf.
North Korean hackers automated phishing and malware with AI tools, stealing up to $12 million over three months.
On Forge Global, Anthropic's valuation reaches $1 trillion while OpenAI stands at $880 billion.
OpenRouter's Qwen3.6 Plus input, output, and cache prices have been cut.
In Claude, you can call 1,000+ apps via the IFTTT MCP.
Google is building its first Alpine data center in Kronstorf, Austria.
Google has released the Gemini Deep Research Agent as an API.
With llama.cpp ngram-mod speculative decoding, Qwen-3.6-27B speed rose to 136.75 t/s.
Beyond AI scores and guides, human judgment in interviews was what truly separated real understanding.