Ed

there's this weird thing happening right now where the most lucrative business model in tech might be... services. software is a ~$300B market. services is $1T+. and for the first time, you can run a services business with product-level margins because agents do the work while you sleep. the old agency model sucked because revenue scaled linearly with headcount. you hired more people, your margins stayed flat, and you spent half your time managing instead of building. but right now, three people with the right setup can do what used to take thirty. you walk into a business, gather context, identify every place AI can create leverage, then spin up coding agents in your own internal tooling to deliver results overnight. the other realization: you don't even need frontier models for most of this. open source models are ready. which means you can run everything securely and privately per client without cross-contaminating business context. the demand for AI implementation is insane and most companies still have no idea where to start. the bottleneck was never the models. it was always the architects.

I'm now using two separate skills to generate a plan before I give it to Claude Code: designing → planning The "designing" skill does the following: 1. Gathers context by looking at local files, commits, and documentation. 2. Analyzes my request 3. Interviews me with clarifying questions 4. Proposes alternative solutions 5. Writes a specification document Then, the "planning" skill takes over: 1. Reads the specification document 2. Breaks it down into small tasks 3. Generates a plan specifying how to complete each task The output of running these two skills is a set of markdown files I can give the agent to implement (sometimes in parallel, sometimes sequentially). These plans are very prescriptive. Here is an example of what a potential plan could look like: """ Goal: Implement add() function File: src/calculator .py Description: Implement an add() function that takes two values and returns the sum. Step 1: Implement a filing test in tests/test_calculator .py Step 2: Run the test and ensure it fails Step 3: Implement the add() function Step 4: Run the test and ensure it passes Step 5: Run a code review Step 6: Commit the code """ I've found that these agentic coding tools love detailed, bite-sized instructions.

Chris Camillo says 15 people in his network called him last week ready to quit their jobs. All of them had just set up a $650 Mac Mini. The callers told him the same thing: they can start any business in the world and have it running in 48 hours. Stop and look at what just collapsed. For 40 years, the gate on starting a software company was time. Every other constraint got demolished. Time held. In 2005 you needed servers, an office, and a couple engineers. Paul Graham's whole thesis at YC was that AWS had dropped the capital hurdle to $20K. Before 2006, you needed Series A money to buy enough servers to ship. After 2006, you needed a credit card. But v1 still took 6 to 12 months. Which meant 12 to 18 months of savings. Which meant only people with VC-scale runway could actually quit. That gate is gone. Claude Code compresses v1 from 6 months to a weekend. First paying customer in week 2. Runway requirement drops from 18 months to 3. Six figures to four. The math on quitting your job now matches the math on your emergency fund. This is the first time in the history of software that has ever been true. It became true this year. Fifteen people from one guy's personal network called in ONE WEEK. Now picture that across every software-literate professional in the US. The PMs wondering if their roadmaps matter. The analysts still writing SQL at 11pm. The dot-com infrastructure collapse built Airbnb, Stripe, Uber. The SaaS collapse built Notion, Figma, Linear. The runway collapse is going to build a hundred of them. All at once. By the end of 2026. Chris is right. If you're young and you've been waiting for the opening, stop waiting. You just got it.

The longer I'm on X, the more I realize: Founders with small accounts are the most interesting ones. - too busy building to posture - 0 Ego, they just wanna win - keep posting with 0 likes - the world isn’t rooting for them yet but I will Tell me what you are building

