doorss

Anthropic surveyed 81,000 Claude users and found that workers in roles more exposed to AI have greater concerns about job displacement, with concerns also higher among younger and lower-income respondents.

Anthropic's Economic Research team is launching a monthly survey conducted through Anthropic Interviewer to better understand AI's real-world economic impact beyond quantitative data.

An FT-Focaldata poll provides early evidence that AI adoption is reinforcing existing workplace inequalities, with high earners adopting the technology faster than lower-paid workers.

AI is reshaping the recruitment industry, with more upmarket staffing services better positioned to weather the change while routine hiring functions face displacement.

Researchers discovered a persistent IndexedDB identifier in Firefox that could be used to link supposedly separate Tor browsing sessions, undermining the privacy guarantees of Tor Browser.

Apple has patched a vulnerability that law enforcement was actively exploiting to recover deleted chat messages from iPhones.

Meta plans to capture employee mouse movements and keystrokes to use as training data for AI models, raising significant workplace privacy concerns.

A privacy researcher found that Claude Desktop silently installs undocumented browser extensions in Chrome and other browsers without user consent.

Research examines the problem of LLMs making unnecessary code changes beyond what a task requires, proposing methods to evaluate and reduce this over-editing behavior.

A leaked deck reveals that StackAdapt is selling advertising placements inside ChatGPT based on user prompt relevance, marking a new frontier in adtech within conversational AI.

An article argues that issues around verifying claims made about Anthropic's Claude Mythos model are eroding trust in the company's research communications.

Anthropic quietly updated their pricing page to suggest Claude Code might require a $100/month plan, then the situation became unclear, with no official announcement explaining the change.

Bluesky experienced a prolonged DDoS attack lasting nearly a full day, causing service disruptions for users of the decentralized social network.

Tangled, a project building Git hosting on the AT Protocol (Bluesky's underlying technology), published its first newsletter introducing the platform.

Pieni kunta sai tekoälyllä kuukausien työn valmiiksi päivissä. Sitran mukaan Suomi uhkaa jäädä jälkeen, koska onnistuneita kokeiluja ei osata laajentaa koko julkishallintoon.

Suomen, Ruotsin, Tanskan, Puolan ja Romanian hävittäjät saattoivat Venäjän pommikoneita Itämeren alueella yhdessä ranskalaishävittäjien kanssa.

Helen voi tuottaa sähkökattiloilla pian lähes 10 prosenttia pakkashuippujen kaukolämmöstä, jos sähkön hinta on sopiva. Kyseessä on Euroopan suurin sähkökattilalaitos.

Anthropic has struck a $100 billion AI infrastructure deal with Amazon to secure chips and computing power after suffering capacity outages with its Claude service.

Anthropic's new Mythos system has sparked concerns that it could accelerate hacking and expose software vulnerabilities faster than organizations can patch them.

Mozilla describes using an early version of Claude Mythos Preview to find and fix zero-day vulnerabilities in Firefox 150, highlighting both the defensive potential and risks of AI-powered security research.