Hi, Till here, worked on the DuckDB-NSQL model on MotherDuck side. 1. Definitely training data (for me), we explored about 10 different directions before settling on the current approach. It's easy to underestimate the effect of training data on the quality of the model. Starting point was the benchmark dataset though, which we assembled manually (to avoid data pollution and also because there was simply no... - Source: Hacker News / 5 months ago

The implementation of Text-to-SQL can be achieved through the use of Vanna.AI, an open-source 🐍Python library that allows the training of an RAG model with queries, DDL, and documentation from a database. - Source: dev.to / 5 months ago

Nice! I’m going to try it out and possibly integrate it into my Python package: https://vanna.ai. - Source: Hacker News / 9 months ago

Nice job! We're building something relatively similar at Vanna AI: https://vanna.ai/. - Source: Hacker News / 10 months ago

One thing that keeps coming up is that SQL equals low productivity. I don't think this is true. I think the culprit is that most developers are using to heavily abstracting SQL using ORMs like Prisma that hides the database and SQL logic. Since building a SQL generator (https://aihelperbot.com) as a side project, I have become much more proficient in SQL and even though I am also locked into Prisma, I use the... - Source: Hacker News / 8 months ago

A few things I have learned over the years and in particular launching and growing my latest project[1]: 1) Track everything including errors. Know what users are using and what they aren't. Remove or rebuild less used features. 2) Find out who your users are and what they value. Ignore non-payers. 3) Economize, market, and document features. You not only need to develop and deploy features, but also to price them... - Source: Hacker News / 8 months ago

It started with me working on a hobby project, aihelperbot.com which enables users to generate SQL using AI. It was build using the comprehensive Prisma ORM. I always found Prisma bloated and it requires almost constantly that you lookup their API for even trivial usage. Source: 9 months ago

Sounds like a combination of an unpolished product with insufficient demand and wrong business model. Users today don't want to pay XXX up front for a complicated product, they want a free tier where they incrementally learn about features. I just looked up what I spend on Google Ads for a small campaign on SQL/AI related keywords for https://aihelperbot.com: - $699.95 (part of a spend $400 to get $400 deal) -... - Source: Hacker News / 10 months ago

Neat setup. Regarding the deploy script. I have just setup a separate VPS for proxying database queries using various Node database drivers for my own project[1] and only used Github Actions managing it[2]: - add build script using Github Action that fails the entire build pipeline if code doesn't build - add deploy script (essentially a few commands ssh into your VPS and pulling, install, building and restarting)... - Source: Hacker News / 11 months ago