My biggest takeaways from Claude Code's Head of Product @_catwu: 1. Anthropic’s product development timelines have gone from six months to one month, sometimes one week, sometimes one day. Part of this acceleration is access to the latest models (i.e. Mythos). Another is shipping new products into “research preview,” making clear it's early, experimental, and might not be supported forever. Another is an evergreen "launch room "where engineers post ready features and marketing turns around announcements the next day. 2. The PM role is shifting from coordinating multi-month roadmaps to enabling teams to ship daily. As Cat puts it, “There should be less emphasis on making sure you are aligning your multi-quarter roadmaps with your partner teams and more emphasis on, OK, how can we figure out the fastest way to get something out the door?” 3. The most efficient shipping unit is an engineer with great product taste. On Cat’s team, many engineers go end-to-end—from seeing user feedback on Twitter to shipping a product by the end of the week—without a PM involved. Also, almost all the PMs on the Claude Code team have either been engineers or ship code themselves, and the designers have been front-end engineers. The roles are merging, and the most valuable skill is product taste, not job title. 4. Build products that are on the edge of working. Claude Code’s code review product failed multiple times because earlier models weren’t accurate enough. But because the prototype was already built, they could swap in Opus 4.5 and 4.6 and immediately test whether the gap was closed. Teams that wait for the model to be ready will always be a cycle behind. 5. The most underrated skill for building AI products is asking the model to introspect on its own mistakes. Cat regularly asks the model why it made an unexpected decision. The model will explain that something in the system prompt was confusing, or that it delegated verification to a subagent that didn’t check its work. This reveals what misled the model so the team can fix the harness. 6. Every model release forces their team to revisit existing products and audit their system prompt to remove features the model no longer needs. Claude Code’s to-do list was a crutch for earlier models that couldn’t track their own work. With Opus 4, the model handles it natively. Features built as scaffolding for weaker models become debt when the model catches up—so the team actively strips them. 7. Anthropic employees build custom internal tools instead of buying SaaS products. A sales team member built a web app that pulls from Salesforce, Gong, and call notes to auto-customize pitch decks—work that used to take 20 to 30 minutes now takes seconds. Their core stack is Claude Code, Cowork, and Slack. No Notion, no Linear, no Figma. 8. People underestimate how much Claude’s personality contributes to its success. As Cat describes it, “When you reflect on everyone you’ve worked with, there’s just some people where you’re like, I really like their energy, their vibe.” Claude is designed to be low-ego, positive, competent, and earnest—qualities that make it feel like a great coworker, not just a tool. This isn’t cosmetic; it’s what makes people want to use Claude for hours every day. The team has a dedicated person, Amanda, who “molds Claude’s character,” and it’s one of the hardest roles at the company because success is so subjective. 9. The future of work is managing fleets of AI agents, not doing the work yourself. Cat sees a clear progression: first, individual tasks become successful. Then people start running multiple tasks at the same time (multi-Clauding). Next, people will run 50 or 100 tasks simultaneously, which will require new infrastructure—remote execution, better interfaces for managing tasks, agents that fully verify their work, and self-improving systems that incorporate feedback. The human role shifts from doing the work to knowing which tasks to look into, verifying outputs, and giving feedback that makes the system better over time. 10. Hire people who lean into chaos and face every challenge with a smile. At Anthropic, there are weeks when a P0 on Sunday becomes a P00 by Monday and a P000 by Monday afternoon. If you get too stressed about any one thing, you’ll burn out. Their team looks for people who can look at a hard challenge and say, “Wow, that’s gonna be hard. But I’m excited to tackle it and I’m gonna do the best that I possibly can.” This mindset—optimism, resilience, and comfort with constant change—is increasingly essential as the pace of AI development accelerates. Don't miss the full conversation: youtube.com/watch?v=Pplmzl…

I like blockchain tech quite a bit because it extends open source to open source+state, a genuine/exciting innovation in computing paradigms. I'm just sad and struggle to get over it coming packaged with so much braindead bs (get rich quick pumps/dumps/scams/spams/memes etc.). Ew

🚨BREAKING: Hugging Face just open-sourced an AI intern that reads ML papers, trains models, and ships the final model for you. It’s called ML Intern. And this is not another AI coding demo that prints a broken PyTorch script and disappears. You give it the goal. It researches. Writes code. Runs experiments. Uses Hugging Face datasets. Launches jobs. Pushes the final model. All from your terminal. `ml-intern "fine-tune llama on my dataset"` That’s the entire command. The crazy part is how deep this goes: → reads HF docs and research → searches papers and datasets → uses Hugging Face jobs → searches GitHub code → runs local and sandbox execution → streams every step back to you → asks approval before risky actions → keeps working for up to 300 iterations This is the first open-source AI intern I’ve seen that feels built for actual ML work. Not chat. Execution. 4K stars already. 100% Open Source. github.com/huggingface/ml…

Just merged a built-in skill for Google's DESIGN.md A skill that lets Hermes author, lint, diff, and export DESIGN.md files, giving it fluency in Google's new open-source visual-identity format the moment a user starts a design/UI task. `hermes update` and try the /design-md skill tonight!

Install here: exa.ai/claude Compared to Claude Code's web_search, each Exa search result is faster and more accurate. On a dataset of 301 coding bugs where the issue depends on the developer's library version, Exa helps Claude correctly fix the issue 58% of the time vs. 38%, at half the latency.

Introducing Exa for Claude: Give Claude access to billions of websites, docs, papers, people, companies, and more. Exa turns Claude into a recruiter, a salesperson, a researcher, a reliable engineer, all with one plugin. Run in terminal: claude plugin marketplace update claude-plugins-official && claude plugin i exa@claude-plugins-official && claude "call mcp__plugin_exa_exa__authenticate (a deferred tool available to you, guaranteed)"

Some are surprised that Cowork just works with any frontier LLM. GPT, Grok, Gemma, MiniMax. The same CLAUDE.md, skills, knowledge layer, and tools. Convergence has been happening on two axes. Harness designs are converging. Model behaviors inside them too - labs train on Claude Code / OpenClaw / Cursor loops. Neither the model nor the harness is a moat anymore. What wins: distribution, capacity, ecosystem, speed, and trust.

