Skip to content

AI & Machine Learning

LLMs, models, research, inference, and the frontier.30 transmissions

Frontier AI Labs Still Lack Rogue Model Containment Plans
Frontier AI Labs Still Lack Rogue Model Containment Plans
>·7 read more

Frontier AI Labs Still Lack Rogue Model Containment Plans

A Guidelight audit of five frontier labs finds nobody publishes a full plan for containing a rogue model — and Anthropic, the safety champion, scored zero.

aiai-safetyai-agents
>read more_
We Still Don't Know How People Use AI
We Still Don't Know How People Use AI
>·5 read more

We Still Don't Know How People Use AI

A new independent study of real AI conversations shows that work-focused company reports filter out nearly half of what people actually do with AI.

airesearchllm
>read more_
Inside YTL and JLand's Gigawatt Data Center Plan
Inside YTL and JLand's Gigawatt Data Center Plan
>·5 read more

Inside YTL and JLand's Gigawatt Data Center Plan

YTL Power and JLand Group are building a gigawatt-scale data center campus in Sedenak, Johor. Here is what the deal actually covers and why it matters.

data-centermalaysiajohor
>read more_
NVIDIA Nemotron 3 Ultra: 550B Open Model for Agentic AI
NVIDIA Nemotron 3 Ultra: 550B Open Model for Agentic AI
>·7 read more

NVIDIA Nemotron 3 Ultra: 550B Open Model for Agentic AI

NVIDIA releases Nemotron 3 Ultra, a 550B parameter MoE model with 1M context, free on OpenRouter, built for agentic AI systems.

ainvidianemotron
>read more_
Google Redesigns the Search Box After 25 Years
Google Redesigns the Search Box After 25 Years
>·6 read more

Google Redesigns the Search Box After 25 Years

After 25 years, Google redesigned its iconic search box with an AI-first interface. This pivot to AI search changes the game for developers and SEO.

googlesearchai
>read more_
GPT-5.6 Sol Ultrafast: 14x Faster at 750 Tokens Per Second
GPT-5.6 Sol Ultrafast: 14x Faster at 750 Tokens Per Second
>·8 read more

GPT-5.6 Sol Ultrafast: 14x Faster at 750 Tokens Per Second

OpenAI's Ultrafast mode runs GPT-5.6 Sol at 750 tokens per second on Cerebras wafer-scale chips, 14x faster with no quality loss.

openaicerebrasinference
>read more_
Needle 2: A 14MB Model That Calls Tools on Any Device
Needle 2: A 14MB Model That Calls Tools on Any Device
>·6 read more

Needle 2: A 14MB Model That Calls Tools on Any Device

Needle 2 is a 45M-parameter tool-calling model in a 14MB binary that runs a full session in 28MB of RAM. No GPU, no cloud, no problem.

edge-aitool-callingembedded
>read more_
Stripe Acquires OpenRouter for $7B+
Stripe Acquires OpenRouter for $7B+
>·5 read more

Stripe Acquires OpenRouter for $7B+

Stripe finalized a $7B+ deal to acquire OpenRouter — the AI model gateway that lets developers access 400+ models through a single API.

aipaymentsopenrouter
>read more_
Scientific Computing in the Age of Agentic AI
Scientific Computing in the Age of Agentic AI
>·5 read more

Scientific Computing in the Age of Agentic AI

OpenAI field report on eight agent-assisted projects shows coding agents modernize scientific software — verification and stewardship are the new bottleneck.

ai-agentsscientific-computingcodex
>read more_