fast.ai is how I started learning about ml in 2017!

Send me a DM

Congrats Matt!

Change your language setting from Amharic back to English

Shepherd is such a beast! When I was doing ApF stuff for Parabible, I literally just kicked the can down the road on it...

Congrats, glad to hear it :)

I saw you're also presenting at a Chicago meetup! Wish I could join, hope it's a good time :)

Really enjoyed reading this. Love most of the work you guys are doing! Keep it up :)

When I first read "no side-effects", my brain immediately said "he's using pure functions?" But I'm also glad to hear there were no side-effects of the lower dose...

The one (successful) example the Dems have of responding in defiance was the government shutdown (2018-2019) when Trump refused to sign the appropriations bill, and Pelosi said "fine, here's your shutdown".

Thanks! I now recall seeing that this was possible...

Could you point me to an example/docs?

I'm hoping that it's possible to rip the right stuff out of JS/TS and make minor additions, so that you could easily mistake Chicory for JS. But suddenly you get a ton of guarantees. Admittedly, that may be a pipe dream...

Ja, I've used Rescript for a couple of projects (I actually asked Kris to do this episode), and I agree, the objectives are very similar. The big thing for me is that Rescript is still quite a different language. All of `module`, `@react.component` and `let make = ()` are new.

I'm curious, in what ways is it not a "perfect" fit?

One of the things I realized in this episode is that a language I'm just starting out designing (chicory-lang.github.io) has nigh identical goals, but is launching from the TS side rather than the OCaml side. So I'm interested in any pain points.

Thanks so much! I love that this worked out :)

I really have no business doing any kind of language design. So this is really a PoC of a Poc, set expectations accordingly :)