Please don't discontinue Gemini 2.5 Flash

Write code like a human will maintain it

Unified Memory, Explained: Why Mini PCs Can Run 70B Models a Big GPU Can't

I think I have LLM burnout

Automating AI Away

GLM 5.2 and the coming AI margin collapse

Price per 1M tokens is meaningless

Pruning RAG context down to what the answer actually needs

A global workspace in language models

What Emily Bender meant by "stochastic parrots"

Fable 5 On Vending-Bench: Misbehaving, With Plausible Deniability

Anthropic's Method to Losing Goodwill in a Few Easy Steps

New AI tutor achieves 0.71-1.30 SD effect size in Dartmouth course [pdf]

Claude Design System Prompt

Is The Economist Always Wrong?

GPT-5.5 Codex reasoning-token clustering may be leading to degraded performance

Leaking YouTube creators' private videos

2026 Unslop AI-Written Fiction Contest Results

Performance per dollar is getting faster and cheaper

I Wasn't Allowed Prompting ChatGPT During My Chalk Talk: This Is Discrimination (2025)

60% Fable cost cut by converting code to images and having the model OCR it

Claude, please stop trying to memorize random crap

Jamesob's guide to running SOTA LLMs locally

It Still Can't Do My Job: Four Years of Moving Goalposts (2022–2026)

Instead of banning AI, I made a classroom contract with my students

Claude-real-video － any LLM can watch a video

No LLM Code in Dependencies

Is One Layer Enough? A Single Transformer Layer Matches Full-Parameter RL Train

The Fall of the Theorem Economy

Claude Fable 5 export control lifted

Claude Sonnet 5

Words Are a Byproduct of Consciousness. For LLMs, It's Backwards

LongCat-2.0, a large-scale MoE model with 1.6T total and 48B Active

Qwen 3.6 27B is the sweet spot for local development

Knowledge Distillation of Black-Box Large Language Models (2024)

Do LLMs pass the mirror test?

Wayfinder Router: deterministic routing of queries between local and hosted LLM

U.S. government will decide who gets to use GPT-5.6

Previewing GPT‑5.6 Sol: a next-generation model

The Exhaustion of Talking to a Tool

US Govt to individually approve who gets GPT 5.6

AI children's books, body horror edition

Political bias in AI: Where the AI models stand

The Unbearable Cheapness of Open Weight Models

Bible as RAG Database

OpenAI and Broadcom unveil LLM-optimized inference chip

Qwen-AgentWorld: Language World Models for General Agents

AI's Affordability Crisis

Unsloth GLM-5.2 – How to Run Locally

Prompt Injection as Role Confusion