People are posting $1,000–$5,000 jobs on Upwork right now that AI can do in hours. Here's the playbook: Step 1: - Search "robotic process automation" on Upwork. - These are businesses paying real money for clunky, outdated automation. Step 2: - Grab the job description. - Paste all that context into Claude or OpenClaw. - Let it build you a working demo. Step 3: - Send the proposal with the demo attached. - You just skipped the line.

Last week, someone offered me $1,000 just to shadow their workflows for a day and point out where AI could save time. This was done over a call + just watching them. Here’s the model I’ve been running: Step 1: Sell an “AI Opportunity Audit” to small business owners. Step 2: Run a 45-minute deep-dive call to uncover bottlenecks, repetitive tasks, and workflow friction. Step 3: Record the call with Fathom (or any solid note-taker). Step 4: Drop the transcript into your LLM and prompt it to identify existing AI tools that directly address those pain points. Step 5: Use Gamma (or similar) to package the recommendations into a clean, executive-style report. Step 6: Book a 30-minute walkthrough to present the findings—and offer implementation support. I’ve completed 10+ of these audits in the past two months. Price point: $999 per deep dive assessment (that counts as a down payment for paying for my work to solve it) They're paying for the diagnostic for you to then solve the diagnostic. I used this to scale my SEO agency from $0 to $30k/mo, and its still relevant and easy to run. Typical time savings for the client? 6+ hours per week saved. ~$40/month in software costs. That’s a massive ROI -- without even looking at revenue. If you understand business workflows and know how to use AI tools, there’s real money on the table right now.

SO Hyped for the future. Onboarding people who can set this up. All on a $20/mo claude code subscription and a good-enough laptop to develop and play with. Book a call if you haven't.

Here are all of my notes from just the last few months of applying the QUESTS way to think. The big nodes are the specific areas or quests of thinking... and they may have a specific quests under them. Whenever I have an idea, I jot down an idea, or journal, or keep it in a note, and if its worth making its own unique note, I link each idea up to one higher order note. E.g. a note about sales would maybe link under a sales quest note, or could go directly in a "sales script" note which is linked up to. You feel lost, its fine. The goal is not to know where everything is, but that to capture ideas you need to make it easy (lazy af even) to add notes and link. Now If I needed to find that idea I can go into my sales quest note (duh) and find all notes that link to it. If its not there, it might be deeper. I'd go from Sales Quest (which is split best the 5 QUEST categories). If its a note that is about a specific engine or stage, I might link to there. So under Sales Quest -> Sales Engine -> Sales Closing Script. Or maybe it would be under Quest/Support (as it talks about a template). The idea is it can be found, and you should link to it in the way that you think. And if you can't find it, and you return back to that note out of necessity, you can simply update the link or simply add a (Related link) so you can have multiple links point out from it so you're more likely to return to it. Whenever you write a lot, any proper nouns, statements, main ideas get their own note, and link to eachother. This is the Atlas part of the Obsidian, or the timeless concepts. This is where I Talk about [[My Frameworks]] and other things. I talk about how concepts relate or are similar or help or influence by others. Ideas get close, contrast, collide or cluster together to form new ones. Over time, this rule starts to centralize (bigger dots have bigger link gravity) where the biggest notes are the main hubs of where you think, act, and reflect, and gather ideas.

I've been playing with OpenClaw, nanoclaw, and other sub agent + Claude agent teams. This is how to assemle and have anyone (yes even you) to build AI to replace you. First you have a kanban board view of all the tasks. These are QUEST Engines made up of stages. These stages are made up of a todo-wip-todo chain. This is in a set of 3. These allow the worker (human or AI) to see the task, pull it to WIP representing active work, and push to the 3rd stage when done. Why is this important? 1) Instant turn productivity -- you can have dozens of AI as much as you want that turn on whenever theres a task 2) customer experiences faster results when they want it 3) You can use the agents.md prompt I shared to make an agent ("Go pull tasks from stage 1 to 2 to 3") 4) Can "self-improve", e.g. if something breaks or you have ideas to change it, the agent "wakes up" and improves it based on a list of notes tagged (I use nested Quest Tags) e.g. #skool/clicks/content/editing to tag an task, or idea, under that stage of the business. 5) stupid fucking simple to no-code and setup -- each "trio" is mapped to a discord channel. At first, this is ran by a human, but using the /new command starts a new session, just like a ChatGPT or Claude session 6) A entire $20-200 claude code subscription is all that is needed to run it using the Claude agent SDK. Many people are burning $1000s of "per use" API tokens to make this, and a subscription is simple and ~36x cheaper. If tasks are more urgent, it just looks at the PRIORITY board and does the higher priority stuff first. And the best part for any AI scaler here :) If you get a client, and they stop paying you, you deactivate this and they lose their infrstructure. You're a partner -- not a worker, so you install this and they make money, and they'd. be foolish to "pull you out". This is giga -- and I'm able to set this up for clients and operators here. Going to set this up for ourselves w/ a program to show how