3 ms·

This feels like an ad. The author created a tool that makes it easy to ditch ORMs. The tool is referenced 10 times in the article, including the article's first

by thih9 4y ago

This feels like an ad. The author created a tool that makes it easy to ditch ORMs. The tool is referenced 10 times in the article, including the article's first sentence and the last sentence. The article also includes examples on how to use the tool. I'd be fine if it was a "show hn" post. But I thought this is an objective article about ORMs and their pros and cons; this is not the case.

tomhallett 4y ago
I agree that this blog post is part "orm" and part "lead gen for my saas app which helps you write sql queries". I think a blog title which reflects this, ORM and AiHelperBot, would be more helpful. That being said - the faq's on https://aihelperbot.com/ https://aihelperbot.com/ don't answer the main questions I have: 1) I saw a banner somewhere, can't find it now, which said "your schema is only stored in the browser". While that is nice, is my schema transmitted over the wire to your cloud service's servers? 2) If your servers get it, do you send my schema off to OpenAI's ChatGPT? Or some other AI as a service? 3) Do I own the intellectual property of all sql queries written by your tool? 4) How do you verify the "validness" of a query? For example, do you take my schema, load it into a small postgres/etc database and have your code run the query your AI generates? Do you parse/lint the query? Or is the AI good enough that it always generates valid sql? 5) I'm also interested in the "correctness" of a query, from the perspective of "user with the name of john", will the resulting sql returns results with name = "John", "john", " John ", "Johnny", etc - but I am fine with that being left up to the user. 6) Is snowflake/duckdb/clickhouse on your roadmap? Rather than replacing ORM's, helping business users write analytics/BI queries seems like a home run use-case.

3 ms·

This feels like an ad. The author created a tool that makes it easy to ditch ORMs. The tool is referenced 10 times in the article, including the article's first

by thih9 4y ago

tomhallett 4y ago
I agree that this blog post is part "orm" and part "lead gen for my saas app which helps you write sql queries". I think a blog title which reflects this, ORM and AiHelperBot, would be more helpful. That being said - the faq's on https://aihelperbot.com/ https://aihelperbot.com/ don't answer the main questions I have: 1) I saw a banner somewhere, can't find it now, which said "your schema is only stored in the browser". While that is nice, is my schema transmitted over the wire to your cloud service's servers? 2) If your servers get it, do you send my schema off to OpenAI's ChatGPT? Or some other AI as a service? 3) Do I own the intellectual property of all sql queries written by your tool? 4) How do you verify the "validness" of a query? For example, do you take my schema, load it into a small postgres/etc database and have your code run the query your AI generates? Do you parse/lint the query? Or is the AI good enough that it always generates valid sql? 5) I'm also interested in the "correctness" of a query, from the perspective of "user with the name of john", will the resulting sql returns results with name = "John", "john", " John ", "Johnny", etc - but I am fine with that being left up to the user. 6) Is snowflake/duckdb/clickhouse on your roadmap? Rather than replacing ORM's, helping business users write analytics/BI queries seems like a home run use-case.