David Moosmann

Need some opponents. Does anyone dare? learnclash.com/c/david-2fBI

@FlutterDev Genkit Dart is the part Im actually waiting for. My app already runs Gemini for question gen and a chat tutor. The AI flows live in TS, the rest is Dart. Switching prompts means jumping languages every time. If Genkit Dart hits stable, that pain just goes away.

Flutter developers can now write backend logic in the same language, and with the same tooling, as their frontend 🎉 Follow along as we build a complete serverless application live, using full-stack Dart across client and cloud environments → goo.gle/48zqPnR

@theo Same lesson Anthropic is learning right now. Message count means nothing once an agent loops for hours per prompt; the cap stops mapping to cost. Compute-based pricing wins the second time around, always does.

- 15 messages - $221 of tokens - 1.6% of my $40 plan used It's obvious that GitHub couldn't keep this model for billing on Copilot.

@betomoedano Big one missing: how you write the form matters more than people think. Reviewers read hundreds of these a day. One repro step plus a customer complaint screenshot beats paragraphs of context. Skip the sales pitch, write it like a triage ticket.

Apple can review your app in minutes Expedited reviews allow you to ship quick updates to your app. Here's how to avoid getting banned: - Rule #1 – Don't submit excessive requests otherwise future requests get denied. - Rule #2 – Use this for critical bugs. e.g. Users cannot access the app because of a bug I've requested expedited reviews multiple times and it takes ~5 minutes to get approved if no issues are found. Link below

@scaling01 60% by 2028 sounds bold given how often the first attempt still misses today. Watching my agent run, the slow part is noticing when its wrong and starting clean. I still nudge it three times before it lands. Self-improvement needs that loop closed first.

Jack Clark co-founder of Anthropic says there is a 60%+ chance of an AI system that can autonomously build it's own successor by the end of 2028

@giladbuilds @gregisenberg At one point the remaining subs run out. Many will switch to their own solution the better the models get

@damoosmann @gregisenberg Agree - I think a lot of older SaaS cos are basically coasting at this point

I don't know why more people aren't buying dead SaaS companies and turning them into AI agent companies. 1. Use OpenClaw, Hermes, Perplexity Computer etc to build an automation that scans Product Hunt, Acquire, and app stores for dead SaaS products. Filter for ones that launched 2019-2024, had real customers, and went quiet. 2. Reach out to the founder on X. Most of them will respond within a day because they've been wanting to sell for a year and nobody asked. 3. Buy it. $5-30k. Sometimes less. 4. Export the database. Feed it to Claude or GPT. Map every workflow their customers were trying to do. 5. Read the support tickets. This is the goldmine. 200 strangers already told the last founder exactly what they needed and he couldn't deliver it. 6. Build an agent-native version that actually does those workflows instead of giving people a dashboard to do them manually. 7. Upload the old email list to Meta. Build a lookalike audience. Those old customers have moved on. You're not selling to them (realistically). You're using their data to find the next them. 8. Run $20/day ads targeting people who look exactly like the customers who already validated this market for you. 9. Build content around the exact pain points you found in the support tickets. Post on X. Post on YT. You already know what to say. 10. You now have the customer profile, the pain points, the pricing sensitivity, the churn reasons, and a lookalike audience. Your competitor who's starting from scratch has a landing page and a guess. The dead SaaS acquisition playbook is going to be one of the biggest quiet wealth builders of the next 5 years. Most SaaS products are a collection of workflows that can be rewritten as agent skills. Many will die. The top ones will pivot to agent companies. Build agent companies.

@adamlyttleapps App Store conversion and CPI sit on the same lever. A 10% listing means youre paying twice what a 20% listing costs per install, same ads. Easiest wins are usually the screenshot order and the icon. Same ad spend, more installs.

Actually, no My biggest issue is cost per installation I think I'm incorrectly combining App Store conversion with cost per installation… There is possibly a correlation there (a bad converting app store listing will leak users) But really: my issue is the ad spend right now

@VicVijayakumar Same shape here. 4 months building, 168 backend files. Naming-collision week was rough. What kept it from collapsing was writing down my folder layout in a skill the AI reads on every run. Slows the entropy enough to keep shipping.

hot take: this is fine your app either dies pre-launch / pre-revenue or lives long enough for you to need to tackle technical debt

Maybe the credits arent the problem. Hit zero the same way last month. What helped me was stopping to rewrite the brief before kicking off the next run. Tokens burn fastest when the first try doesnt land. You keep nudging instead of starting clean. aHit zero the same way last month. What helped me was stopping to rewrite the brief before kicking off the next run. try doesnt land. You keep nudging instead of starting clean. Hit zero the same way last month. What helped me was stopping to rewrite the brief before kicking off the next run. try doesnt land. You keep nudging instead of starting clean. Hit zero the same way last month. What helped me was stopping to rewrite the brief before kicking off the next run. Tokens burn fastest when the first try doesnt land. You keep nudging instead of starting clean. Tokens burn fastest when the first try doesnt land. You keep nudging instead of starting clean.

