mit

Jesse Thaler has been named director of the MIT Laboratory for Nuclear Science, effective Aug. 1. Thaler is a theoretical physicist who combines quantum field theory and machine learning, and has been leading the NSF AI Institute for Artificial Intelligence and Fundamental Interactions.

MIT News reports on Rachel Sava's winning essay for the Envisioning the Future of Computing Prize, highlighting the dual-use risks of neurotechnology, where neural implants could evolve from medical aids into surveillance tools.

A study found that participants writing essays with ChatGPT showed weaker brain connectivity and poorer recall compared to those using a search engine or nothing, suggesting reliance on AI may reduce cognitive engagement.

A study by MIT, Stanford, NYU, and Princeton found that while people expect AI to save time on simple tasks, it actually reduces perceived effort rather than completion time, due to a 'speedup illusion'. The study was limited to short tasks under 5 minutes, so findings may not apply to complex, expert workflows.

A free textbook 'Mathematics for Computer Science' is available under CC BY-SA 3.0, covering key topics for computer science students.

This paper from MIT proposes an adversarial generator-discriminator framework that combines verifiable rewards with a learned signal from human demonstrations to address issues like diversity collapse, unnatural responses, and reward hacking in RLVR training of language models.

This article debates whether AI should be banned for children under 15, referencing a French law banning social media for that age group and citing MIT research showing cognitive debt from AI use.

Introduces the classic MIT open course "How to Speak," where Professor Patrick Winston emphasizes the importance of communication skills and provides practical rules such as how to open a talk, control information density, and use analogies. This course remains a secret weapon for many elites, and some have even turned its rules into AI prompts.

MIT President Sally Kornbluth and ASU President Michael Crow discussed the critical role of federal funding for curiosity-driven research and the integration of AI with ethical, foundational education at a Washington Post summit.

MIT News interviews Professor Phillip Isola about agentic AI, covering its definition, differences from generative AI, challenges like lack of training data, and promising applications such as coding agents.

MIT held its inaugural Music Technology Research Showcase, featuring projects from the first cohort of its new graduate program, including AI co-improvising agents, sound-art installations, dance-generated music, and EEG-based tune identification.

An interview with MIT alumnus Alexandros Haridis about his exhibition 'Beyond Data-Driven Aesthetics,' which explores the historical and philosophical intersections of computation, AI, and aesthetic judgment through physical installations and visualizations.

MIT physicist Sanjoy Mahajan's textbook 'The Art of Insight in Science and Engineering' is available for free on MIT OpenCourseWare, teaching nine mental tools for tackling complex problems effectively.

Gilbert Strang presents an updated vision for teaching linear algebra through a YouTube playlist and MIT OpenCourseWare site.

The MIT textbook 'Mathematics for Computer Science' has been translated into Japanese and is available at inzkyk.xyz.

MIT CSAIL researchers developed Masked Inverse Reinforcement Learning (IRL), which uses large language models to clarify ambiguous instructions for robots and focus on key environmental details, reducing the need for extensive demonstration data.

A Twitter thread shares a curated list of 10 free, legally downloadable textbooks from MIT, Stanford, Berkeley, and Harvard, covering topics like linear algebra, machine learning, probability, and data science.

The article reports on a Scientific American special section featuring MIT faculty who advocate for curiosity-driven science as a key driver of America's success and call for sustained public investment in research.

MIT team released a paper on self-evolving skills for Claude Code agents, achieving 71.1% pass rate, surpassing Anthropic's skill-creator by 37 points through a Generate-Test-Verify-Co-Evolve framework.

Researchers from MIT and Microsoft developed an intelligent system that automatically optimizes agentic workflows, reducing computational resources and energy usage while maintaining performance.