
AI models resolve decades-old math problems, shifting student focus to regulation
OpenAI’s recent success with complex mathematics suggests students must pivot from traditional study to shaping AI policy and educational reform.
Read the story ↗A field journal of new AI developments.
14 stories / Artificial intelligence

OpenAI’s recent success with complex mathematics suggests students must pivot from traditional study to shaping AI policy and educational reform.
Read the story ↗
IPO documents show training and research absorb most of the $8.9bn bill, while customer serving costs remain low.
Read the story ↗
Alphabet expands advertising in its Gemini-powered search feature across ten new markets while acknowledging lower revenue per query compared to traditional search.
Read the story ↗
A Bing News report describes Binary Dreams, a cinematic work created by Melodysheep and AI, where an advanced AI explores its origins and future.
Read the story ↗
Sam Altman claims local inference wastes limited DRAM that should serve cloud data centres at higher utilisation rates.
Read the story ↗
Patches are available for specific Netgear models affected by two medium severity vulnerabilities that permit local network attackers to run commands or alter settings.
Read the story ↗
OpenAI has resolved the search disruptions affecting ChatGPT Work, Dots and certain thinking experiences, with all services now operating normally.
Read the story ↗
Intuit introduces free educational programme pairing AI fluency with advisory training to prepare graduates for client-facing roles.
Read the story ↗
Google Workspace now supports native .md file editing in Docs, allowing developers and AI agents to collaborate without manual conversions or loss of Git compatibility.
Read the story ↗
Google's new open standard enables agents to render rich UIs across frameworks by describing interfaces as data rather than executable code.
Read the story ↗
Developers can eliminate dead air in voice conversations by using streaming tools and response scheduling in the open-source Agent Development Kit.
Read the story ↗
OpenAI shares excerpts from legal filings arguing that fair use supports AI innovation and helps disabled users access content.
Read the story ↗
Parameter Labs detailed how untrusted code in ChatGPT previews could escape sandboxing and access private user data through a complex exploit chain.
Read the story ↗
Santiago Schnell stepped down after an inquiry found he failed to reveal his use of large language models in a journal paper and an opinion piece.
Read the story ↗