doorss

An article examining how advertising technology is being embedded directly into web browsers, forming what the author describes as a cartel-like structure. The piece discusses the implications for user privacy and the open web.

Helsinki edistää fossiilisten polttoaineiden mainosten kieltämistä kaupungin kilpailuttamilta mainospaikoilta. Kielto ei koske yksityisiä mainostajia, mutta voi vähentää kaupungin mainostuloja.

Gen Z uses AI more than any other generation but many fear it is weakening their job prospects and career development. The article examines growing negativity among young people toward AI technology.

An FT article exploring the hypothesis that remote working, rather than AI, may be the primary cause of weak junior hiring in tech and other industries.

Discusses ARC-AGI-3, a new AI benchmark that drops models into a computer game with no instructions to test their ability to figure out mechanics on the fly. The author reflects on what this means for how we think about automating work.

A Stanford Law study found that AI outperformed law professors on certain legal tasks, raising questions about AI's growing role in professional knowledge work.

Researchers discovered that websites can fingerprint visitors by analyzing SSD activity patterns through the browser, creating a new privacy threat.

Mullvad VPN argues that mandatory age verification for social media threatens to end a free and open internet, as it would require identity verification infrastructure that undermines privacy.

A security researcher details a VSCode vulnerability that allowed stealing GitHub tokens with a single click, highlighting risks in developer tooling security.

Anthropic presents results from a survey of 1,260 social scientists about AI and coding agent use. 81% of respondents have tried AI chatbots in research, particularly for coding tasks.

Anthropic shares an initial update on Project Glasswing, part of their ongoing research program into AI capabilities and safety.

Simon Willison highlights Anthropic's detailed documentation on how they sandbox Claude across their products, noting it as a rare example of thorough sandboxing documentation.

A catalog of telltale signs that text was written by an LLM, covering stylistic patterns and linguistic tics that reveal machine-generated content.

An essay examining how the flood of AI-generated low-quality content is reshaping expectations around quality in software and writing.

The FT examines how migration policy and labour market reforms are affecting joblessness in Spain and Finland in very different ways, causing their unemployment rates to converge.

Michelin-ravintolaopas julkaisi uusimman pohjoismaisen oppaansa. Suomessa on nyt yli 30 Michelin-ravintolaa ja toinen kahden tähden ravintola.

Andrej Karpathy announced he has joined Anthropic, the AI safety company behind Claude.

Bluesky experienced a sustained DDoS attack lasting nearly a full day, disrupting access to the decentralized social media platform.