I remember reading pg's article on Lisp and startups, and at the time questioned if it was just mere luck. Then having a cursory look at Lisp, I questioned its relevance to the modern world.
... fast forward a decade later and I'm reading books on functional and logical languages for work. After the first chapter of The Little Schemer, I was at first blown away with the content, but then sad after I realised I had put off reading it late in my life.
If you're reading this comment and thinking Lisp and what's the point? Take a deep dive. It you're still questioning why, I highly encourage you to read The Little Schemer (and then all the others in the series). Scheme, Lisp, and now Bel, are a super power... pg's article was spot on.
You and I both like Rust. Concrete performance is one key reason why I choose to use Rust for various things, and all Lisps I know of are just generally slower and use more resources; I’m not aware of any Lisp attempting to address this problem, and from what I do understand of Lisps (though I’ve never seriously coded in one) they seem at least somewhat incompatible with such performance. I’m interested in whether you have any remarks on this apparent conflict of goals.
(I’d love to be credibly told I’m completely wrong, because syntax aside, which I could get used to eventually, I rather like the model of Lisps and some of the features that supports.)
There was a systems type of Lisp called PreScheme that was closer to what you're envisioning. Carp also aims at no-GC, real-time use. Finally, ZL was C/C++ implemented in Scheme with both their advantages compiled to C. Although done for ABI research, I've always encouraged something like that to be done for production use with tools to automatically make C and C++ library bindings.
https://en.wikipedia.org/wiki/Scheme_48