Fast move on the workhorse tier. Matters if you build anything agentic 👇 The news: 3.7 Flash launched just three weeks after 3.6, at half the token cost — $0.75/M input, $3.75/M output, locked through Dec 31. The gains are real: → DeepSWE coding: 49% → 65.3% → Reportedly beats Claude Sonnet 5 and GPT-5.6 on coding → 1M context, better multi-step planning and tool calls The bigger move: Gemini Spark — Google's 24/7 personal agent for Pro/Ultra — now runs on 3.7 Flash. The agent working between your sessions just got a real bump, especially for Workspace tasks. My take: Google isn't chasing the frontier crown right now — their flagship 3.5 Pro is still delayed. They're dominating the workhorse tier instead: the cheap, fast, good-enough model you actually run at scale. Smart. Most agentic workflows don't need the smartest model — they need a reliable one cheap enough to run thousands of times. For builders: The intro pricing is a defined window to test cost-per-task before rates double in January. Use it. And this is the price war I keep flagging, now three-way — Claude doubled Cowork limits, OpenAI gamified resets, Google halved Flash pricing. Every lab's fighting for the builder. You win. Stay model-agnostic and ride whoever's cheapest for the job. Caveat: still no EU availability. Check your region first. Running Gemini Flash in any workflows yet, or all-in on Claude/OpenAI? 👇

New agent from xAI + Cursor — the easiest way I've seen to get a team of AI agents working for you. No canvas, no nodes, no terminal: you hire a bot like a person (name, job title, one-line description). Each gets its own cloud computer and signs into your tools like you do, so it doesn't need an API. - What's genuinely good. Bots read each other's job descriptions and hand work off on their own — no wiring. And instead of writing an SOP, you hit record, do the job once on screen, and it writes its own instructions. - The catch. $200/mo via Cursor Ultra (a swap, not an addition — stacked on a $200 Claude plan that's $400 for overlapping tools). Trial won't get you far, it runs Grok only, and every skill you teach lives in their cloud. The lock-in isn't the subscription, it's the training. My take: never touched Claude Code? Best on-ramp there is — but run it like an intern, read-only and draft-for-approval for a month. Already in Claude Code/Codex daily? Don't switch. On Hermes/OpenClaw on a $5 VPS? You already solved always-on. Three questions decide it: weekly browser work with no integration, mistakes cheap to catch, and $200/mo less than that time costs you. 📎 Full guide PDF pinned below — setup, pricing math, the three verdicts, and the limitations. 👉 What are you running now — and is this pulling you over? Drop it below. [Watch the video here ▶️]

Most of you in here can already build the thing. That was never the bottleneck. The bottleneck is that you can't answer four questions in one sentence each: **who you sell to, what you fix, what you charge, and how you reach them.** Almost everyone stuck is missing at least two — and no tutorial you watch this week is going to hand them to you. So we're running the challenge again. Three days, one decision each day 👇 - Day 1 — Your baseline and your path. What you can already demonstrate today, what proof and relationships you already have, the hours you actually have, what you can spend. Then you pick one lane: build it, deploy it in a business, or sell it to clients. Most people are stuck because they're quietly trying to do all three. - Day 2 — Your niche and your buyer. One market. Who signs, what the expensive problem is, whether a customer is worth enough to make the math work, whether you can reach them, and whether you can deliver after the sale. This is the day that kills most AI businesses before they start — and it takes an afternoon to get right. - Day 3 — Your offer and your route. A one-sentence outcome. What's included and what isn't. A price. A risk reversal limited to what you actually control. One proof asset. One acquisition route. And a 14-day activity commitment with a real number on it. - Day 4 — You submit it. Your Launch Audit gets reviewed properly — against a rubric, with either an approval or the exact sections to fix. Then you get a personalized Roadmap Review: your biggest constraint named, your next 30–90 days mapped. Annual members and the strongest submissions come straight to me. Then the 90-day roadmap unlocks and you go execute the plan you just built. I'm not promising you a paying client in three days. That's not in your control or mine. What you *will* walk out with is a defined market, a proven offer, a price, an acquisition channel and a reviewed plan — which is the exact thing you don't have right now, and the reason the building hasn't turned into money yet.

No plugins, no code. Three folders on your computer — Inbox, Brain, Output — plus one file called CLAUDE.md that tells the AI where everything lives. Here's why it matters 👇 You're sitting on years of pricing, proposals, SOPs, client notes and meeting recaps that AI can't touch, because it's scattered across a laptop, three inboxes and a Drive folder nobody's opened since 2024. So every chat starts with you re-explaining your own business. The fix is the file everybody skips. CLAUDE.md is what Claude reads first, every time — what your business is, what's in each folder, and which file is the real one when you've got five versions of your pricing. Skip it and it opens your folder and guesses, then quotes a 2024 price to a live lead. I put together a short step-by-step guide with the exact folder setup, a copy-paste CLAUDE.md template, the prompt that sorts your whole Inbox into a clean Brain, and the one rule that makes every answer name the file it came from. 👉 Grab it here: https://claude.ai/code/artifact/f4da52f7-5dc8-4d41-a06a-6b88ac5755eb Build it, then tell me the first question you'd ask it about your own business 👇

OpenAI shipped Record & Replay. You hit record, do a task on your computer one time, and it turns what you just did into a skill it can run for you from then on. No prompt. No script. No developer. Here's why it matters 👇 Think about the weekly report someone on your team spends two hours pulling from the same three dashboards. Or the onboarding checklist only one person knows how to run — the one that stalls the whole business the week they're on holiday. That's not a hard job. It's just a job that lives in someone's head. The fix is picking the right task. It has to happen every week, take more than 20 minutes, and be the same steps every single time. All three, or skip it — anything that changes every run still needs a person. I put together a short step-by-step guide that shows you exactly how to pick the task, how to record it so it actually replays, the prompt that tightens the skill afterward, and a fallback if you're on Windows or in the UK/EU where it hasn't rolled out yet. 👉 Grab it here: https://docs.google.com/document/d/1lEDnzJnoNkjsQ1B4fi08ddNE02HzS4cV5XYDJejsu3M/edit?usp=sharing Do your three worst weekly jobs and you'll get a full day back every month. Drop the one task you'd record first 👇