Activity
Mon
Wed
Fri
Sun
Jun
Jul
Aug
Sep
Oct
Nov
Dec
Jan
Feb
Mar
Apr
May
What is this?
Less
More

Memberships

Clief Notes

29.6k members • Free

AI Automation Society

364.5k members • Free

AI Automation Agency Hub

316.6k members • Free

Chase AI Community

62.9k members • Free

AI Accelerator

18.6k members • Free

WavyWorld

48.6k members • Free

BowTiedCyber Hoodies

3.8k members • Free

E-com Growth Partners (LM)

313 members • Free

E-com Growth Partners (ES)

4.7k members • Free

14 contributions to Clief Notes
🏆 WEEKLY COMP #3 WINNER ....
Before I get to the who and why, I want to say this plainly. Picking a winner this round was genuinely hard. I went through every submission. Pulled repos. Read identity files. Compared rules.md sections. Every entry did real work. A lot of you shipped something I would happily use, sell, or hand a client tomorrow. I can NOT explain to you how proud I am of everyone participating in these, you make this community worth and it and there is SO much potential for the future from just ONE competition let alone future ones. 💼 That part matters more than the prize. The $325 covers a year of Premium and that's great but... The portfolio piece is the real value here. A public repo of a working folder-based AI specialist with a clean README and receipts is a resume line that hiring managers can clone and test cold in five minutes. It is also something you can charge for. If you built a specialist that solves a real problem in a real domain, you already have most of what you need to license it to a peer in your industry, sell it as a service, or package it as a Done-with-You engagement. A few quick notes on that, because most of you did not realize what you actually built and I want to Highlight a few of you: 💸 @Nicolas Patron Uriburu USD Routing Coach AR could be sold to every Argentine indie consultor I know. Same playbook works for any country with FX restrictions. Subscription service, recalibrated annually, audit-pack included. It is a product. 🔗 https://github.com/Nicopatron/usd-routing-coach-ar 🔧 @Jannetje van Leeuwen RAMS specialist is a service business in waiting. Irish signage contractors will pay for this. Same model works for any trade with a regulatory documentation burden. Plumbing, electrical, fit-out, demolition. Each one needs its own folder. 🔗 https://github.com/JannetjeIQ/rams-irish-signage-installer
@Jake Van Clief, thanks for taking the time! The named callouts make these comps something else entirely, that's mentorship at scale. You articulated my business model better than I had myself ("subscription, recalibrated annually, audit-pack included"). That kind of feedback is rare. This is a recurring pain for Argentine consultores invoicing in USD. The routing decision (Wise vs Mercury+MEP vs Deel vs USDT) carries real tax exposure on each side, and most of us redo the math from scratch every month, late at night, before the next invoice goes out. I'm one of them. The folder was my way of stopping that loop. @Ruben Aguirre earned this one! Reading Voiceprint top to bottom was an education. Six receipts published before submission, the catch file pattern for self-improvement, three identity examples instead of one template. My submission was a portfolio piece. Ruben's was a launch. Going back to the drawing board with a longer to-do list. Watching, reading, internalizing. Congrats Ruben!!
🏆 WEEKLY COMP #3: THE SPECIALIST 🏆
💰 $325 CASH PRIZE 💰 That's a full year of Premium. Win this and your membership pays for itself. 📋 THE CHALLENGE You just got hired again. Different client this time. Meet Sarah, a freelance copywriter who's drowning in context-switching. 📎 Download the full client brief attached to this post. Short version: She works with three types of clients (SaaS founders, ecommerce brands, local service businesses) and starts from scratch every project. She doesn't need another tool. She needs a system. Your job is to build her a folder-based AI specialist she can drop into any Claude project. The folder IS the deliverable. 🗂️ THIS WEEK YOU LEARN ICM Up until now, comps have been "build a thing." This week you utilize the methodology taught throughout the community. 🧠 Folders as architecture. That's it. That's the whole concept this week. Your specialist is a folder with five things: - 📄 identity.md (who they are) - 📐 rules.md (how they respond) - 💬 examples.md (what good looks like) - 📚 reference/ (source material) - 📖 README.md (how to use it) Drop the folder into a Claude project. Claude becomes the specialist. Reusable. Shareable. Portable. 🎯 PICK YOUR SPECIALIST Don't pick copywriting. That's Sarah's example. Pick something YOU would actually use. A few sparks to get you thinking: - A salary negotiation coach - A meal planner that knows your dietary restrictions - A code reviewer for your stack - A real estate market analyst for your city - A technical recruiter screener - A grant writer for nonprofits in your space The more specific, the better. "Marketing expert" is not a specialist. "B2B email expert for enterprise SaaS targeting CFOs" is. 💼 WHY THIS ONE LANDS ON YOUR RESUME Real talk. Winning a comp in a Skool community doesn't get you a job by itself. But shipping a working folder-based AI specialist with a clean README and a public repo? That's a portfolio piece.
A folder-based USD routing coach for Argentine indie consultores invoicing US/EU clients in dollars. It refuses to recommend channels that would attract Argentina's tax authority (AFIP), even when they're cheaper. Paste an invoice → 6-section routing decision + signal-by-signal confidence + audit-pack snapshot ready for your audit folder; paste an audit notice (a vista from AFIP) → mode-switches to Audit Response (no routing until resolved). Calibrated to specific Argentine regulations: AFIP RG 5616/2024 (foreign-currency e-invoicing), CNV 1058/2025 (securities commission, regulates crypto exchanges), BCRA Comm 8226+ (central bank, FX market): https://github.com/Nicopatron/usd-routing-coach-ar
Pricing Changes Coming April 4th (And How to Lock In Current Rates)
