The post gives an idea of how we’ve been approaching this challenge. It would be great to hear what the community thinks about our approach and whether we are missing something impactful that we should be focussing on.
Also, I hope you give OCaml 5 a try if you haven’t already
I don’t see any option to split threads on this Discourse instance, otherwise I would have followed this recommendation. In any case, let us try to cool down a bit. @Chet_Murthy and @jbeckford could agree to disagree (or to be talking about different things, training vs. inference, etc.) and that would leave more space for other kinds of feedback on the post.
(I found the various works listed in this blog post very impressive in variety and depth. I knew each piece, but seeing it all together gives a strong impression.)
It’s a little hard to find, but you click on the spanner of the top-right of the post, click on “select posts”, and then have the option to move them (either to a new topic, as I’ve just done, or to another existing one)