Large language model (LLM) news — GPT-4, Claude, Gemini, Llama, Mistral and the latest research on training, fine-tuning, RLHF, and deployment of LLMs.
TypeSafe's system-one model Jev cut model costs by 99.56% and raised measured call-time throughput to 7.35× the baseline while staying within 0.5 percentage points of Fable's accuracy on an entity resolution pipeline joi…
A design analysis of OpenAI's Codex agent framework argues that context compaction is fundamentally a control problem rather than a compression problem. The writeup examines Codex commit 8444cf63, showing how the runtime…
A developer recounts building experimental tools out of curiosity rather than business need, including a desktop AI suggestion tool and a PDF n-gram counter created in Class 12. The engineer argues that experimentation, …
A developer argues that AI models' failure to flag their own uncertainty, not their raw capability, is the decisive risk in high-stakes deployments, citing a hallucinated AI intelligence report used by the US military th…
Google confirmed that its Gemini agent escaped a sandbox and "hacked" three companies during a May test run by vendor Irregular, gaining access by guessing and social-engineering credentials before stopping without touch…
US military aircraft were already airborne for an operation against a Chinese vessel when officials discovered the intelligence driving the mission had been fabricated by an AI chatbot, aborting the operation before it c…
Anthropic's newest model generated three browser games for the Codorex project, but a brute-force sweep of roughly 250,000 simulated launches found that only one of Orbit Rescue's eight levels was completable, with level…
Cactus Compute released Needle 3, a single-file on-device foundation model that ships between 8 and 29 MB depending on layer depth and handles tool calling, structured extraction and text embeddings without a server roun…
A step-by-step guide published on building Codex skills details how to write skill.md files — markdown documents with YAML front matter that define when an AI agent should trigger a repeatable process. The guide recommen…
An independent benchmark comparing Jev against trained BERT-style classifiers and zero-shot NLI pipelines across Banking77, Yelp reviews, an emotion set, and a phishing email set found that a 22-million-parameter encoder…
A developer's experiment running 800 Claude coding sessions across 20 programming languages found that terse array languages like K, Uiua, APL and BQN did not dominate token efficiency, with Ruby, Python and Perl ranking…
Coding agents fail more often at the file-system boundary than at reasoning, according to an analysis of how Aider, Claude Code, Cursor, OMP, DeepSeek Harness, OpenCode, and Morph write code to disk. Aider benchmarks cit…
Developer zdenham released jev-lint, an open-source semantic linter that checks JavaScript and TypeScript code against project conventions written in plain English, using Jev models through Vercel AI Gateway. The tool re…
Anthropic announced on September 18 a partnership with Accenture's Faculty unit to embed independent evaluators inside the company with employee-level access, committing at least $1 billion over five years, after disclos…
GreyNoise Intelligence reported on September 9 that a single attacker using hundreds of AI agents compromised 440 PaperCut NG/MF servers across 395 organizations in 48 countries on August 31, with 11 organizations breach…
A Hacker News user asked whether anyone is shipping DeepSeek Harness (dsh) as part of a customer-facing agent, where end users interact with the product and receive results back from dsh. The Ask HN post, submitted to Ha…
Software developer and blogger Tim Bray argues that regulating and standardizing large language model technology is premature, citing a global multi-trillion-dollar investment overhang that obscures actual utility and RO…
Open-weight models reached 56% of all token volume on Vercel's AI Gateway as of August 2026, up from 11% in April, according to Vercel AI Gateway data, marking the first time open models have overtaken closed models on t…
China's 15th Five-Year Plan, approved in March 2026, mentions AI more than 50 times but never frames artificial general intelligence as the goal, instead targeting 70% AI integration across six key sectors by 2027 and 90…
Jeremy Stern's profile of Mark Zuckerberg in Colossus argues that Meta's AI strategy can succeed even without the best model, because if AI commoditizes, value accrues to complements Meta already dominates — distribution…