Videos

Keeping the full conversation history in your AI agent's context is often cheaper and more effective than compacting...

Qwen's 27B model ships with a multi-token prediction head that can nearly triple throughput for free—no extra...

Alibaba's Qwen 3.8-27B replaces 75% of its attention layers with Gated DeltaNet, slashing KV cache memory from 244...

Alibaba's Qwen 3.8 27B dense model claims to beat Opus 4.6 on several coding benchmarks while being Apache 2.0...

Qwen 3.8 27B is a huge leap over 3.6 for local AI agentic coding. In a zero-shot test, it built three playable retro...

The creator argues that Claude's native project feature is best used only for grouping chat sessions, not for...

Four new AI models hit the market in three days — Grok 4.6, DeepSeek V4 Pro, Nemotron 3.5 Lightning, and Muse...

A GitHub repo with two commits and no runnable code is outrunning every other tool on the board—because it's a...

16 GB VRAM just became the most common GPU config on Steam, which means the local AI tier list has shifted. Qwen 3.5...

Nvidia's Nemotron 3.5 Lightning is 88% not a transformer — it's a hybrid of Mamba state-space layers and...

SpaceX bought Cursor and merged with XAI to launch Grok Bot, a $200/month AI agent platform that lets non-technical...

Nvidia dropped Nemotron 3.5 Lightning, a 30-billion-parameter mixture-of-experts model with only 3 billion active,...

GrokBot is a new messaging-platform-style AI agent system from SpaceX (the company that acquired Cursor for $60M)...

Qwen 3.8 (27B model) drops in 20 hours and promises to redefine local AI, but the creator's AC failing at the worst...

Meta's Muse Glimmer is a 30B parameter coding agent that runs in 14GB of RAM thanks to dynamic quantization, which...

Meta released Muse Glimmer, a 30B-parameter open-source model (Apache 2.0) designed to fit on a 24GB GPU, built via...

Five free Claude skills fix the biggest pain points: watching videos, doing real market research from comments,...

NVIDIA's Switchboard is an open-source routing library that sits between your agent and its models, deciding...

Meta's new Muse Glimmer 30B local LLM, released under Apache 2.0, delivers surprisingly strong visual reasoning and...

Meta is back in the open-weights game with Muse Glimmer 30B, a dense model released under Apache 2.0 that directly...

A developer's $219 monthly AI subscription stack can be replaced by 10 self-hosted open-source tools, covering...

Ling 3.0 Tiny, a 7.9B parameter mixture-of-experts model with 1.3B active parameters, impresses in coding tests...

DeepSeek V4 Flash achieved 14-cent million-token pricing through native sparse attention, compressed KV cache, and a...

Bijan Bowen tests Traycer, an open-source multi-agent orchestration tool, by having large models like Fable 5 and...

LM Studio's Bionic is a local agent that runs open-weight models on your machine, offering a free alternative to...

The top 10 AI repos of July 2024 are all scaffolding around existing models, not new models themselves. Only two...

Waste is a 6,000-line C engine that runs the 2.78-trillion-parameter Kimi K3 mixture-of-experts model on a laptop...

DeepSeek V4 Flash 0731 is a powerful local AI model that excels at reasoning and benchmarks but tends to overthink...

DeepSeek V4 Flash is a newly released official version of a small but powerful AI model with 284B parameters (13B...

Bottle Cap AI's ThinkingCap fine-tune of Qwen 3.6 27B reduces reasoning tokens by ~46% while preserving benchmark...

Abacus AI's supercomputer, using GPT-5.6 Sol and Fable 5 in max mode, autonomously built and deployed a full-stack...

Hermes Agent's Quicksilver update introduces smart approvals, durable background jobs, delivery ledgers, profile...

Ling 3.0 Flash from Ant is a surprisingly competent coding model at 124B total/5.1B active parameters, outperforming...

Anthropic's Claude Opus 5 was released, outperforming the larger Fable 5 on most benchmarks while costing about half...

Poolside Laguna S2.1 is a 118B parameter Mixture of Experts model (8B active) that excels at creative writing and...

Local and open-source AI models are rapidly improving in capability density, with newer efficient models...

Claude's token limits can be avoided by optimizing token consumption and model usage without increasing cost. The...

Prism ML's Bonsai 27B models (ternary and 1-bit) dramatically shrink a Qwen 3.6 27B base while retaining surprising...

Combining Fable 5 with Hermes Agent enables powerful, cost-effective AI workflows by using cheaper models for data...

Austin Marchese presents a six-step roadmap for using Claude Code to build 10x faster, based on the 'T-shaped...

Tencent released the full version of Hy3, a 295B parameter mixture-of-experts model with 21B active parameters and a...

The Fable 5 model powers a new 'agentic operating system' across five levels: integrating into the core OS, unifying...

Tencent HY3, now fully released under Apache 2.0, is a 295B parameter mixture-of-experts model with 21B active...

MiniCPM-5 is a 1B parameter dense model from OpenBMB that aims to be a 'cognitive core' — a small, on-device model...

Building a self-improving system with Claude Code requires a five-step framework: setting up a knowledge base and...

Ornith 1.0, a family of fine-tuned local coding models (9B dense and 35B MoE variants tested), shows promising...

Ornith 1.0 from Deep Reinforce introduces self-scaffolding LLMs for agentic coding, where the model learns to...

Hermes Agent allows swapping different AI models for different tasks to optimize cost and performance, breaking the...

The Hermes agent integrated with Obsidian creates a powerful second brain system where markdown files become "living...

Claude Code, combined with a free open-source plugin, can perform a full SEO and AI Search (GEO/AEO) audit and...