Jonathan Kemper / The decoder:
Alibaba Technical Report: Qwen3-VL Beats GPT-5 and Gemini 2.5 Pro on Visual Tasks and Has 100% Accuracy on “Needle in a Haystack” Tests for 30-Minute Videos— A few months after the launch of Qwen3-VL, Alibaba released a detailed technical report on the open multimodal model.
Trending
- Luta Security CEO says US government restrictions on Mythos follow jailbreak report by Amazon researchers and calls restrictions ‘completely overreacting’ (Amrith Ramkumar/Wall Street Journal)
- After years of false dawns, Big Tech, startups and governments are betting on commercially useful quantum computers by 2030, as skeptics worry about the hype (Michael Peel/Financial Times)
- Ukrainian extradited from Ireland to US pleads guilty to conspiracy to commit wire fraud for role in Conti ransomware attacks between 2021 and 2022 (Lawrence Abrams/BleepingComputer)
- Moonshot AI releases Kimi K2.7-Code, claiming 30% lower reasoning token usage than K2.6, available under a modified MIT license (Sean Michael Kerner/VentureBeat)
- Paramount-WBD Merger Gets DOJ Approval
- Sources detail frustration within Meta’s applied AI team, formed in March to support superintelligence labs, over menial projects, ‘overwhelming’ work, and more (Wired)
- American households are paying a $3,100 Trump tax and they don’t know it
- Roku is in talks to sell itself; its shares have risen about 24% this year, giving the company a market value of $19.9 billion (Bloomberg)

