Qwen3.8-27B tops Hugging Face trending models
The open-weight model, released on the 14th, topped the charts in downloads and likes
20
Hyunkook Kim
김현국
One in five US employees now assigns work to AI instead of a colleague
Epoch AI/Ipsos survey finds 57% of software development work uses AI, with many accepting outputs unedited
30
Musk: Orbital Compute, the Only Way to Scale AI
Elon Musk says orbital data centers are the only way to scale AI by 2029, as power shortages and permitting delays stall ground-based buildouts
20
Yutori's Browser Agent Runs Twice as Fast on Together AI
Browser agent that repeats screen captures and clicks dozens of times cuts inference costs 4-5x
10
NVIDIA halves guarantee for OpenAI data center
Under investor pressure, guarantee cut from $25B to under $12B; Anthropic counters bubble talk with 2.4x quarterly revenue jump
30
Qwen confirmed as No. 1 in Hugging Face's local inference rankings
Hugging Face's "State of Open Models" report shows Qwen leading Gemma in real-world local inference use
10
SpaceX Officially Completes Acquisition of Coding Startup Cursor
The deal, announced in April at up to $60 billion, has closed, with Cursor emphasizing access to GPU infrastructure
10
Companies That Replace Entry-Level Hires With AI Keep the Gains While the Industry Splits the Bill
Nolan Lovett of NATO's Special Operations University reframes AI adoption as a "tragedy of the commons." His warning: when individually rational company decisions add up, the shared reservoir of expertise in a profession dries out — and the bill arrives in the 2030s. We dug deep into the paper's five diagnostic criteria and its prescriptions.
190
World Labs Turns One Robot Task Into Thousands of Simulated Runs
Fei-Fei Li's startup unveils an engine that clones a single real-world robot task into thousands of simulation variants for training
370
GitHub's Spec Kit Forces AI Agents to Draft a Blueprint Before Writing Code
As vibe-coded projects keep falling apart, GitHub responds head-on with an open-source spec tool
00
Even the best AI models can't score above 60% on pure vision
Moonshot AI's new benchmark ranks GPT-5.6 Sol first at 59.7%, revealing that many "reasoning errors" are actually perception failures
20
Cursor CEO's "must become multi-product company" remark resurfaces
a16z revives Cursor founder's 2025 strategy comments, sparking renewed attention
00