Latest
A $7 ESP32 Gives an AI Agent Hands on a Real iPhone
An ESP32-S3 board plugs into an iPhone and acts as an AI agent as eyes and hands over AirPlay and USB HID. Here is the full setup, and the limits that bite.
Docker sbx: What a Sandboxed Coding Agent Can Still Delete
Docker Sandboxes runs coding agents in a microVM with full autonomy on by default. Here is what crosses the trust boundary, and what your agent can still delete.
Claude Duet vs Grok Bot vs Muse vs Dots: What You Pay
Four always-on AI agents, priced and compared: Duet is free on Claude Pro but US-only, Grok Bot starts at $30/mo, Dots needs a $100 ChatGPT Pro plan.
text-embedding-3-small vs large: Which to Start With
A guide prices text-embedding-3-small at $0.02 per 1M tokens with 1,536 dimensions, and the large model at 5x that. When the upgrade is worth it.
The M4 Mac mini Linux GPU Driver Was Mostly Written by AI
Two developers shipped an OpenGL ES 3.0 conformant Linux GPU driver for the M4 Mac mini in about a month, with coding agents writing most of the code.
Kimi K3.1 Pro vs Claude Opus 5.5: Half the Price Per Token
Kimi K3.1 Pro lists at $2/$10 per 1M tokens, half of Claude Opus 5.5 and level with Sonnet 5.5 — but cached input pricing flips it for agent loops.
Only 19% of New Pages Get Cited by ChatGPT or Claude
A 4,000-article study puts ChatGPT median time to first citation at 3.5 days and Claude at 9.5 - but four in five pages got nothing in 60 days.
Anthropic's OSS Scanner Sends AI Vuln Reports Unverified
Anthropic's OSS Scanner mails model-written vulnerability reports straight to maintainers with no human validation first. Who qualifies, and what it costs.
OpenAI's Decisions API Goes Live, Taking Direct Aim at Jev
OpenAI's Decisions API is now generally available, cutting latency to 100-180ms — but public benchmarks still rank TypeSafe's Jev above it.
1Password for Agents Signs In Without Sharing Passwords
1Password for Agents lets AI agents complete logins they never see the credentials for, using a digital key and per-sign-in approval on your own device.
Claude Duet: Anthropic's Always-On Agent Hits Preview
Anthropic's Claude Duet works between your prompts on Claude Opus 5.5, with 6,000+ connectors, a cloud VM, and an approval gate on every sensitive action.
DeepInfer vs vLLM on 8x B200: One Run, a Split Verdict
GPUStack benchmarked DeepInfer-server against vLLM on eight NVIDIA B200 GPUs: 25.2% more throughput one way, 2.09x the time to first token the other.
A New Tokenizer Cut One Team’s Inference Cost by 77%
Dharma-AI says rebuilding the vocabulary of a Qwen3-4B essay grader cut measured inference cost 77% and raised quality — with the same weights and recipe.
