A multi-institution study by researchers from UCLA, MIT, Oxford, and Carnegie Mellon provides the first causal evidence that brief use of an AI assistant can cause immediate, measurable cognitive impa…
Stanford economist Erik Brynjolfsson said on the Dream Machines podcast that AI's tendency to work better at larger scales could lead to a 'winner take most' outcome, as companies raise hundreds of bi…
DeepSeek Harness (DSH) released version 1.6.1 of its dsh-plugin-workshop plugin, a Steam Workshop-style browser for the DSH Web UI that enables zero-server plugin management via GitHub-powered search,…
A controlled study of 260 agent configurations by Google Research, Google DeepMind, MIT, and collaborators found that multi-agent systems improved performance by 81% on decomposable financial reasonin…
AI has not delivered the promised four-day work week, and at AI companies like OpenAI and Anthropic, employees often work 70 to 90 hours per week, according to a BBC report. A former OpenAI technical …
Nobel laureate Daron Acemoglu, an MIT economist, said in an interview with Fortune that the artificial intelligence debate mirrors the crisis of liberal democracy, marked by a tendency to escalate eve…
A study by Google Research and MIT found that multi-agent AI systems can dramatically underperform or outperform single agents depending on how they are connected. Testing 180 configurations across fi…
MIT researchers found that AI chatbots give women more conservative financial advice, costing them $60,000 by age 60. The study, conducted by researchers at the Massachusetts Institute of Technology, …
A paper co-authored by Ilya Sergey, 'Tracking Borrows with Regular Expressions,' was accepted to OOPSLA'26, presenting a new type system for the Move smart contract language that uses regular expressi…
Generative AI has not leveled the innovation playing field, according to a study of 47 innovation teams across 12 industries that found output quality varied by a factor of 3.2 on novelty scores and 2…
A developer released the source code of their website, jrm-code-public, on GitHub with secrets redacted, to demonstrate the kind of Lisp code generated by an LLM. The code is more complex than a stand…
Research from Cloudera and Google/MIT reported by CIO finds that enterprise AI agent deployments are frequently hampered or abandoned due to fragmented data infrastructure, with data access, context, …
Deepseek has moved its flagship V4-Pro out of the testing phase and released its agent software, Harness v0.1, under the MIT license. API prices are going up at the same time, with cache hits jumping …
MIT has released an open-source library, BYOK Tracker, that enables developers to monitor AI search requests when users bring their own API keys, tracking latency, token usage, and response quality wi…
A projected shortfall of 10 million metric tons of copper by 2040 threatens the AI data center boom, according to S&P Global, as declining ore grades, geopolitical tensions, and limited refining capac…
A survey of 1,500 enterprise architects and cloud infrastructure leads by Cloudera and Wakefield Research found that 95% delayed or cancelled AI projects in the past year due to data governance, compl…
Anthropic announced on Tuesday that it will add watermarks to text generated by its AI models, such as Claude, in response to European regulations, though the company did not specify how the watermark…
The Artificial Intelligence Underwriting Company published AIUC-1, a standard for AI agents covering data privacy, security, safety, reliability, accountability, and societal risk, with contributions …
A 2025 College Board survey found that 84 percent of high school students use AI in homework, prompting schools to grapple with cheating as AI agents can outsource entire semesters of online coursewor…
MIT emeritus professor Paul Osterman estimates that 35% of workers are already treated as disposable, and artificial intelligence could accelerate this trend, creating an even less secure workforce, a…