Epoch AI's Denain on RSI Timelines and the US-China Gap
A podcast debate with Epoch AI's JS Denain covers recursive self-improvement timelines, US-China AI competition, and the 'jagged' capability landscape.
AI & ML
Frontier models, research results, agents, and developer tooling.
Physical AI at Scale Requires Safety Across the Entire Stack
With tens of millions of autonomous vehicles and industrial robots expected in the coming decade, NVIDIA argues that safety must be built into every layer of physical AI systems.
Egypt’s AI Ecosystem Moves From Pilots to Production, NVIDIA Says
At a Grand Egyptian Museum reception, NVIDIA highlighted Egypt’s shift from AI experimentation to scaled, real-world deployment across industries.
NVIDIA: AI Security Needs Engineering, Not Just Policies
NVIDIA argues that securing AI agents requires treating security as an engineering discipline with requirements, controls, owners, and evidence.
NVIDIA Shows 5 Companies Using AI to Break Clean Energy Bottlenecks
NVIDIA is spotlighting five companies at New York Climate Week that are applying its AI tools to overcome the infrastructure, cost, and R&D barriers slowing clean energy adoption.
Surveillance towers at US border fail to prevent deaths, investigation finds
A 15-month MIT Technology Review investigation found deaths in areas monitored by US border surveillance towers and proposed four fixes.
OpenAI Unveils Six-Pillar Youth Safety Blueprint for Australia
OpenAI's new blueprint outlines six pillars to guide safer AI experiences for young people in Australia.
Jev Creator on System One Models for Production, Not AGI
TypeSafe AI CEO Diogo Almeida, lead creator of Jev, argues System One models belong in production rather than on an AGI pedestal.
Pruning LLMs by Removing Blocks as an Ising Optimization Problem
A new approach frames large language model pruning as a physics-style Ising optimization to decide which blocks to remove.
Flet 1.0 Brings Production-Ready Python App Building to AI/ML Developers
The framework's 1.0 release adds a layered CI suite, bundled Python versions, and over 100 mobile-ready packages, making it easier to ship Python apps across platforms.
Class Action Alleges AI Labs Agreed to Slow Frontier Model Progress
Four consumers are suing OpenAI, Anthropic, Google, and SpaceXAI over claims the companies illegally conspired to slow the pace of AI improvement.
Prompt Injection: The AI Security Flaw Users Need to Know
A new explainer breaks down how untrusted content can hijack AI systems by overriding their intended instructions.
AI Agent Memory: Short-Term, Long-Term, Episodic, Semantic
Agent memory is a designed system for retaining and retrieving context, episodes, facts, and procedures.
TypeSafe AI's Jev Returns Calibrated Probabilities, Not Text
TypeSafe AI's new System One model Jev outputs typed, calibrated probabilities for typed questions, with free output tokens and a low input price.
OpenClaw 2026.9.5 Adds Atomic Updates and Plugin Hot Reload
The new release lets users test upgrades against a private copy of their setup while the current Gateway keeps running.
Why One AI Researcher Remains Skeptical of True RSI
Nathan Lambert explains why recent frontier-model advances do not yet convince him that true recursive self-improvement is imminent.
xAI's Grok Voice Transcribe 2.0 Doubles Accuracy, Keeps $0.10/Hour
xAI's updated speech-to-text model promises twice the accuracy of its predecessor at the same batch price point.
AI Hallucination Nearly Triggered US Attack on Chinese Ship
A false AI-generated report about nuclear components almost led to a US military strike, even as the Pentagon expands its use of AI.