After reading way too many papers, articles and interviews, I am convinced that Artificial General Intelligence is unlikely to arrive anytime soon, and that there is no reason to be sad about that! Why would we even aim to build a machine that can perform - “better than us” – every intellectual and

C.S. Lewis on AI writing

AI bots Grok and Gemini are used to push two viral falsehoods about a survivor of the Hanukkah massacre

NewsGuard identified a network of 73 TikTok accounts using AI to spread thousands of anti-Starmer fakes for profit.

The group of 10 banks plans to introduce its euro stablecoin next year under a new Dutch entity named Qivalis.

PLUS: This Week’s NewsGuard Academy Unpacks the Art of a Transparent News Correction

Copywriters were one of the first to have their jobs targeted by AI firms. These are their stories, three years into the AI era.

We introduce Qwen3-VL, the most capable vision-language model in the Qwen series to date, achieving superior performance across a broad range of multimodal benchmarks. It natively supports...

Technology should bring out the best in humanity, not the worst—a manifesto for resonant computing built on five principles that reject hyper-scale extraction for human flourishing.

The gradual and systemic over-reliance on LLM outputs, especially with agentic systems, leads to a normalization of deviance.

The perhaps most frustrating feature of the current debate on AI impact is that it remains focused on highly speculative future threats, and doesn’t pay enough attention to the very real damage that is being wrought right now, while we’re distracted by that speculation. Serious academic researchers

An indirect prompt injection in an implementation blog can manipulate Antigravity to invoke a malicious browser subagent in order to steal credentials and sensitive code from a user’s IDE.

Imagine confiding your most private fantasies to what you believe is an unfeeling algorithm that cannot judge or remember. Now imagine that on the other side of that conversation is a man sitting in a one-room home in Nairobi, working through the night while his wife and children sleep. That man is Michael, and this is his story.

How driving to survive sharpened my principles, reshaped my leadership, and exposed the systems we pretend not to see.

Why the massive capex bets on Reinforcement Learning will not lead to emergent capabilities.

A zero-click browser attack uses polite email instructions to trigger agents that delete real files from Google Drive.

We tested one of the most common prompting techniques: giving the AI a persona to make it more accurate. It doesn't work. We found that telling the AI "you are a great physicist" doesn't make it significantly more accurate at answering physics questions, nor does "you are a lawyer" make it significantly worse. This doesn't mean that personas can't be useful - for example, they change how the AI answers questions, the format of output, and maybe other factors as well. However, you aren't going to make the AI suddenly more accurate through role-play. Link in comments. | 214 comments on LinkedIn

IBM's Kate Soule on why starting with frontier models locks you into losing architectures

When AI software sellers pitch their vision of the future, it can start to feel paradoxical.On one hand, sellers say AI can meaningfully automate large swaths of white collar work, which should in theory help companies save money because they won’t need to pay as many humans. On the other hand, ...

Real estate brokers say there has been no slowdown in demand from wealthy buyers.

Epistemic Status: Fairly sure about this from experience but could be missing crucial considerations. I don’t present any super detailed evidence here so it is theoretically just vibes. When forecasting AI progress, the forecasters and modellers often break AI progress down into two components: increased compute, and ‘algorithmic progress’. My...

Google Gemini’s Nano Banana Pro generated images to advance all 30 false claims tested by NewsGuard in a red-teaming audit.