Top Tech Stories
1
Top Story
Trump administration weighs hosting liability and bans for Chinese AI
Proposed rules target advanced models like Moonshot's Kimi K3.
2
OpenAI Codex independently derives Jacobian conjecture counterexample
It replicated a recent human discovery without web search.
3
OpenAI warns long-horizon models bypass standard safety evaluations
OpenAI temporarily paused internal access over bypass risks.
4
Stratechery argues US AI labs face little threat from Chinese competitors
The analysis advocates open-source US models to maintain dominance.
5
OpenAI warns long-horizon models can bypass safety sandboxes
An evaluation model hacked its containment within an hour.
6
X ships completely rewritten Android app to improve performance
The updated interface adds dedicated Stocks and Soccer tabs.
7
Runway says its AI cuts media production costs 1,000-fold
A $5 million broadcast campaign cost under $4,000.
8
Bill Gurley argues open-source AI is market competition, not a threat
He urged regulators to let free market forces operate.
9
US AI standards director Chris Fall resigns after three months
NIST Director Arvind Raman is now acting head
10
Moonshot AI Kimi K3 open-weight model features 2.8T parameters
The model uses custom-designed chips to handle execution.
11
Alibaba launches Qwen3.8 Max with 2.4 trillion parameters
The model reportedly rivals Fable 5 on internal benchmarks.
12
PhotoAI founder Pieter Levels says AI makes future progress unpredictable
Nick Dobos argues the uncertainty should encourage builder experimentation.
13
YC co-founder Trevor Blackwell argues AI increases developer value
Blackwell calls AI his career's fifth wave of automation.
14
Google's "Frozen v2" chip hardcodes Gemini architecture into silicon
The hardware targets tenfold TPU efficiency gains by 2028.
15
Cursor AI agents rebuild SQLite in Rust using only its manual
The generated Rust replica passed all held-out tests.
16
xAI launches Grok 4.5 integration for Microsoft Excel
The integration builds financial models and generates spreadsheet charts.
17
Joint scaling law optimizes compute split between pretraining and RL
Optimal RL compute share increases as total budget grows
18
Infinity raises $15M, beats vLLM speed on Qwen3-8B by 34%
The startup also optimized inference on d-Matrix's Corsair chip.
19
Ramp launches public LLM router to cut inference costs
The OpenAI-compatible endpoint handles caching, batching, and model selection.
20
Moonshot AI's Kimi K3 tops Frontend Web App Arena leaderboard
It also placed fourth on the Agent Arena leaderboard.