I agree that universal libraries by themselves are insufficient without proper curation and filtering of content. Training AI on data that has been carefully fact-checked and made more truthful could be one step towards improving information access. However, this is just one part of the solution. Just leaving an open ended question for everyone here: What other changes to our information ecosystems do you think are needed to surface more novel and useful insights from the data we have access to?

A very well written article. Clear explanations and we'll structured examples. Thank you for sharing this. 😀

Thanks for sharing this detailed breakdown! Sure it'll be useful for a lot of folks building similar apps in AI

Good one! I think one big difference is also if you build an application or a framework. I do understand that using TS in frameworks can get quite complicated. I use one repository still in JS and it is so easy to miss basic things like missing imports, not available parameters, etc. I can't think of working only with JS again.

This is a lovely article! I switched over to Svelte a couple of months ago as I was tired of all the complexity I was having to deal with. I'm happier and I know exactly what's going on much more than I did before.