Lukas Hermann

I'm going to get rich and this is how I’m going to do it. Part 2. Now don’t take me wrong, I’m not rich yet. But I'm closer. 4 years ago I was laid off, and this is what happened... x.com/_lhermann/stat…

@philipyoungg This stuff is terribly itchy

Every time I went to sleep, my AC cries so hard because of radiating heat from uninsulated ceiling. Seriously thinking to buy fiberglass batt and install it myself.

@martinleblanc No credit whatsoever?

2000 artists each created a 5 second video of a ball descending. Then all shots were combined.

@forgebitz @thdxr Which everyone knows who's tried to hire a good dev

@thdxr the pool of really good devs who love building and keep learning is extremely small AI is going to curb stomp a lot of the average devs

what's your most controversial opinion? looking for things that provoke me to think hard about something i'm taking for granted DM me if not comfortable sharing publicly

@thdxr Coffee tastes awful

@_lhermann The city is famous for Mercedes and Stagetimer 🔥

I was wondering what are the best tech cities in Europe right now. Based on what I see online I’d probably say Stockholm, Paris and…? Where is the cool stuff happening right now? I really hope to see a future when people don’t feel the need to move to the US.

@d4m1n I got a Product of the Day badge back in the good times. Holding on tight to it!

you only have a 4% chance of getting featured on Product Hunt in 2026. thx slopstack goblins

@romancm It's great to build product. It's hard when trying to launch a product in between so many other new products.

@_lhermann Grateful I got to build from scratch with AI as a senior. huge advantage. You move fast, iterate cheap, and code becomes disposable. But without pre-AI experience, it’s tough IMO. So many times ive caught agents going in the wrong direction or taking shortcuts. Experience is key.

I'm so happy my product got successful before AI. Scaling an already working business with AI is fun! Building a new product right now and competing with 200 identical slop-products launched each day is brutal!

@xmaxsmi @IurySza Unfortunately

@_lhermann @IurySza Last updated 4 years ago, or am I missing something?

I added a German Law tool to my home-brew OpenClaw alternative. Now it can look up any regulation and help me win against German paperwork ⚔️

@paulo_creates You can build it based on this: github.com/bundestag/gese…

@_lhermann Where can we get access?

@tobimori Stuck in the straight of hormuz

seriously wie soll ich weiter leben

@samiralibabic I'm not going down without a fight

@_lhermann Neither OpenClaw nor Chuck Norris can win against German bureaucracy.

@inxanity06 Be snarky with me.

@_lhermann What can it that perplexity can't ?

@atbrakhi It's a sad state of affairs in Germany. Few invest. Even fewer invest wisely.

People in the EU, especially Germany: are you investing? I'm genuinely surprised the majority isn't. The numbers are even worse among women. Please ladies, I beg you. Invest. DM me if you're scared to start, I am not a professional but can help you with whatever is blocking you from getting started!

@IurySza Official Bundestag Repo: github.com/bundestag/gese…

@_lhermann Hey, where do you source the data from? Curious about your setup.

@robin_liquidium But it's not vector search. Agent has to know the exact words.

@_lhermann found this MCP as well: github.com/Ansvar-Systems… haven’t tried myself but did some light research and it looks pretty good

@yipengsss For once the bundestag did it themselves: github.com/bundestag/gese…

@_lhermann Will you gonna make it open source?

@options_g The war has been lost long ago...

@_lhermann you're winning the battle but lost the war

@atbrakhi It's always human-in-the-loop. And I run the important stuff by my Steuerberater. But it beats me doing it alone.

@_lhermann How much of human verification does this need in general? Ps: doing something similar but more in investment/taxes side. I end up doing lots of human verification due to lack of trust in AI.

@GhostOfKaido Set up my own vector DB and embedded everything in chunks. Plus a lean MCP tool so the agent can discover it.

@_lhermann Welches mcp nutzt du dafür wenn ich Fragen darf?