Every month a cheaper model ships and the group chat says 'switch.' The rate card is the wrong number to switch on: an agent's real cost is tokens-per-task times price times a retry penalty, and only one of those three is on the pricing page. Here's the reusable test — freeze your tasks, measure completed-task cost, decide in an afternoon — with Gemini 3.6 vs 3.5 Flash as the worked example.
Palantir and Torq veterans took an $8M seed to discover every agent running against your systems, profile its behavior, and pull a kill switch when it drifts. The round is early; the gap it names is not.
Five verified moves for a team of one: a self-propagating npm worm that hunts AI-coding-agent credentials, DeepSeek warning it will raise the cheap-token floor, Claude Code turning auto mode on by default Aug 14, Rippling shipping a spend console after its own AI bill grew 80% a month, and the EU quietly slipping its high-risk deadline to 2027.
Spot GPUs are the same H100s at 60–90% off — until the provider reclaims one mid-job. The discount isn't the number that matters. The notice window is.
Three ways to extend Claude Code, and founders keep picking the wrong one — building an MCP server when a skill would do, or writing a skill for something that needs live data. The rule of thumb is one sentence, and the 2026 answer is usually 'all three, layered.'
A copy-paste setup that wires three coding agents to the same searchable memory over MCP — so a fact one of them learns is a fact all of them know. Ten minutes, one npm package, no API key required.
A San Francisco startup 11 months old raised a $35M Series A to move AI agents from a working demo to production, with a drop-in Router it claims cuts agent runtime cost up to 10x. If your agent looks great in a demo and falls over at scale, this is your category now.
Kimi K3 tops the Frontend Code Arena but is a rack to self-host and priced like a flagship. The right way to capture the win is task-based routing: send only your UI calls to K3, keep everything else where it is. Here's the router, the cost guardrails, and the math.
Alibaba's 2.4-trillion-parameter model is slated to open its weights this month. The headline is smaller than Kimi K3, but the number that sets your token bill — 95B active — is nearly double. Here's the serving math, and why it pushes the rent-vs-own line further toward 'just use the API.'
Within 48 hours in early August, OpenAI's Codex CLI and Anthropic's Claude Code both shipped installable plugins — and picked opposite defaults. Codex bets on a searchable catalog; Claude Code bets on a raw ZIP over HTTPS with a checksum. If you're publishing an extension, the choice decides who finds it and how they trust it.
The Agent Skills format is now an open standard that 30-plus tools read. So the same SKILL.md folder can run in five different coding agents — but only the frontmatter and body travel cleanly. Here's the install path for each, and the three things that quietly break portability.
The Aug 4 release makes Ollama's /v1/chat/completions streaming match OpenAI's wire format byte-for-byte: role on the first chunk, finish_reason on its own chunk, usage in a separate one. If you kept a fork of your streaming parser for local models, you can delete it.
A $1B acquisition just made 'non-human identity' a real budget line. Here's what it means when your AI agent needs credentials — and the five moves that give it access without handing it a password you can't revoke.
Meta's Muse Code just made it a four-way race. Here's the actual buying decision for a team of one — price, data terms, and background-agent throughput, not a benchmark leaderboard.
Meta shipped Muse Spark 1.2 on August 5 at the same $1.25/$4.25 price as 1.1, but the three points it added on the intelligence index landed almost entirely in agentic work: its real-world-task Elo jumped 260 points and Terminal-Bench climbed to 82.9%. For founders, the question isn't whether it's frontier — it's whether a same-price, better-at-agents backend earns a slot in your router.
You run Claude Code, Codex, and Cursor on the same repo — and each one starts from zero. Four open-source tools fix that by sharing memory across agents. They disagree on one thing that decides which you want: who controls what gets remembered.
Six weeks ago you could run all three agent-memory layers on your own hardware. You can't anymore — Zep deprecated its self-hostable Community Edition, so the choice is now as much about where the code runs as how it remembers.
Kitesurf throws out Chromium and runs the whole browser in V8 isolates on Workers. It's 3–7× cheaper on CPU and memory and ~1.7× slower per page. For an agent firing thousands of short page loads, that trade is the point.
Every tool you register rides in the model's context on every turn, so verbose schemas quietly inflate your input bill. Trim each description to its load-bearing job, measure the drop, and A/B for accuracy — the same move that cut a Deep Agents turn's input tokens ~65%.
The complaint is never that skills give bad instructions — it's that they never fire. The one field that decides whether a skill loads is the description, and most are written too vague and too polite. Here's how to write one Claude reliably picks up.
Claude Code's August build moves any main-conversation MCP tool call that runs past two minutes into a background task, so a slow database query or deploy call stops locking up your shell. Here's exactly what changed, the one environment variable that controls it, and when to turn it off.
Claude Code v2.1.224 shipped self-hosted environments in public beta: cloud sessions started from the web, mobile, desktop, or a scheduled routine now execute inside your network. Here's what it is, who it's for, and the exact setup — plus the one-runner-per-user rule that decides your fleet size.
Qwen3.8-Max shipped on August 3 speaking both the Anthropic and OpenAI wire formats, so you can run your existing agent CLI on it by changing three environment variables. Here's the exact setup — plus the one Codex gotcha that will waste your afternoon.
llm 0.32 shipped a primitive that most agent frameworks make you build by hand: a tool can raise llm.PauseChain to stop the loop before it does something irreversible, hand control back to you, and resume later without re-running the calls that already finished. Here's the exact pattern — pause, persist, approve, resume — in about 40 lines.
Whole-value masking hides a bare token fine — but it corrupts a JWT your code decodes or an AWS key the SDK signs with. Claude Code v2.1.224 adds three structured fields (extract, decode: jwt, awsPairs) that keep the tool working while the agent still never holds the plaintext. Here's the exact config for each.
Deep Agents v0.7 dropped a default turn from 5,395 to 1,895 input tokens with no quality regression. The savings came from deleting prose that duplicated the tool schemas the model already sees — a tax your own harness is almost certainly still paying.
v2.1.224 (August 7) deleted the hard per-session ceiling that made long orchestrations fail at agent 201. It didn't make fan-out unbounded — it moved the real limits to concurrency, nesting depth, and a budget cap that finally halts running background agents. Here's the new mental model and the three env vars that set it.
Claude Code v2.1.224 added cross-session SendMessage — one running session can now message another, on any of your machines, and discover peers with ListAgents. Here's how it works, the two settings that gate delivery, and when to reach for it instead of a subagent.
The headline is a vanity metric. For a founder, the milestone matters only because it settles a strategic question you were probably still hedging on — whether ChatGPT is a channel you ship into, an answer surface you get cited on, or a competitor you build around. It's now all three, and you have to pick.
OpenAI's Atlas browser stops working August 9 with no automatic data migration. If you wired an agent to it, here's the export checklist and the honest decision between ChatGPT's desktop app, Comet, Claude in Chrome, and the open-source escape hatch.
🎧 Listen
The Stack
