Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
A survey of Thai-language AI resources on Hugging Face highlights underused datasets and models, including the OpenThaiGPT evaluation dataset and the WangchanThaiInstruct multi-turn conversation set, alongside SCB 10X's …
A developer behind the Edge0 project demonstrated a 35B-parameter language model running on an iPhone with only 1–2.5 GB of peak memory by using Mixture of Experts (MoE) routing to load only the needed expert weights fro…
Google's Threat Intelligence Group documented a mass credential-harvesting campaign in which a financially motivated attacker deployed an autonomous multi-agent framework inside compromised cloud infrastructure, compromi…
Fatih Kadir Akın's prompts.chat, launched at the start of the consumer AI boom, turned unstructured LLM system directives into version-controlled, machine-readable artifacts, according to the article. The project grew fr…
GoatCode released version 2.1.10, an open-source terminal AI agent distributed as a single 85 MB binary that connects to more than 180 LLM providers and automatically retries and switches providers mid-turn when a quota …
OpenAI's GPT-6 Astra solved the final remaining problem in the elite FrontierMath Tier 4 benchmark, taking the model from a sub-2% success rate to total saturation in under two years. In response, twenty-five Fields Meda…
OpenAI shipped GPT-5 to Azure and DeepSeek open-sourced a 236B-parameter Mixture-of-Experts model, according to the headline of The Newsroom EP01. The source body contains no further details beyond YouTube platform boile…
OpenAI's claim to have solved the Navier-Stokes Millennium Prize Problem in 88 hours using roughly 10,000 AI agents and 130 billion output tokens is being challenged on the grounds that the work may have leaned on prior …
OpenAI published a case study on Cognition's Devin using GPT-6 Astra to autonomously test its own code, with one AI-generated summary claiming the model reduces engineers' manual review burden by 20–30%. The post present…
OpenAI scaled its online storage system to serve over 1 billion ChatGPT users and handle 22 million requests per second, according to an OpenAI engineering post. The company evolved an internal Python library into a glob…
OpenAI has paused new sign-ups for its $200/month Pro plan because demand for its new Astra model is straining the company's infrastructure, TechCrunch reported on September 10, 2026. The pause affects only new customers…
A developer writing as Nokka has published a guide detailing five Ollama settings that should be tuned before running local models in sustained heavy use, arguing that slow responses usually stem from untuned defaults ra…
An independent developer has open-sourced Beatriz Epistemic Gate, a lightweight defensive proxy that detects data poisoning during LLM fine-tuning by verifying generated text against an immutable anchor corpus via embedd…
Three Thai open-weight LLM projects — SCB 10X's Typhoon, the Thai AI Entrepreneur Association's OpenThai, and VISTEC's Pathumma — offer locally runnable, commercially usable alternatives to foreign models, each different…
Writer Robert Vesco published a transcript of an early-2025 ChatGPT session in which the LLM generated novel biological attack scenarios against an AI-protected nation state, saying the model's ideas were "even better an…
A blog post on autonomous agent reliability argues that fully autonomous agentic workflows are "extremely improbable" to be reliable, citing arXiv paper 2510.27630 showing reliability tends to stabilise when humans are i…
An open-source VS Code extension called Antigravity Usage Intelligence gives developers visibility into token consumption, cache savings, and subscription quotas when using Google's Antigravity agentic coding IDE. The to…
A Hacker News user posting under the handle steammaho asked whether anyone has tried to build a game engine purely for LLMs, proposing an engine that would make models efficient, fast, and cheap through how data and work…
A developer built Antigravity Usage Intelligence, an open-source usage tracker and analytics dashboard for Google's Antigravity agentic development environment. The tool breaks down token consumption into fresh input, ca…
A developer built a 5-layer, 512-wide TinyStories language model that runs natively on Apple's M4 CPU by treating ARM Neon instructions—including XOR, popcount, and SDOT—as trainable architectural primitives rather than …