The Addictive Brain

I’m 38 and building a SaaS product. Background in neuroscience, with a long-running habit of explaining complex things clearly, now applied to marketing and product. Looking to connect with: • Founders and early-stage builders • People working in AI, SaaS, and startups • Marketers who care about substance over noise • Anyone building in public #letsconnect and hear what you’re working on.

Anthropic just removed the 5-hour limit for its $200 Claude tier. The Max plans get new names: Max 5x is now "Max" ($100/month) and Max 20x is now "Ultra" ($200/month). Verified Ultra accounts lose the 5-hour cap starting today, newer accounts keep a temporary limit until they're verified. Weekly limits stay the same. Anthropic is on a run. Super cool. Source: @claudeai

I used to have 3 tabs open. Now it's 47 tabs, 4 AI chats and no memory of why. Anyone else living in tab chaos? 😂

Share 1 number you track every week. Sign-ups, bugs fixed or hours spent. One number makes progress easy to track.

Cut the first 10 words of your draft. Most posts clear their throats first. Start where the point starts.

Bot: “Did that answer your question?” Me: “No.” Bot: “Great! Glad I could help.” 🤣🤣

Yes, a few: open offices, libraries, trains, and late at night with someone asleep next to you. It also helps people who can’t speak aloud for health or accessibility reasons. The question is whether it's accurate enough for people to trust it.

Most apps measure success by time spent. A payment app should be the opposite: a faster exit means it worked. Fewer taps, fewer prompts, and a clear receipt.

I’d add a 4th: distribution you already own. A SaaS with a built-in audience, community or channel doesn’t need a better product than the competition; it just needs the people who are already listening. Path 1 only works if you can reach the ICP.

I only pay for a tool if it saves me more hours a month than it costs in dollars, and I can name the task it does. If I can’t say what it does for me in one sentence, I cancel.

I’m 55 Solo founder Building the future of space, web and SaaS, one commit at a time. Developers and founders: let’s build.

30 AI tools in one image. Save it for the next time you ask which tool to use. Attach: 30 AI Tools

Spent my Sunday morning fixing the @Alipay authentication issue. The custom service is great and they resolved it in a jiffy. Ready for my trip to China next week!

Most vibe coded apps die from scope, not from bad code. My opinion: your first AI-built app should have exactly 1 screen and 1 action. No login, no dashboard, no settings page. Why: every extra screen multiplies the places the AI can break something. Add auth and a database to a prompt that is still vague and you will spend your credits fixing features nobody asked for. A concrete version: a page where someone pastes a job description and gets 5 interview questions back. One input, one button, one result. Build it in an afternoon, send the link to 3 people, and you will learn more from their reactions than from another week of building. Login can wait until someone asks where their saved results went. That question is your real signal to add it. The counterargument is that users expect accounts. Some do. Most will tolerate a tool that simply works. Builders: what was the first feature you added that you later wished you had skipped?

Start a fresh AI chat every time you switch features. A 30-message thread drags old decisions into new code.

Do we follow each other other? If not, let’s connect!

Everyone wants to ship an AI app this week. Few want the unglamorous checklist: 🔐 Keys out of the frontend 🧪 Test the broken path 💾 Back up the database 📏 Add rate limits 👀 Read the diff

What are your thoughts on this? @X growth isn't a secret. It's boring: Post original content daily. Reply to bigger accounts early. Answer every comment. Repeat for 90 days. Most people quit at day 14.

Learning question 📚 What's the one skill you wish you'd learned earlier? 💻 Coding ✍️ Writing 📣 Marketing 🧪 Data and analytics What would you pick?