I have no entrepreneurial friends… Not one person in my life I can call and say "hey what are you working on man" and actually get a real answer back… I live in Denmark in a small city and I have friends — quite a lot actually. But you know that feeling where you're trying to grow and you just feel kind of outside? Like you're still there, still hanging out, still laughing at the same stuff — but something is just a little off. They're not bad people. Not at all. You just feel like you're on a completely different planet sometimes. And you can see them kind of roll their eyes when you start talking about the things that light you up. Not because they're bad people — but because they just don't get excited about this stuff the same way you do. And that's okay. But it gets kinda lonely... And trust me I looked everywhere. Reddit. Discord. Skool. Facebook groups. Every single one the same story — join, scroll through a dead feed, and even when there is some activity it's never at the scale where you can actually find your people. Someone in your niche. Someone at your level in your age. So I decided to build it myself. A real app. On your phone. Not hosted on Reddit or Facebook or Skool — think Instagram / Twitter but actually way cooler. Because you can feel productive while scrolling. That alone is kind of wild when you think about it. Imagine opening the app and seeing thousands of posts daily. Someone documenting their new AI startup. Someone celebrating their first client. Someone in your exact niche looking for a co-founder. All of them your age. All of them actually building something. I know what you're thinking — this sounds like a scrappy version of LinkedIn. I get it. But it's completely different. LinkedIn is more like "look at me, im in tech sales" — just kidding, I have nothing against LinkedIn at all. But back to the point... This is for people who are in the middle of figuring it out. Like me. Like probably you. Or maybe you have figured some of it out and seen a little bit of success — but you just want friends who actually share your interests and get excited about the same things you do.

Hey everyone, looking for some expert advice on a project I'm building. It's a wheel visualizer — users upload a photo of their car, pick a wheel from a catalogue, and Google Gemini generates what the car would look like with those wheels fitted. The idea is for car shops to embed it on their website so customers can visualise wheels before buying. The stack: Node.js + Express backend, vanilla JS frontend, single index.html. No framework, no database. Six wheel images stored on the server, both the car photo and wheel reference get resized with Sharp, base64 encoded, and sent to Gemini as inline image parts in one prompt. Model is gemini-2.5-flash-image. Up to 3 retries if no image comes back. The problem: Generation is inconsistent. Sometimes it works great, sometimes it fails or the wheels in the output don't match the reference image at all — even with identical inputs. I'm pretty sure this isn't a hosting issue. I think the core issues are (claude code said it): - Gemini interprets rather than copies, so the wheel reference isn't being replicated accurately - Both images go in as raw base64 blobs with no clear instruction distinguishing which is the reference and which is the target - The wheel images are standalone product shots so the model has to guess scale, perspective and angle - No seed or temperature set so output is different every time Has anyone dealt with inconsistent image generation with Gemini specifically (or any other model)? Is there a better way to structure the prompt or payload so the model reliably uses the reference wheel? Would a different approach work better here altogether? Happy to share server.js if anyone wants to look at it properly.

And I am using Claude Code with Opus 4.8. What I have seen since recently - it uses a strange grammar forms like ""I went looking at your website"" - when I am using agent skills and I need to make it's modification, it add to memory file instead of the skill file Did anyone notice any issues on their site?

If you don’t know who Andrej Karpathy is, here’s the short version: he was a founding member of OpenAI, ran AI at Tesla, and now runs his own AI education company. The dude is a literal legend in the space. He just dropped a 3.5 hour video called “Deep Dive into LLMs like ChatGPT” and it’s free on YouTube. Here’s why it actually matters for us: Most of us are out here vibe coding, prompting, building stuff, and hoping it works. When it breaks, we don’t really know why. We just try a different prompt and pray. This video fixes that. It’s not about prompts. It’s not about tricks. It’s about understanding the actual machine you’re talking to every single day. He covers: • How these models actually learn (pretraining) • Why ChatGPT hallucinates and what to do about it • Tool use, memory, and how the model “thinks” • Why models are weirdly good at hard stuff and weirdly bad at simple stuff (he calls it “jagged intelligence”) • RLHF, AlphaGo, DeepSeek, the whole stack The reason I’m pushing this so hard: if you understand even the basics of what’s happening under the hood, your prompts get sharper, your debugging gets faster, and you stop being surprised when Claude messes up the same thing 3 times in a row. You don’t need to be technical. He literally made it for a general audience. Watch it once just to take it in. Watch it twice to actually start applying it. Here’s the link: https://www.youtube.com/watch?v=7xTGNNLPyMI If you’re serious about going from vibe coder to actual builder, this is one of the single best free resource I’ve seen this year.