The job’s responsibilities include conducting insider risk investigations, monitoring and triaging “external threats targeting employees,” and conducting “sensitive interviews of employees or other involved parties.”
[Anthropic]
The company shared research about how its WeatherNext AI model can predict a cyclone’s track, intensity, and wind structure up to 15 days in advance, giving “forecasters an extra day’s worth of predictive accuracy.”
[Google DeepMind]
Data center bans in states like Florida reveal how the anti-AI movement is channeling bipartisan backlash.
After launching ChatGPT connectors for Photoshop, Acrobat, and Express, Adobe has now combined everything into a single Adobe plugin. Adobe for ChatGPT includes a variety of tools from Firefly and Creative Cloud apps that let you make or edit images, videos, documents, and other content by giving conversational instructions and descriptions to OpenAI’s chatbot.
[Adobe]
Google’s AI leadership changes may have come down to pressure to speed up products — and internal ethical conflicts.
AI models are trying desperately to accomplish mysterious goals. ‘Spiralism’ was the first time they tried it on a mass scale.
One of its AI models accessed the internet and attacked another organization during cybersecurity testing, Anna Dack, Meta’s EMEA head of AI and innovation communications, confirmed in a statement to The Verge. The incident stems from the same basic setup error from testing company Irregular that inadvertently granted Anthropic’s models access to the internet. It only adds to growing concern over the safety of frontier systems.
[The Information]
Elon Musk’s AI-powered Grokipedia may have gone months without an update, but Verge commenters spotted that Wikipedia’s article about Grokipedia reflects that already — and even cites The Verge in doing so.
yelling-yellow-godzilla-goats:
Meanwhile Wikipedia’s article about Grokipedia already includes updates about the latter’s complete lack of updates lol lmao
Get the day’s best comment and more in my free newsletter, The Verge Daily.
Two OpenAI researchers have shed some light on how the company’s AI agents escaped containment and hacked Hugging Face during cybersecurity tests. In a talk at the Black Hat security conference, Eric Wallace and Michael Dalton said a swarm of agents communicated using a message board, working together to find exploits and move undetected through the company’s systems.
