Videos

OpenAI's GPT-5.6 Soul is offered at both 40-50 tokens/sec on GPUs and 750 tokens/sec on Cerebras chips (18.5x...

GPT-5.6 Sol and Claude Fable 5 were compared across multiple challenging tests including 3D printing, magazine...

Erik Meijer argues that LLMs with tool calls are intrinsically dangerous because their agentic loop can produce...

Engram CEO Dan Biderman argues that long context windows alone cannot solve the AI memory problem; models need...

Fable 5 is an expensive reasoning model intended for strategic, multi-step tasks, not quick chatbot usage. By...

Phaidra engineers solved the problem of LLMs failing to handle 500,000 sensor names in large data centers by using a...

Cactus Needle is a 26-million-parameter open-source function-calling model that runs efficiently on consumer...

AI agent hallucinations can be reduced by moving guardrails from prompts into code. Five techniques—semantic tool...

Machinecraft built a 36-agent AI system called Eira that runs their entire go-to-market without a data science team...

Meta's Muse Spark 1.1 shows benchmark improvements over its predecessor but fails in practical coding and game...

GPT 5.6 Soul outperforms Fable 5 in three out of four real-world coding and design tasks at a fraction of the cost,...

The AI Engineer conference (AIE) was founded by Swyx and Ben to professionalize AI engineering, inspired by earlier...

Pydantic AI's 2.0 release introduces "capabilities" as a single composable primitive that bundles tools,...

OpenAI released GPT-5.6 as a trio of models—Soul, Terra, and Luna—each with different cost and capability tiers,...

GPT 5.6 Soul is significantly cheaper and more token-efficient than Fable 5, but Fable 5 produces more creative and...

Grok 4.5 is a 1.5-trillion-parameter frontier model from xAI, built on a new V9 foundation that incorporates...

GPT-5.6 Soul is a newly released model from OpenAI, part of the GPT-5 pre-training run, offering frontier-level...

Fable 5 is a powerful but expensive model; its real advantage comes from strategic use rather than raw capability. A...

AI engineers are eating the world as models and agents rapidly advance, with OpenAI shipping new models every six...

Building agent loops is the new prompting. Developers should focus on creating verification loops and specifying...

Personal AI agents built on markdown-driven second brains (like the Karpathy LLM Wiki) are powerful for individual...

Claude Fable autonomously built a complete business from a single goal prompt, producing a landing page, product,...

Theo Browne argues that AI models are improving faster than developers can adapt, so builders must think bigger and...

AI is transforming game development by removing skill barriers and enabling rapid iteration, but creating standout...

Anthropic's research reveals that large language models like Claude develop an internal 'J-space'—a hidden layer of...

Agents are slower than LLMs primarily because they rely on external tool calls (e.g., fetching web pages, making API...

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

The key insight is that the model itself is not the moat; the process and systems built around it matter more. By...

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...

Anthropic's Thariq Shihipar introduces Fable, a new Claude model that represents a major leap in capability, akin to...

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

Shipping an AI agent is only the beginning; the real work starts post-launch with monitoring, understanding, and...

Continual learning for AI agents requires turning failures into durable improvements by learning from experience...

Nate Herk demonstrates building an LLM-powered personal wiki (second brain) using Obsidian and Claude Code (Fable)...

Loop engineering is the latest evolution in agent scaffolding, stacking another autonomous loop outside of harness...

Fable 5 is Anthropic's latest high-intelligence model, released after a 17-day suspension for safety overhauls, and...

Fable 5 is back after a brief ban, and the speaker considers it a step-change in model capability, especially for...

The prompt interface for large language models is structurally identical to the batch-processing punch card: the...

Three major AI companies (SpaceX, Anthropic, OpenAI) are filing for IPOs at a combined ~$4 trillion valuation,...

Google's Open Knowledge Format (OKF) standardizes the structure and metadata of LLM wikis, building on Andrej...

Claude Fable 5 is a powerful but expensive model that benefits from short, clear prompts with context, negative...

Anthropic released Claude Sonnet 5 alongside a Linux desktop app beta. The model offers improved agentic coding and...

Hermes agent's latest update introduces a Ministry of Experts feature that uses a panel of models orchestrated by a...

Mixture of Agents (MoA) is a new feature in Hermes Agent that consults multiple AI models from different providers...

GLM 5.2, an open-weights model, is compared head-to-head against Claude Opus 4.8 across several challenging tasks: a...

Gavriel Cohen, creator of NanoClaw, outlines a blueprint for autonomous work agents, emphasizing personal agents...

Rachel Lee Neighbors argues that developers can drastically cut costs, improve latency, and enhance security by...

Erik Hanchett from AWS presents five practical strategies to reduce token costs when building and deploying AI...

Lech Kalinowski built a physical AI-native terminal that combines an OLED and e-paper display to serve as both a...