Kolibri is an English-German Mixture-of-Experts model with 78B total and 3B active parameters, a context of up to 1M tokens and open weights under Apache 2.0, specialized for sovereign, mission-critical work.
See how Cerebras powers GPT-5.6 Sol in OpenAI's Ultrafast mode at up to 750 output tokens per second for time-sensitive AI and agent workflows.
This study explores the neural and behavioral consequences of LLM-assisted essay writing. Participants were divided into three groups: LLM, Search Engine, and …
Kimi K2.6 advances open-source coding, featuring long-horizon coding, coding-driven design, agent swarms, proactive agents, and the Claw Groups research preview.
Meta fixed the bug that let anyone trick its Meta AI chatbot into resetting the password on Instagram accounts that didn't have two-factor authentication.
I compiled a RISC-V CPU into Qwen 3.5 4B's weights, using its Gated DeltaNet layers as RAM. It runs C and MicroPython inside the forward passes, and Qwen uses it for exact arithmetic.
Claude Code vs OpenCode token overhead measured at the API boundary. Out-of-the-box baselines, instruction file weight, MCP schema tax, subagent multipliers, and cache-write behaviour.
Today we're announcing Bonsai 27B, our multimodal flagship: ternary at 5.9GB for laptops, 1-bit at 3.9GB for an iPhone 17 Pro, with a 262K-token context.
Qwen offers comprehensive functionality spanning chatbot, image and video understanding, image generation, document processing, web search integration, tool utilization, and artifacts.
This post is an elaboration on a comment I made on Hacker News recently, on a blog post that showed an increase in volume and decline in quality among the “Show HN” submissons. I don't actually mind AI-aided development, a tool is a tool and should be used if you find it useful, but I think the vibe coded Show HN…
Qwen offers comprehensive functionality spanning chatbot, image and video understanding, image generation, document processing, web search integration, tool utilization, and artifacts.
...or how to make Qwen learn query optimization via agentic reinforcement learning
GLM 5.2 is the first open weights model I'd call a genuine competitor to Opus and GPT for agentic work - at ~15-20% of the price. Part one of why AI inference margins are about to collapse.
xAI is renting huge amounts of GPU capacity to Anthropic and Google. Financial engineering ahead of the SpaceX IPO, a real compute shortage, or a genuine datacentre advantage? Probably all three.
Everyone is talking about Jev - here it is in 25 lines of Python.
Agents made implementation cheap, so I started doing far more of it in parallel. My GitHub history since 2019 shows what that did: as VP of Technology I am back at my full-stack engineer pace, with bursts, burnout and a low season that is no longer low.
Pi now supports MCP. Why we changed our minds, what changed in MCP and in Pi, and what Codemode is.
An overview of how Multi-Token Prediction (MTP) drafters are making Gemma 4 models up to 3x faster at inference.
When AI handles 95% of your incident response, do you get worse at handling the 5% that actually matters?
I generated 1,000+ SVGs across 7 frontier models to test whether AI labs are training on Simon Willison’s pelican-riding-a-bicycle benchmark.
The delirious rantings of Frederick Vanbrabant. A blog focused on the intersection of Enterprise Architecture, product, and business strategy.
Chinese AI lab DeepSeek’s last model release was V3.2 (and V3.2 Speciale) last December. They just dropped the first of their hotly anticipated V4 series in the shape of two …
Google AI Ultra subscribers in the U.S. can now try out Project Genie.
If you liked this piece, you should subscribe to my premium newsletter. It’s $70 a year, or $7 a month, and in return you get a weekly newsletter that’s usually anywhere from 5,000 to 18,000 words, including vast, detailed analyses of NVIDIA, Anthropic and OpenAI’s
How I tricked Claude into leaking your deepest, darkest secrets
Compression and LLMs are trying to solve the exact same problem: predicting what comes next. Learn the fundamentals of compression and how better prediction leads to better shrinkage.
Muse Wonder is Meta's frontier AI model that keeps track of what matters to you with Active Memory & Active Focus to work towards your goals at work & life.
I trained a small transformer from scratch in 1.5hrs on a 5090 Beats many LLMs, and scores the same as TRM/HRM
I am a dermatologist. I tested the ability of GPT-6 Astra and Claude Fable 5.1 to diagnosis skin cancer and found that they were less accurate than I expected. I compared to a 10-year old CNN and to my surprise found that this was superior.
