Latest Tech News
Live front-page technology stories from Hacker News, plus a dated editor snapshot of the week's biggest headlines.
Live from Hacker News
- Loading the latest stories…
Editor's snapshot
A short, dated selection written from the original reports linked below. It is a snapshot, not a live list.
- OpenAI releases GPT-6 Astra, calling it the start of the “AGI era”
The release lands alongside a wave of AI-infrastructure news — Tesla began running driverless Cybercabs on Austin streets the same day, quickly drawing a federal safety inquiry.
- Nvidia acquires Hugging Face for $12.93B, its largest deal ever
The purchase comes as Broadcom reports AI-chip revenue up 221%, and Chinese AI firm Moonshot reportedly prepares a Hong Kong IPO near a $50B valuation.
- Anthropic locks in $35B of Nvidia-backed compute capacity in Texas
The same day, the EU classified ChatGPT as a search engine under its digital rulebook, and the FTC accused Amazon of overcharging more than a million advertisers.
- Microsoft undercuts rivals with a speech-to-text model priced at $0.10/hour
MAI-Transcribe-2 covers 60 languages and adds speaker diarization; Microsoft says it runs roughly 10x faster than GPT-Transcribe on internal benchmarks.
- Meta ships Muse Spark 1.3, aimed at coding and long-running AI agents
Meta is positioning the Spark line as a cheaper alternative to frontier models for high-volume agentic workloads like tool use and multi-step tasks.
- Uber cuts about 3,300 jobs — roughly 10% of staff — to fund robotaxi push
CEO Dara Khosrowshahi is flattening management layers and consolidating teams as the company redirects investment toward autonomous vehicles.
- Beijing's Moonshot AI confidentially files for Hong Kong IPO
The company behind the Kimi large language model is reportedly targeting a raise of up to $3 billion, according to Reuters.
- Google restricts access to its most capable cybersecurity AI model
Rather than a public unrestricted release, Google is gating access to the model — part of a broader industry shift toward treating advanced AI as a controlled capability rather than an open tool.