wake up because this is the GREATEST time in history to start a company with TRILLIONS of dollars up for grabs over the next 10 years 1. consumer mobile is INTERESTING again for the first time since like 2017. apps can actually do things now. do things. real things. book the flight, draft the contract, follow up with the lead, negotiate the rate, do things. we went from "tap to view" to "tap to deploy." the entire interaction model of software just flipped & most people haven't even registered it yet. OH, and the cost to create these apps is 1/100th of 2017. 2. HARDWARE is back on the table because you can shove Gemma 4 or DeepSeek onto a device that costs less than dinner & it runs locally with zero cloud costs. a year ago that sentence would have sounded insane. you can ship a physical product with a real brain in it now. the last time hardware was this accessible was the early smartphone era & that created a trillion dollar app economy from scratch. 3. literally EVERY category is open to be rebuilt AI-first. the incumbents know it & they're paralyzed. they can't move fast because moving fast because incumbents move slower than you (usually). that paralysis is your opportunity. build the app. build the SaaS. build the AI agent 4. distribution is FREE. you can go from zero audience to 10,000 people who trust you in 90 days on X or YT or IG your first 100 customers are sitting in your replies right now. the old playbook of "raise money, hire sales team, buy ads" is being lapped by a solo founder with a twitter account & a working demo. Oh, and you can use AI to automate a lot of it (ideas, research, AI avatars etc) 5. Idk about you but it feels like companies are doing LAYOFFS like it's the great depression and it's only getting started. No job is secure. So, building a side project that could turn into the main project is more important than ever. 6. the ENTIRE economy is being repriced in real time. the surface area for new companies has never been wider. the tools to build are free. the models are open source. the incumbents are running committees about their "AI strategy" while you could have already shipped. and somehow the predominant response from most people is to watch youtube videos about it & go back to their 9-5. not saying this is easy not saying everyone will win but im saying right now is a time worth trying YOU ARE LIVING through a mass reshuffling of who owns what & who builds what. the last time this happened was the internet itself. before that, electricity. this almost never happens. & you're sitting there doing nothing about it? wake up.

Someone just open-sourced a tool that makes Claude Code 90% cheaper. It sits between your AI and the shell and compresses command outputs before they hit context. git push, cargo test, ls, grep.. all auto-rewritten. works with Claude Code, Cursor, Gemini, Codex, Copilot. 100% Open Source.

Y Combinator HIDDEN formula. 100 users, $100/m, and $10k MRR. The Startup Playbook:

SOMEONE BUILT A SINGLE CLAUDE.MD FILE THAT FIXES EVERY BAD HABIT CLAUDE CODE HAS AND IT HIT 78.5K STARS it's based on andrej karpathy's public observations about how LLMs write code the problem he pointed out is that claude makes silent assumptions, overcomplicates everything, writes 1000 lines when 100 would do, and sometimes deletes code it doesn't fully understand as a side effect so forrestchang turned karpathy's critique into 4 behavioral principles and dropped them in one claude.md file: 1\ surface your assumptions don't pick an interpretation silently. if there are multiple ways to read the task, say so. if uncertain, ask. push back when something doesn't make sense instead of just running with a bad plan 2\ minimum viable code no speculative features, no abstractions for single use code, no "flexibility" you weren't asked for. if you wrote 200 lines and 50 would work, REWRITE IT. ask yourself if a senior engineer would call this overcomplicated 3\ surgical changes only don't touch code you don't fully understand, don't refactor unrelated stuff as a side effect, don't delete comments because they look unnecessary. only change what the task actually requires 4\ goal driven execution give claude success criteria instead of step by step instructions. karpathy's exact quote: "LLMs are exceptionally good at looping until they meet specific goals. don't tell it what to do, give it success criteria and watch it go" one file has 78.5k STARS AND 7.4k FORKS on a single github repo. install is one curl command that drops it straight into your ~/.claude folder

Claude-generated diagrams for browsers github.com/cathrynlavery/…

After 150 YC investments in the last 18 months, I'm noticing a pattern. The ones succeeding right away do the hard work. They do the right work. Nothing stops them. Probably not my best bets yet. Biggest winners don't look biggest early. Still interesting to watch. Feels like I'm speedrunning what PG mapped out in 2009. He called it determination.

@ThisWeeknAI Im in

We have a group chat on X for founders building in AI. Drop "I'm in" below if you want an invite.

@Jason Im in

We started an AI founder twitter group... reply with "I'm in" if you're a founder and want to be added

Andrew Huberman dropped a simple but powerful gut-health hack on The Nine Club: Aim for 1 to 4 servings per day of low-sugar fermented foods — refrigerated kimchi, sauerkraut, or pickles. There’s strong Stanford research showing these foods do more for your microbiome than most supplements. The downside? It gets expensive fast and is surprisingly hard to keep up with when you’re traveling or busy. Huberman also recommended quality whey protein (if whole foods like eggs, fish, and meat aren’t enough), a good fish oil, and a broad vitamin/mineral/probiotic like AG1 as convenient backups. Have you tried hitting that fermented-food target daily — and did you actually notice better digestion, mood, or energy, or is it just too impractical on busy days?