901
Ternary Bonsai: Top Intelligence at 1.58 Bits
Today, we’re announcing Ternary Bonsai, a new family of 1.58-bit language models designed to balance strict memory constraints with high accuracy requirements.
Top posts
Every blog post that took off on Hacker News, ranked by the votes of its best submission. 1,521 posts.
902
We might all be AI engineers now
The models are good now. But most people still miss the point.
903
Wolfram Language and Mathematica version 15
AI-related upgrades like the built-in AI Assistant, extended computation-augmented generation and the Wolfram Agent Tools framework. Advances in TimeSeries & Tabular. New categorical data, ModelFit automated model selection, symbolic music. Efficient handling of gigabyte-sized notebooks. Gains in the already-robust…
904
AI for American-produced cement and concrete
Meta is continuing its long-term roadmap to help the construction industry leverage AI to produce high-quality and more sustainable concrete mixes, as well as those exclusively produced in the Unit…
905
Claude Token Counter, now with model comparisons
I upgraded my Claude Token Counter tool to add the ability to run the same count against different models in order to compare them. As far as I can tell …
906
OpenAI Losses Increased Nearly 8X in 2025, with Spending Hitting $34B
Soundtrack: In Flames - Colony To further support my independent journalism, please subscribe to my premium newsletter. It’s $7 a month or $70 a year. If you’re subscribed to the free newsletter and logged in, you should see at the bottom right hand corner of your screen a little
907
Eight more months of agents
I wrote up my experiences programming with LLMs a bit over a year ago , and updated it for the world of agents eight months ago . A lot has changed since then, so here is an update.
908
909
OpenAI DayBreak – GPT-5.5-Cyber
OpenAI introduces new Daybreak tools, including Codex Security and GPT-5.5-Cyber, to help organizations find, validate, and patch vulnerabilities at scale.
910
Cohere Transcribe: Speech Recognition
Unmatched accuracy and speed. Transcribe converts your business’ audio data into precise text for search, analytics, and automation.
911
Running Kimi K3 on MI355X at Better Performance per Dollar Than B300
Running Kimi K3 at ~952 tok/s/node, AMD continues to prove its case as the winner in performance per dollar.
912
Choosing an AI model: one prompt, 11 models, different results
Netlify now runs any OpenRouter model, including Kimi K3, GLM 5.2 and DeepSeek V4. We tested 11 of them on the same build prompt to see how they differ.
913
Kevin Kelly's 2016 AI Predictions, 10 Years Later
Kelly predicted AI heavyweights while advocating for a future of open weights
914
The New AI Superpowers: Focus and Followthrough
Burnout is on the rise again, with an ironic twist.
915
Slightly reducing the sloppiness of AI generated front end
I've been trying to make progress on a herculean task; I want decent looking programs for personal use which I can generate quickly using AI agents yet I'm a person without taste controlling an AI without taste. I found "one weird trick" that does enough for me, the results are not extra nice but it doesn't make me…
916
Real-time LLM Inference on Standard GPUs: 3k tokens/s per request
Today, Kog AI launches a tech preview of the Kog Inference Engine (KIE): 3,000 output tokens/s per request on 8× AMD MI300X GPUs and 2,100 on 8× NVIDIA H200 (FP16, no speculative decoding). This preview runs a 2B model, with support for large third-party MoE models coming next at similar speeds.
917
Consistency diffusion language models: Up to 14x faster, no quality loss
Standard diffusion language models can't use KV caching and need too many refinement steps to be practical. CDLM fixes both with a post-training recipe that enables exact block-wise KV caching and trajectory-consistent step reduction — delivering up to 14.5x latency improvements
918
Gemini Robotics-ER 1.6
Gemini Robotics ER 1.6 upgrades spatial reasoning and multi-view understanding, unlocking new capabilities like instrument reading for autonomous robots.
919
Introducing Toast 1
Meet Toast 1, Mixedbread's search agent for knowledge-intensive tasks, matching or outperforming Claude Opus 5 and GPT-5.6 Sol while being up to 10×…
920
AI, Lean, and the Future of Formal Verification, with Lean Creator Leo de Moura
Leo de Moura on Formal Proofs, AI Agents, and the End of “Trust Me” Code. Subscribe: Apple • Spotify • … More
921
922
500B tokens later: Letting AI agents decompile a first-person shooter
During the last 3 months, I spent some of my time and tokens decompiling a popular first-person shooter. The goal was not to reach a simple proof-of-concept state. Instead, we really wanted an accurate, stable and feature-complete recreation of the game. The avid reader of my blog might have noticed that I had…
923
The classifiers Anthropic puts in front of Fable are too zealous
COMBINE-lab develops algorithms, data structures, and software for high-throughput genomics.
924
AI children's books, body horror edition
Last week, I posted a visual demonstration of the sameness of AI-generated content.
925
Making an AI bid writer refuse to lie
A year of failure postmortems from building document AI for public tenders: phantom partners, silent coverage collapses, broken truth-meters, and why the refusal became the product.
926
The first 40 months of the AI era
Lzon.ca. A personal blog, by a programmer and IT expert.2026216 points144 commentsAI & Machine Learning details
A personal blog, by a programmer and IT expert. Essays, Articles, Guides, and Recipes. As well as Code, Quotes, and Links.
927
Training our own AI models
I really think we're on the verge of some of our best work through the next six months. Over the past year, we've started building more AI-powered…
928
Ornith-1.5: From Self-Scaffolding to Self-Improvement
Introducing Ornith-1.5, an open-source model family that extends self-scaffolding into an end-to-end self-improvement loop.
929
LLM Usage in Debian: Three Proposals
This proposal aims to expressly forbid any contributions to Debian written with the use or assistance of large language models (LLMs) or other generative AI tools.
930
Vibe coding my way to a healthy family: Introducing Gamow Labs
Owen arrives On September 23rd, 2021, my first son Owen was born. Clearly inheriting his mom’s type-A personality, he arrived on his due date at a chunky 8.75 lbs. We were over the moon. Until we weren’t. At a few … Continue reading →