Inside: 5 'Grok Bot' Guides
Top Insights
xAI's 5 Guides to Running Teams of Grok Bots
xAI's team wrote five guides to running Grok Bot, where each bot is an always-on agent with its own computer. A product manager, a sales lead, a designer, and a solo game developer each show the bot setups they've built and what those bots do all day.
💡 Key Insights:
- Your to-do list stops going stale by lunch...a Chief of Staff bot watches the product manager's Slack, email, and meetings, keeps a live "attention list" of what's in play, and archives email that isn't.
- Teaching a bot a job takes as long as doing it once...Meta wouldn't approve the developer's ads API, so they recorded one ad upload by hand, and the bot replays that click-path daily.
- Six bots ran a game studio's ad testing on their own...they saw players arriving as couples, swapped the party ad for a partner one, and drove installs from $15 to $1.
Using AI to Increase Your Intelligence & Enrich Humanity
Dr. Fei-Fei Li, the Stanford computer scientist behind ImageNet, the photo library that set off modern AI, joined the Huberman Lab podcast. She walks through how AI learned to see, what it can and can't do with the internet it was trained on, and where creative work goes next.
💡 Key Insights:
- Which part of your work AI can copy and which it can't...a model mimics anything ever uploaded, and Li uses Picasso to show the judgment that never was.
- How to get a sharper answer from any model...tell it who you are, let it learn your writing so it sharpens your own, and prompt the way Socrates asked questions.
- What's coming for anyone who makes content...AI already assembles near feature-length films, Li's startup World Labs turns a sentence or sketch into a 3D world, and re-skilling creators are who she builds for.
Watch/Listen to the Full Episode
🚀 WATCH: How These AI Copy Bots Are Producing World-Class Sales Copy 50X Faster Than Even The "BEST" Copywriters On The Market…
(Plus… They Don't Get Sick, Miss Deadlines, Or Ask For Raises Either!)
Watch the full AI Copywriting Tell-All Video Here
Prompt of the Week
This Week's Feature: The Agent-Ergonomics Rewrite
Your agent just handed you a project plan. It designed that plan to build the thing, and gave zero thought to the system it'll have to navigate and steer once the code exists.
Jeffrey Emanuel's follow-up prompt fixes that. You paste it right after the plan lands, and the agent redesigns everything from the driver's seat: it rewrites your design docs so the whole system becomes coherent, modular, and legible to the agent operating it.
The payoff is an agent that understands state accurately and controls the build with far fewer wasted tokens and wrong turns.
THE FULL PROMPT:
OK, now I want you to think deeply about how to make this entire system as agent-intuitive, agent-ergonomic, and agent-accretive as you can possibly imagine. Put yourself in the driver's seat and imagine that YOU are the one using this system and driving it. What would most enable you to do an awesome job understanding the situation accurately and optimally controlling everything to drive the best and most accurate results possible, with the least expenditure of resources?
Then make all the requisite changes to the various design documents and plans accordingly. Don't just think of the project as an assemblage of various parts or components: really try to profoundly and deeply conceptualize it as a synthetic SYSTEM that is maximally coherent, cohesive, modular, and interconnected, forming a tower of linked abstractions that are maximally legible to you as an agent. Really ruminate and meditate on all of this incredibly deeply before responding or taking any actions.
Tool to Try
Poth Labs is a diagnostic layer for your customer data, backed by Y Combinator. Ask it anything about your customers... why new accounts are stalling, which ones are at risk, what a feature is really doing... and it answers with the exact call excerpts, tickets, and usage rows behind the claim.
AI Tool Highlights:
🔎 Ask Anything About Your Customers: Type a plain question like "why are new accounts stalling"... Poth reads across your connected calls, tickets, and rows and comes back with an answer in seconds.
📎 Every Answer Comes With Receipts: Each claim ships with cited source cards you can open... click one and read the exact call excerpt, support ticket, or usage row it came from, so you can check the answer instead of taking it on faith.
🗂️ The Filing Cabinet With the Analyst Inside: Poth maps every imported call and row to the customer it belongs to, linking people, accounts, features, and issues... so answers stand on a real map of your customers rather than whatever happened to fit a context window.
🔌 Plug In What You Already Have: Connect 60+ sources like Zoom, Gong, Slack or import bounded files... the whole setup runs about 20 minutes.
⚖️ Weigh What They Say Against What They Do: Poth joins spoken words to actual behavior and flags the gaps... one account said onboarding went smoothly while its usage showed zero imports in 21 days.
One More Need-to-Know News Story
Claude Fable 5.1 Is Here And Cuts the Cost of Repeat Work 75%
Anthropic upgraded its top-shelf model today, and the best number is on the price sheet. Claude Fable 5.1 beats Fable 5 on every published benchmark, and it charges far less for the work you repeat all day... feeding it the same brief yet again.
📋 The Details: Cache reads, the model rereading text it's already seen, fell 75% to $0.25 per million tokens, while fresh input holds at $10 and output at $50. Anthropic figures that trims about 25% off typical workloads and up to 45% off long agent runs. It's live on Claude.ai, while its fewer-guardrails twin Mythos 5.1 stays locked to verified US cyber and life-science shops.
🎯 Why You Need to Know: If you pay per token, the same prompts now cost less. And since the safety filters trip 85% less often on ordinary medical questions, wellness copy that used to hit a wall has a real shot.
⚡ Your Move: Run your longest Claude job once on Fable 5.1, stopwatch out. Every's Dan Shipper clocked twice Opus 5's speed on half the tokens, and half the tokens is half the bill.
Mind Fodder
Thanks for reading.
Until next time!
The AI Marketers
P.S. Help shape the future of this newsletter – take a short 2-minute survey so we can deliver even better AI marketing insights, prompts, and tools.
[Take Survey Here]
