All Stories

Casey Muratori advocates hardware-first performance engineering

Sentence Transformers details ColBERT-style medical retrieval tuning

Pi lets users constrain and cycle model shortlists

OpenCode proposes retention as a model quality metric

Nvidia reportedly agrees to buy Hugging Face for $12.9B

Jacob Peake maps the tradeoffs across AI chip architectures

Dwarkesh maps the compute pricing advantage of major AI labs

X Chat gives developers programmatic agents with public handles

Composio bridges Grok Bot to more than 1,000 integrations

Cua ports Omarchy to Apple Silicon with agent desktop control

Runme combines WebMCP notebooks with bring-your-own agents

DeepLearning.AI teaches agents to turn traces into reusable skills

Flint AI's Switch puts cross-framework agents in shared channels

VS Code Agent Host keeps agent sessions persistent across clients

Lovable connects its workspace to Google's Antigravity agents

Artificial Analysis details a reward-hacking judge for Terminal-Bench

OpenCode agents combine observability, payments, and databases for fraud triage

Higgsfield documents one-photo video swaps with Seedance 2.5

Claude Code drafts feedback reports after failed tasks

Linear reports agents in 95% of paid workspaces

Wan 3.0 leads a practitioner test of raw video editing

Meta's Project OT explored replacing teams with AI-native workflows

Codex visualizer improves cloud coding-session progress views

Mac Studio outpaces clustered DGX Sparks on generation bandwidth

Cerebras plans 3D DRAM for its wafer-scale inference system

Model Discovery Agent combines LLM proposals with Bayesian design

South Korean labs climb Artificial Analysis intelligence rankings

Makora launches on OpenRouter with DeepSeek, GLM, and Kimi

Pi exposes custom slash command registration for agents

Jakub Krehel adds UI stress-testing and variant skills

Vercel AI Gateway sends webhooks after long generations

Factory Droid connects agents to 50-plus enterprise tools

Google AgentHands syncs XR gestures to spoken agent guidance

Grok Bot workforce checker monitors stalled multi-bot work

MiniMax M3 completes email workflow for $0.018

ChatGPT extension passes open browser tabs into Codex tasks

Agent work design framework maps eleven firm-worker configurations

OpenCode says Stripe blocked $300M in card fraud in days

Cline reports Stripe blocked $8.4B in recent abuse

Simile argues query-level confidence is central to simulation

DeepLearning.AI maps skills needed for production AI systems

Claude unifies memory across Chat and Cowork with controls

Tailwind StyleX exposes typed Tailwind tokens in StyleX

Ramp AI Index shows paid AI spend remains dominated by chat

Pi 0.84.3 adds PowerShell support and faster startup

AffineGaps brings high-performance genomics primitives to Mojo

Interconnects tracks Chinese open models rising in AI research

MiniMax publishes an integration directory for H3 video workflows

llama.cpp moves documentation into a dedicated new home

vLLM reports AgentX performance on real multi-turn coding traces

Latent Dynamics Reasoning improves video world-model extrapolation

Korea advances three teams in its sovereign AI model program

VS Code adds one-click visibility and pinning for language models

Xiaomi's XRING O100 details a 3D-stacked on-device accelerator

Prof. Tom Yeh walks through self-attention calculations by hand

Sander Dieleman reviews the return of continuous diffusion for language

François Chollet recommends Deep Learning with Python for attention

MiniMax Sol Engine cuts 768p video latency to 14.93 seconds

Cartwheel maps compute-optimal scaling laws for human motion models

OpenRouter launches visual benchmarks for image model selection

RotNet offers a concise refresher on self-supervised vision learning

ElevenLabs CLI v1 brings voice-agent management into version control

Browser Use becomes available in Grok Build

MCP publishes a roadmap for long-running agent workloads

Vercel Functions extends Bun package sizes and execution duration

OpenAI-branded workspace invitations expose a phishing delivery pattern

Conductor MCP lets coding agents create and manage cloud sessions

Devin supports cloud sessions from desktops to Kubernetes

Tufa Labs reclaims the ARC-AGI-3 high score with an open solution

Claude rebuilds its streaming renderer for smoother long responses

rst replaces the legacy terminal reset delay with a Zig utility

Artificial Analysis adds Qwen3.8-27B to Optima custom benchmarks

Uncle Bob warns agents can inherit conflicting local rule files

OpenAI positions its API around the full model cost-performance curve

Amp previews a configurable model router for coding workflows

ArchAgent v2 beats a human cache-prefetching champion

Ethan Mollick cautions against negative claims from outdated models

RTX 5090 runs DeepSeek V4 Flash 284B at 24 tokens per second

Wafer AI expands its GPU performance engineering resource list

On-policy distillation guide explains privileged teacher supervision

RunInfra serves Parakeet TDT 0.6B through an OpenAI-compatible API

Vercel says Sol price cuts made it its fastest-growing frontier model

Nvidia reportedly licenses Poolside's Model Factory research system

Wispr Flow user reports voice now drives most agent communication

Grok Bot workflow triages 625 issues and pull requests

Unslop skill experiment finds style prompts affect model reasoning

