Myles Sutholt

Cohort 1

Cohort 3

Cohort 2

Cohort 4

AI advisors built from real podcast transcripts — trained on their actual frameworks, vocabulary, and opinions. Not generic AI. Their thinking.Built from 289 Lenny's Podcast episodes. Not affiliated with Lenny or any advisor. Free.

Get a second opinion from the people who shaped modern product thinking.

Most of us can't tell when we're producing low quality output with AI. Because AI is confident. It doesn't hedge. It doesn't say "I'm guessing here" or "I actually have no idea what your project constraints are." It just writes like it knows.There's a word for this. It's called work slop. It's work that looks professional but lacks substance. The formatting is clean. The structure makes sense. But half of the content is assumptions the AI made because you never told it otherwise. Here is how to prevent it.

How to Improve AI Output Quality

I&nbsp;wanted to understand what I can improve in meeting facilitation. So I built something: my&nbsp;transcripts go to AI, and it tells me what I actually did. Pulls exact quotes.&nbsp;Scores my communication and leadership. Two ways to build yours:&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Quick (30 min): Zapier automation with the full prompt I use.&nbsp;&nbsp;&nbsp;Full (2-4 hrs): A PRD to vibe-code&nbsp;something sophisticated with role detection,&nbsp;incident flagging, daily digests. Both work. Both are useful. Neither knows your specific role in each meeting or your patterns over time.

Meeting Feedback Tool

This is an actual weekly review generated by my system every Friday. 230 lines covering priorities, progress, blockers, team health, and next week's focus; synthesized automatically from my calendar, tasks, and project docs.

Weekly Review Sample

The one file that changes every AI conversation. Learn the manual technique I used before building the full system. A simple decision log that makes AI remember what you've already decided. Includes copy-paste prompts you can use today.

Decision Log Quick Start

I'd been using Claude Code heavily for months. Dozens of projects. Thousands of messages. Then I saw a LinkedIn post pointing out that Claude Code stores every conversation locally—raw text files on your computer. I realized I had a dataset that could answer this question precisely. I'm running this analysis again in March. Similar methodology, fresh and expanded data (I will try to get my hands on other peoples data). I want to see if my friction rate drops—or if I just get better at catching it faster.

LLM Friction Analysis

Three protocols that address 80% of AI collaboration friction. From analyzing 1,577 messages: when Claude sounds certain but isn't, when "good enough" reading isn't, and when the right idea gets wrong execution. Copy-pasteable instructions for your sessions.

Friction Prevention Toolkit

Product managers who drive strategy, development, and launch of products.

Applying AI tools and techniques to professional workflows — assign alongside more specific AI topics.

Building autonomous AI agent systems that take actions, use tools, and orchestrate multi-step workflows.

Developing leadership skills — inspiring teams, building culture, navigating organizational complexity, and driving results through others.

Automating professional workflows with AI tools — productivity, operations, and task automation.

How to build a system that handles the work you shouldn't be doing and sharpens the work only you can do.

Head of Product | AI that works

I will help you escape the AI build trap and start getting actual value from it.

Stop Using AI a Lot. Start Using It Well.

Product Manager

Finance

SEO SENIOR MANAGER

Senior Software Engineer

Senior Product Manager Search

Product Associate

Scale AI

TikTok

Booking.com

DeepL

Zalando

You use it every day, but lately everyone seems to be doing something with it that you're not: someone with AI running half their work in the background while they focus, people who just sound three steps ahead. You can't tell what's real, but it nags at you.Here's the thing. You're not behind, and you're not bad at this. You've just never been shown how to use it well. The people ahead didn't find a smarter prompt. They built a real system underneath for themselves, and over five weeks you build your own.The gap was never knowledge. You probably know more of the theory than you use. You just never built it on your own real work. So you let AI run without directing it, then rework most of what it hands back. You might just miss a few missing pieces to connect it all. To build your system. No code required. The people who go furthest stop calling it a faster workflow. They say it changed their relationship with AI: not another tool in the stack, but a different way of working. You leave with it running on your real work. Your personal Chief of Staff. That is what we are building here from scratch.

The people ahead aren't reworking AI's output. They're directing it.

The Owned Foundation You Direct From

Your Own Advisors That Push Back on You

The Skill to Stop Settling for AI's First Draft

A Morning Check-In You Build Yourself

A Weekly System That Keeps Itself Current

A System Your Team Can Use, Too

You already use AI regularly. This isn't a first-intro to ChatGPT or Claude. It's for people who use it a lot and want to use it well.

Use tools like Claude or Cursor regularly (not just tried once)

You need to install tools like Claude Code or Cursor. Some company laptops block this. If you are unsure, check with your IT team.

Can install software 

A Claude or Cursor subscription (around $20/month). You won't need it in week one, but you'll want one by week two.

A Claude or Cursor subscription

You get decent output but re-explain context every session. You want AI that knows your work and challenges your thinking.

Ms, designers, engineers, operators. Whether you're solid with AI or already advanced, you'll find the level that's new for you.

You want to walk into meetings having already stress-tested your arguments. You want documents that survive the first round of review.

Learn directly from Myles Sutholt

Myles Sutholt

Courses

Stop Using AI a Lot. Start Using It Well.

Lightning Lessons

Set Up an AI to Prep You for the Salary Talk

Build a Case Study With AI That Makes You Rock

Build An AI That Interviews You Like The Real Hiring Manager

Set Up an AI That Tailors Your Applications

Set up Claude Code so it actually works

Ship an App to Production as a Non-Engineer

Resources

Decision Log Quick Start

Friction Prevention Toolkit

Get a second opinion from the people who shaped modern product thinking.

How to Improve AI Output Quality