This is the situation I am in with Codex: Pro 20x account, 2 days left until the weekly reset, and I am down to 0%! 😅 Now I’m burning through my API credits! Though, I think I’ll defer making another Codex pet until May 5th! Burn, tokens, burn!🤣

xSame fear my first month. Then I asked AI about auth and Firestore rules upfront. App is locked tighter now than the SaaS I used to pay for. Public supabase tables happen when somebody skips that step. AI will write the fence if you ask. Then I asked AI about auth and Firestore rules upfront. App is locked tighter now than the SaaS I used to pay for. Public supabase tables happen when somebody skips that step. AI will write the fence if you ask.

I’m terrified of using vibe coded products 99% chance all my info is on a public supabase table any “hacker” with half a brain cell can access

@steipete z

Released 🚦RepoBar 0.4.0. This one makes the GitHub menu a lot smarter: persistent SQLite caching, fewer wasted API calls, visible rate limits, better Issues/PR loading, archive fallback support. Tiny menubar app, increasingly useful daily tool. github.com/steipete/RepoB…

@r0ktech Ran 4 Max subs to dodge this exact pain. Each one resets at a different hour, so something is always available. Even then, hard problems eat the whole stack by lunch. That 7pm wait stings most when youre one fix from shipping.

Claude Code: "You've hit your limit, Limit resets 7pm". Me from 5 - 6:59pm . 💀

@jasonlk Built my whole backend on Firebase. Never used the word headless once. Turns out the label was doing more gatekeeping than the actual concept. Half the words in this stack are just velvet ropes.

"Headless" was pretty nerdy before vibe coding Only a developer could use a headless product Today? Anyone can

@steipete Building skills for everything is what made this click for me. Reusable instructions I write once per workflow. One does my X engagement, another my git pull-request flow. Each works better than the SaaS tool it replaced and I can tweak it on the fly.

Seems I have to build all the tooling for the future of software myself. With Claws and Tokens!

@bridgemindai Been running 4 parallel sessions on different branches for months. Feels like four chess games at once. What multitask adds is parallel inside one chat with shared context. That kills the part I hate most: re-explaining the same code to each fresh session.

Cursor's new Multitask mode is insane. Submit prompt after prompt in the same chat. It works on all of them simultaneously. I'm running Opus 4.7 with 1M context firing off parallel fixes while querying the production database at the same time. Claude Code and Codex don't have this. Cursor just jumped ahead.

@gdb Skills are great for stuff you do over and over. Idea testing happens once per idea, so it usually doesnt earn its keep as a skill. Best test I've found is still shipping the thing and seeing if mum keeps opening it. That signal beats any pressure-test.

codex for startup ideas

@sama From the user side, resume is the bit saving me the most time. First run almost never lands clean. And picking up from a snapshot beats starting fresh by miles. You only feel it once an agent crashes four hours in and you have to start from zero again.

Agents SDK 2.0 is underrated

@anumness Apple Search runs lean. You can learn something at $30/day. Meta wants real money before its algo even sees you. That part surprised me.

@damoosmann Meta needs a lot of money poured into it for it to be profitable

Marketing has been by far the biggest frigging frustration of being an indie dev. - I tried Apple Search ads. They were not bad. I either broke even or profitability was kinda low. - Tried meta ads: lost money - Tried SEO: yet to see results - Spoke to several influencers. They want an arm and a leg and Im skeptical about how well it would work. Im exhausted and burnt out. But today I finally feel like I had a breakthrough.

@buildwithsid Yeah those first paying users are brutal. Mine came when mum showed the app to her friends, slow build at the start. Past ten and each new sub already knows someone paying. Whole different pace.

$2,500 MRR is just 10 users paying $25 a month

@jasonlk Felt this last month. Tools I used every day, now its weeks between logins. The agent did the work and I never bothered telling the tool I left. Renewal email is the only signal something happened.

We stealth-churned off Notion months ago. We didn't realize it. We didn't tell Notion. It's just ... our agents have no need for it. They're doing the work now. Stealth churn is the under-discussed agent risk. Usage drops to zero quietly. The renewal email never even goes out.