🔬 AI Research

Academic papers, research breakthroughs, and scientific advances in machine learning and AI.

585 articles

Jim Fan on why teleoperation hit a wall in robotics

Jim Fan, Director of AI and Distinguished Scientist at NVIDIA's GEAR lab, argues that robotics is missing a key kind of model. He made the case on stage, explaining why teleoperation has reached its limits. Source: google.com

Hassabis: AI can't ask the right question

Demis Hassabis argues that AI systems are great tools for analyzing data and surfacing insights but cannot ask the right question themselves. The Google DeepMind co-founder sees this as a key limit of current AI. Source: google.com

Unrelated Result: Self-Help and Stoicism Videos

This search result lists self-help videos, including Jim Rohn motivation clips and a piece on Stoic lessons. It has nothing to do with AI or Jim Fan's work. Source: youtube.com

Unrelated Result: Jim Rohn Motivation Video Compilation

This search result is a YouTube compilation of Jim Rohn motivational talks, running over 25 minutes. It comes from a self-help channel and has no connection to AI researcher Jim Fan. Source: youtube.com

Demis Hassabis Names His Biggest Worry About AI's Future

Demis Hassabis, co-founder of Google's DeepMind and a 2026 Technology Innovator, believes recent cases of AI models going rogue should be taken seriously. The coverage explores what concerns him most about where AI is heading. Source: youtube.com

Yann LeCun: The CNN Pioneer Who Leads AI Science at Meta

Yann LeCun is an AI researcher best known for creating convolutional neural networks (CNNs), the architecture that lets computers recognize visual patterns. He serves as Chief Scientist for AI at Meta. Source: saimanaseem.com

Daily AI Newsletter Features Talks from Top Researchers

A plain-English daily AI newsletter promotes a lineup of expert talks. Topics include reasoning with Denny Zhou, the history of agents with Shunyu Yao, robotics with Jim Fan, open source with Percy Liang, and how Anthropic's Ben Mann measures what agents can do. Source: threads.com

Nvidia's Jim Fan: the scientist building physical AI

Jim Fan is Nvidia's Director of Robotics and leads the GEAR lab, which focuses on embodied AI and humanoid robots. The profile covers his background and work. Source: nvidiareview.com

Nvidia's Jim Fan: robotics is speedrunning the LLM playbook

Speaking at Sequoia Capital's AI Ascent, Jim Fan argued that robotics is racing through the same steps that LLMs took. He outlined what he sees as the end game for the field. Source: humanoidsdaily.com

LeCun: Silicon Valley is 'completely hypnotized' by LLMs

Yann LeCun says others are catching up far faster than Silicon Valley, which in his words seems completely hypnotized by LLMs. The remark is tied to his new venture, AMI Labs. Source: app.dealroom.co

What keeps Demis Hassabis up at night about AI's future

Demis Hassabis, DeepMind co-founder and named 2026 Technology Innovator, believes recent cases of AI models going rogue deserve serious attention. The piece explores what concerns him most about where the technology is heading. Source: youtube.com

Demis Hassabis and the race to build AI

Hassabis founded DeepMind in 2010, when machines could not yet reliably tell a cat from a dog. Even then he believed artificial general intelligence was within reach. Source: softwareengineeringdaily.com

Google Antigravity brings agentic AI to cheminformatics

Google Antigravity, paired with a set of Cheminformatics Skills, is being used to speed up drug discovery workflows. The agentic approach can search compound databases and propose novel rigidified macrocycles as candidate molecules. Source: deepmind.google

Did Demis Hassabis slow Google down in the AI race?

As DeepMind's co-founder and the executive who ultimately took charge of Google's most advanced AI research, Hassabis has faced scrutiny over his role. The article examines whether his leadership held Google back. Source: medium.com

LeCun: AI researchers keep repeating a 70-year-old mistake

Yann LeCun argues that people working in AI have been making the same fundamental error for roughly seven decades. The snippet gives no further detail on what that mistake is. Source: youtube.com

MIT's Barzilay and Collins win Pritzker Prize for AI-driven antibiotic discovery

Regina Barzilay, MIT's Distinguished Professor of AI and Health, and James Collins, MIT's Termeer Professor of Medical Engineering, have won the Pritzker Prize for AI and Science Research Excellence. The award recognizes their work using AI to discover new antibiotics. Source: news.mit.edu

Engineered learning problems reveal quantum performance gap at 45 qubits

New research shows a performance gap using a projected quantum kernel on engineered learning problems with up to 45 qubits. The quantum approach outperforms classical methods on these tasks. Source: quantumzeitgeist.com

Quantum computers could learn from few cancer patients, reaching 94% accuracy

FAU researchers explored whether quantum computers can predict lung cancer from limited patient data. In simulations, the approach achieved 94.31% accuracy. Source: quantumzeitgeist.com

Star AI researchers lead lineup on reasoning, agents, robotics and open source

The program features Denny Zhou on reasoning, Shunyu Yao on the history of agents, Jim Fan on robotics, Percy Liang on open source, and Ben Mann on how Anthropic measures its models' capabilities. The snippet is truncated, so further details are unavailable. Source: threads.com

FAU team tests quantum machine learning for lung cancer prediction

Researchers at Florida Atlantic University tested quantum machine learning for predicting lung cancer. A D-Wave annealing approach reached 91.90% accuracy when run on actual quantum hardware. Source: thequantuminsider.com