Universal Data: Generate

Hello Product Hunters ! 🏴‍☠️ Sometimes, quick and dirty analysis are made impossible because getting access to the desire dataset is really hard, so, I built a small tool to generate any data on-the-fly using AI knowledge. The idea is to lower friction so that quick & dirty analysis are made possible. Kinda of the ChatGPT for the data analyst. ⚠️ Don't rely on this for real stuff. It's just a playground ⚠ It's a common wisdom that LLM data is unreliable, and well... it's true. However, I believe that we should make progress on this, so it's better to address theses issues upfront. Technically, we try to improve few fronts: - Avoid obviously fake data "John Doe" - Avoid cut in response (like "...") - Estimate and extract the correct number of rows As well as providing an api / small website easy to use. It's still falling short on lots of things, but I thought it would be great to hear your feedbacks. So, I invite you to play with it and I would really love to hear your feedback ! Cheers, Ben

Good

Thanks for being honest about the experimental stage. Still sounds useful. Fingers crossed for the development.

A bit slow but works quite well, I tried `pipelines` with some `name`, `source`, `destination`, `published` as fields and got some interesting answers. Maybe more specific answers would be more valuable (the result was quite generic in my case)

@benjamin_derville congratulations on the launch 🚀

Very cool! Congrats Benjamin Derville. Do you have training docs so users can learn how to create better queries? Just played a moment and the output was great actually.

Congrats.