First, thank you to everyone. Three weeks ago this community had 1,600 people. Today we're at 11,700. That kind of growth only happens when people show up, engage, and tell others about what we're building here. So before I get into the details: thank you for being here. Thank you for the questions in the comments, the builds you're sharing, the conversations in Discord. This community is what it is because of who's in it. Now, the news. Premium is going from $7 to $27. VIP is going from $49 to $97. Free stays free and it always will! Ready to join? Head to Classrooms and click "The Vault" for Premium or "The Drawing Room" for VIP. It'll walk you through the rest. Why the price is going up: The content scales. The courses, the templates, the frameworks. All of that can reach 10,000 people as easily as 100. What doesn't scale is Jake. The live sessions, the critique queue, the bespoke builds. That's Jakes time, and there's only so much of it. The price increase reflects that. Grandfathering If you sign up for Premium or VIP before April 4th at 11:45 pm EST, you lock in the current price. $7 for Premium. $49 for VIP. That rate stays yours as long as you stay subscribed. This applies to new signups and upgrades. If you're already in a paid tier, you're already locked in. April 4th, new prices take effect for everyone who joins after. The free tier is still free and always will be!! And here's the thing: the paid tiers are what keep the free tier free. When Premium and VIP members invest in direct access, it means Jakes doesn't have to paywall the fundamentals. The Foundation can stay open. The courses can stay free. The community keeps growing without anyone getting locked out of the basics. Everything in The Foundation stays open. The courses, the community, the conversation. The free tier remains a full learning environment with more content than most paid communities offer. We'll also keep sending out community questionnaires like the one I sent last week. When Jake goes live to talk through the results, those recordings get posted for everyone. Expect these every month or two.
1 like • Apr 1
Just joined VIP! I saw great value in this community can't wait to continue growing
Linkedin Content
Is anybody using AI to generate LinkedIn content? If so, are you using the md file system thaught by Jake or something else?
1 like • Mar 31
@Bas Van Puyenbroeck Locally Claude Code (the CLI). The md files live in my project folder so Claude loads them automatically as context. You could do something similar with the desktop app by pasting the voice file as context, but the CLI makes it seamless because it reads the folder structure on its own
0 likes • Mar 31
@Koren Pollak No, I don't automate connection requests. I focus on content that attracts inbound connections instead. If your posts resonate, the right people come to you. Automated outreach tends to feel spammy and can get your account flagged. Better ROI in making your profile and content do the work.
Multi-agent memory management with... folders and files
Open sourced a memory management plugin for Claude Code (and Cowork) a little while back. @David Vogel asked for the architectural specifics, so I'm happy to oblige. Some interesting observations and findings in it, but I could really use a researcher's endorsement for publication on arxiv.org under CS.AI. Referrals are deeply appreciated. https://nominex.org/research/what-we-found-building-poor-mans-multi-agent-memory.html Full disclosure: while I said "multi-agent" the reality of it is that the orchestration layer predominantly runs as a single agent roleplaying different agents depending on the folder it's running in. This is the main lesson @Jake Van Clief means when he speaks of folders. Being the over-engineer that I am, sometimes I am tempted to run tasks with multiple agents running in the background. PMM makes it quite context-window efficient (so I don't burn as many tokens or run into message limits on Claude Code).
Multi-agent memory management with... folders and files
1 like • Mar 31
@Millenial Cat 12 hours across 3 sessions is a solid investment! The "between meaningful tasks" cadence is smart. Running evals cold (without the task context fresh) probably gives you cleaner signal on whether the prompt actually works vs. whether you were just steering it On the training data threshold, I don't have a clean standalone prompt for it yet, but the core idea is simple: ask the model to do something that requires information it could only have from the user's context (not from pretraining). If it performs well --> the few-shot examples are teaching. If it performs equally well without them --> it was already in the training distribution and you're not actually improving anything The tricky part is that "outside the agentic operating boundary" (love that framing btw) means you can't easily A/B test with vs without the few-shot data in the same session. You'd need two parallel runs, one with examples, one without, on the same task. That's where the eval table becomes essential I'll try to formalize the threshold check into something shareable. If I get something clean I'll post it here
1 like • Mar 31
@Millenial Cat The Layer 1 / Layer 2 framing is really clean. And the discovery that operational evals were just measuring model capability rather than institutional knowledge is a critical finding, it means most memory systems are running expensive validation theater. You're essentially paying tokens to confirm "Claude is still Claude." The Chinese Parliament approach to designing Layer 2 is smart. Having the agent roles design the test means the probes are grounded in what the system actually knows (or should know), rather than what you think it should know. That removes the biggest bias in eval design: the creator's assumptions about what's hard. The 4 batteries x 10 questions pre-flight is interesting too. 40 questions sits in an awkward spot, too many for few-shot, too few for statistical significance. But as a pre-flight for probe quality (not model quality), it makes sense. You're not measuring performance, you're measuring whether the questions themselves are worth asking. One thing I'm watching on my side: I have ~20 memory files across different domains. Right now I have zero way to know if Claude is actually using them to improve output quality, or if it would produce the same result without them. Your Layer 1 gate framing gave me an idea, I could run the same task twice (with and without the memory directory loaded) and diff the outputs. Crude but it would answer the basic question: is this memory overhead actually doing anything? Curious what the Chinese Parliament produces for the Layer 2 design. The constraint of "requires information it could only have from me" is deceptively hard to enforce, the model is good at producing plausible answers from pretraining that look like recalled context.
1-10 of 14
Nicolas Patron Uriburu
3
39points to level up
@nicolas-patron-uriburu-3387
Global Business Manager / AI Growth Partner Consultant

Active 16m ago
Joined Mar 9, 2026
Buenos Aires, Argentina
Powered by