Find your startup: the AI startup finder concept, a voice memo

Source: a voice memo by Dinis Cruz, transcribed by otter.ai, file 84369837-done_AI_Startup_Finder_Concept_otter_ai_transcript.txt, recorded 20 October 2026. Published here verbatim, with the transcription's own wording, on 20 October 2026 as the brief this site starts from. Status: the brief; the site was built from it the same day. Corrections to what was done with it go in BRIEF-CORRECTIONS.md, never into this file.

Dinis Cruz 0:02

So here's an interesting form or article, a article to write in brief to create. And actually, I think we should have a startups AI. I think we have enough material now, especially the book, especially because what I'm about to ask to do next is, I think, an interesting question because I keep coming back to the idea of how do we empower people? How do we empower the people to to focus and to to do the thing right, and to for the startup, and again I want to help people, and and I I think what we can do is a very interesting online questionnaire and very very interesting again powered by fractal semantic crux powered by a an LLM or maybe even not an LLM right because some of this is kind of like can be made deterministic which is kind of what we want to do and I think the question is find your startup right and I think what would be cool to do is to do a mapping, taking into account all the ideas and the concepts I've been pushing here, which is fundamentally the whole idea that it has to be something that will ship, something that is shipped to do, something that is actually can be done right now, and can be done, you know, can be executed. So it's almost like can we create a questionnaire and a form, and this goes back all the way to the idea of creating dynamic questionnaires where the questions changed. It's almost like this who done it of questions, where I think we can now create very powerful state machines and very powerful things. And I think it's two parts here. I think we can make one again. Let's make this so that it works from the website, but also is almost designed to be worked by another lamb like Claude running interviews. Because here's my my point: if somebody are not sure or needs to figure out his startup, like it's almost like what is the person's startup? Where should they focus? What should be their key sort of direction of travel, so we can help with that because we can have either again on the website via maybe even our storyboards that we are working on, but also with an LLM driving it again on a website or somebody's cloud or ChatGPT directly. So this guidance needs to work for humans and for agents, but then the idea is to zoom in into that person's and say, okay, here's your strategy. Here's what we. Here's your idea, and to be honest, the idea could be like here's your focus. To sorry, dude, you you you have to go old school, right? You're going to have to raise a lot of money to pull this off, right? And I think it all comes down to the idea of can you ship it? Can you get it to the hands of users? And I think it's important that I think what's important here is that we're not passing judgement. You know, you know, the right startup for the person is the one they're passionate about. Is the one they want to work. Look, think about, look at my example. Right, I I'm a framework guy. I build stuff refactoring, so I naturally focus on the components. I naturally focus on the infrastructure elements, and I arrive at my, I think my real startups, or my I must say my profitable and my effective startups, you know, three, four startups later and X components later, right? I think that's okay, right? I think we're not passing judgement. I think what we can help is we can help with focus, but also we can help open there for them to ship. So I think there's a spectrum here. There's the person who has not figured out where their startup is, so they just want to start from okay, start from zero. You know, what do you want to do? What do you want to build? How do you want to build it, etc. And then we can start from somebody who already has a very good plan, and and who already has a really good idea. But we can bring focus. And to be honest, just about most people I meet, they have really good ideas. What they're not doing is they're not shipping, right?

Dinis Cruz 3:59

They they're doing the typical thing of it's not ready yet, you know. I want to just get this feature out. Oh, you know, I got you know. They have all those. They can come up with every single excuse in the book, right? Why they're not ready yet, and I would argue that that's the problem. They have to ship it. So I think we can help with that. We can help to bring that focus, and I think what is also is is a great way to start the book, right? Because the first thing you should do in the book is to get this zoomed in, right? It's sort of like how do we help them to find their startup, and ultimately it's about coming up with you know their own words and and coming up with their own description of this is kind of what you should be focusing on, right? And I think here maybe what we can do is we can be quite strategic, thinking about what are stuff that can be done on the web, or maybe stuff that can be done via. Actually, I think that's better. I think what we should do is we because my idea here is to get Claude to interview the person. So I think for now, let's start by this is something that they will is designed to be used by Claude or ChatGPT because you need that data analysis. You need an agent behind the scenes. So what we're going to be doing is creating the primitives, and maybe I'll show some examples. Right, I'll do it myself. Right, so, but I think it's let's build these toolkits that we can add to the thing that basically allows them to find their startup, and for now, assume that you're going to need what's it called. Assume that you're going to need an LLM to run it, and I think once we've done a whole part of this, we we can find a better way to to do it programme some of it programmatically, right? But I think the idea here is that either they upload a whole bunch of documents, or they basically just have have Claude interview them or ChatGPT view them because now they're very those works really really well. Cool. Let's build this

Unknown Speaker 6:11

up

Transcribed by https://otter.ai