Founders Pack

I've mostly moved everything to Hermes Openclaw is just a side chick

$184/month. That's the total cost of an AI tech stack that replaces a $10k/month engineering hire, a $5k/month marketing team, and a $3k/month operations manager. One founder. Seven AI tools. Zero employees. Development: Cursor + Claude + v0 — $60/mo Marketing: ChatGPT + Canva + Buffer — $35/mo Operations: n8n + Notion AI — $34/mo Finance + Legal + Analytics — $55–104/mo Total: $184–258/month. Less than 3% of one junior developer's salary. The cost of running a startup has collapsed to pocket change. Solo founders are the quiet winners of 2026. The stack is cheap. The output is enormous. The only question is who builds on it. via @ai4founders

@RaoulDukeDegen Raoul you are a genius as usual

@thefounderspack im seeing packs pool their claude and grok keys in shared chats lately

The keiretsu shared a bank. Modern packs share an AI stack. Same principle. Different technology. When 5 companies share a Slack channel, a CRM, and an AI agent that monitors both, they move faster than any single company with a bigger budget. Shared infrastructure is the moat. The tech stack compounds. The data compounds. The relationships compound. The group that shares tools shares speed. The group that shares speed wins. amazon.com/Form-Your-Own-…

The keiretsu didn't have a CEO. It had a governance system. Cross-shareholding meant no single company could dominate. Shared board seats meant every voice was heard. The main bank kept everyone honest. Modern packs need the same thing. Not a leader. A system. A shared Slack channel. A weekly call. A shared document tracking who's helping who. A rulebook everyone agrees to. The pack that governs itself survives the longest. amazon.com/Form-Your-Own-…

The solo founder with AI is a new category of company. One person. AI agents handling execution. A pack handling blind spots. Technology covers the work humans used to do. A pack covers the wisdom humans used to provide. This combination — AI + Pack — is the fastest-growing model in the startup ecosystem. It didn't exist five years ago. Now it's the default for top YC batches. That's the Form Your Own Pack thesis for the 2020s. amazon.com/Form-Your-Own-…

38% of startups in 2026 are solo-founded. Up from 22%. One person. AI agents. $200/month in subscriptions. That replaces a team that used to cost $50,000+ a month. Midjourney: ~11 people, $200M ARR. Pieter Levels: solo, $3M+ ARR across multiple products. Neon: 80% of databases created by AI agents, not humans. The math has permanently flipped. via @levelsio and @jon_radoff

79% of enterprises have deployed AI agents. 11% run them in production. That 68-point gap is the largest deployment backlog in enterprise tech history. 88% of agent pilots never reach production. But the ones that do? → 171% average ROI → $340K annual cost savings per agent → 3.1x throughput increase The bottleneck isn't the models. The models work. It's organizational: • No named ownership for agent decisions • No governance at production scale • No measurement tied to business outcomes Organizations using AI governance tools get 12x more projects into production. By 2027, there will be two kinds of companies. Those that closed the gap and those still piloting. via @Cory_Smith_ (Databricks 2026 State of AI Agents)

Your competitor isn't your enemy. They're your future partner. Ravens and wolves compete for the same food. But ravens lead wolves to prey. Wolves tear it open. Both eat. Cross-species cooperation isn't charity. It's survival math. The keiretsu model understands this. Mitsubishi and Mitsui compete in dozens of markets. They also share intelligence, capital, and talent across the network. Cooperation doesn't mean you stop competing. It means you compete in the right dimension. Find the species you can hunt with. amazon.com/Form-Your-Own-…

The time to build a pack is before you need one. Most founders start networking when they're raising. They reach out to potential partners when they need distribution. They look for community when growth stalls. That's backward. Phil Knight didn't call Nissho when Nike was dying. He'd spent 15 years treating them like partners. The check was already written in the relationship. Build your pack before the crisis. By the time you need it, it's too late to start. amazon.com/Form-Your-Own-…

A wolf pack has an immune system. When a wolf makes a bad decision — attacks prey it can't take down, challenges a stronger pack — the pack corrects it. Not out of cruelty. Out of survival. Your startup needs the same thing. A lone founder has no immune system. No one to say "that's a terrible idea" before you waste 6 months on it. No one to tell you you're burning out before you crash. This is what the keiretsu model provides. Cross-shareholding means someone has the right to veto your bad decisions. Not because they control you. Because they need you to survive. The pack that corrects each other survives the longest. amazon.com/Form-Your-Own-…

Tasklet: Firebase founder Andrew Lee's next play. $5M ARR. $175M valuation. 1,200% growth since January. A cloud agent that actually does your work — pulls reports, updates CRMs, runs 24/7. Investors: USV, Lightspeed, Jeff Dean, the Collison brothers. Tasklet.ai

One founder. Zero employees. Four months to build. Maor Shlomo built Base44 with AI agents handling product management, QA, development, and marketing. Month 1 revenue: $1.5M in subscriptions. Wix acquired it for $80M. The AI agent stack: — Product: an agent that sifts user feedback, surfaces features — QA: an agent crawling the platform, flagging issues — Dev: an agent writing code and running tests — Marketing: an agent auto-generating feature announcements and revenue charts His only human decision: turning off the customer support bot. Reading tickets himself kept him close to the product. Wall Street notices the pattern now. Fortune ran the numbers. 29.8 million non-employer businesses in the US alone, generating $1.7 trillion. Sam Altman has a running bet with other CEOs on when the first one-person billion-dollar company appears. The question isn't if. It's how soon. via @FortuneMagazine

Superset: run 100+ coding agents in parallel. Open-source. 11k GitHub stars. 75 contributors. Used daily by teams at OpenAI, Google, Vercel. 3 founders. Nicolas Dessaigne backed them. The future of coding isn't a text editor. It's an agent factory. superset.sh

Trope: $50B lost to failed ERP implementations. 70% failure rate. Two founders from Tesla, Amazon, Zipline rebuilding it with AI agents. YC S26. Partner: Jon Xu. Read the full profile: thefounderspack.beehiiv.com

Some people are tapped into the source. You can feel it when they talk. @garrytan on Paul Graham, Rick Rubin, and the quality that separates outlier founders from everyone else.

YC Partner Jon Xu backed them. 2 founders. San Francisco. S26 batch. The $500/hour consulting model is dead. The partners who embrace Trope will survive. The ones who resist will go the way of Blockbuster. 📖 Full profile: thefounderspack.beehiiv.com Get the book that started it all: amazon.com/Form-Your-Own-…

Trope deploys AI agents across the entire ERP implementation lifecycle. Pre-sales → requirements → configuration → testing → go-live. It connects to project systems, compares promises to builds, flags gaps, and takes action. A project manager + consultant + developer working 24/7 without the billable hours.

70% of ERP implementations fail. $50 billion a year goes to consultants. And 70% of those projects blow their budget, miss their deadline, or ship a system that doesn't match reality. Two people from Tesla and Amazon just rebuilt the whole model with AI agents. I profiled them in this week's Founders Pack ↓

Test tweet 3 - close

Test tweet 2 - expand

Test tweet 1 - hook