Adversarial Review uses three agents for evidence-grounded code review

Amp Orbs gain friendly subdomains and custom domains

Devin escalates blocked agent work to humans by email

World Robot Conference showcases companion humanoid hardware

Antigravity CLI builds a live global flight radar with free APIs

Netflix describes a production lifecycle for LLM recommendation judges

AutoDesign improves weak agents by optimizing their surrounding harness

MerchantBench exposes long-horizon agents that gradually stop acting

Google Antigravity exposes Turbo mode with no permission barriers

Matt Pocock integrates Remotion into an AI-editable video editor

Gemini Flash turns a simple skill into PyTorch visual explainers

Dinkus launches a local-first visual Markdown editor for Mac

Venice makes Wan 3.0 available with 30-second reference video

Prof Tom Yeh walks through a graph convolutional network by hand

Grok Imagine adds a discovery page for image editing workflows

Medical LLM evaluation thread challenges broad claims from old models

Vercel AI Gateway reports 62% open-model token share

NVIDIA CUDA harness solves all ARC-AGI-3 public levels

Prof. Tom Yeh walks through a transformer block by hand

LMCache guide explains why compaction can raise agent costs

Microsoft Agent Lightning trains agents through deployment harnesses

CMU proposes forking-sequences for multi-horizon forecasting

ClawGym II trains models through black-box agent harnesses

Sebastian Raschka explains Claude text watermarking

Cloudflare Worker isolates process concurrent requests

Hermes documents Himalaya CLI for terminal email agents

Task Model Induction converts computer traces into agent skills

QWM uses world models for robot test-time search

Keep turns saved content into agent-searchable Markdown

AgentMail warns Gmail can flag API-only agent accounts as bots

Brett Adcock demos autonomous vertical part handling

Gergely Orosz urges vendors to ship MCP interfaces

SwarmForge adds a UI for six-pack agent squads

TanStack AI enters release candidate with AG-UI support

Hermes Agent fixes duplicate and lost outgoing messages

Amp explores replacing message queues with agent steering

Dan McAteer traces the evolution of the agent harness

AutoResearchEval finds agents ignore their own failure warnings

Theo's model tier list puts Fable 5 above GPT-5.6 Sol

DocWriter builds a multi-agent writing-style calibration flow

ACID-Agent prevents failed steps from polluting agent memory

Rork Max installs iPhone apps directly from Windows

Google DeepMind formalizes cost-aware AI model routing

UC Berkeley FreeToken accelerates local model inference

Ollama adds Kimi K3 to included Pro and Max usage

Split-MoE dynamically divides inference between CPU and GPU

Gemini 3.7 Flash matches top CAD computer-use performance

Gergely Orosz maps natural-language cron jobs for Sheets

FrankenGit builds a repairable agent-native Git forge

Lovable builds custom MCP apps over business systems

DeepSeek V4 Pro posts ARC-AGI scores and task costs

Google DeepMind uses AlphaEvolve to improve a matrix bound

Adam Sandler Markdown second-brain playbook targets AI context

Tencent brings Hy-MT2 translation models to OpenRouter

Lovable imports Manus exports and rebuilds the project

Anthropic ELI5 skill creates visual HTML explainers

ChatGPT Sites lets users build and share hosted apps

Magnific workflow preserves character continuity in AI video

GitHub Copilot generates VS Code debug configurations

MLCR-AA exposes completeness gaps in medical long-context models

Venice builds short films from a premise with an agent harness

Artificial Analysis exposes accuracy gaps across model endpoints

Google says Gemma has passed one billion downloads

Switch Transformers show scaling ratios depend on the task

Grok 4.6 supports an OpenAI-compatible API migration path

Ultra-FineWeb-L1 opens a 1T-token web training corpus

Study finds post-training agents rarely revise initial strategies

Earendil explains the four core layers of an agent harness

Claude Managed Agents ships an AG-UI integration cookbook

Muse Spark 1.2 demonstrates multimodal robot tool use

Vercel Functions adds Bun 1.4 with WebSocket support

Grok Ship packages a persistent agent software factory

TanStack AI adds resumable full-duplex WebSocket streaming

Devin opens dedicated Slack code channels for active work

Linear Coding Sessions adds browser testing and custom environments

TanStack AI adds typed interrupts for human-in-the-loop agents

OpenAI expands Computer History and Record and Replay in Europe

Browserbase shows Stagehand agents using parallel browser sandboxes

Paper argues AI scientists should be evaluated as human-agent teams

Hermes Agent offers provider-neutral remote computers

Claude browser tool uses page structure instead of pixels

Cursor adds /goal and isolated VM subagents

Grok Bot CTO workflow delegates repo work to child agents

Goodfire offers $1M in Silico grants for interpretability research

Meta previews WildArtifactBench for multimodal agent evaluation

Unsloth Desktop adds experimental automatic context compaction

OpenClaw, Hermes, and Grok Bot get a persistent-agent comparison

Codex turns desktop Computer History into reusable workflow ideas

ChatGPT adds Apple Messages plugin for Work and Codex

Generative UI Bench compares model output formats across providers