Just a tiny notice, the github page, nor the website seem to currently contain an “installation” link. The one found by google returns a ‘page not found’ for the current version.
Well, code size is another interesting aspect here. A JIT compiler can effectively create any number of versions for a hot method, based on even very aggressive assumptions (an easy one would be that a given object is non-null, or that the interface only has a single instance loaded). The checks for these are cheap (e.g. it could be encoded as trapping an invalid page address in case of NPEs), and invalidation’s cost is amortized.
Contrast this with the problem of specialization in AOT languages, which can easily result in bloated binaries (PGO does help here quite a lot, that much is true). For example, generics might output a completely new function for every type it gets instantiated with - if the function is not that hot, it actually makes sense to rather try to handle more cases with the same code.
I do think, that in the general case, a JIT compiler is required: you can’t make every program fast, without having the ability to synthesize new code based on only-runtime available information. There are many where AOT is more than enough, but not all are such. Note, this doesn’t preclude AOT/hybrid models as pjmlp correctly says.
One stereotypical (but not the best) example would be regexes: you basically want to compile some AST into a mini-program. This can also be done with a tiny interpreter without JIT, which will be quite competitive in speed (I believe that’s what rust has, and it’s indeed one of the fastest - the advantage of the problem/domain here is that you really can have tiny interpreters that efficiently use the caches, having very little overhead on today’s CPUs), but I am quite sure that a “JITted rust” with all the other optimizations/memory layouts could potentially fair better, but of course it’s not a trivial additional complexity.
Last time I checked, it couldn’t handle expressions that are not just tokens one after the other. For example, German separable verbs. I tried fixing it here: https://news.ycombinator.com/item?id=38915786
I have been working on a similar project on-and-off in my spare time, the only remotely interesting feature that other similar software may not have is that it actually tries to parse/analyze sentences (with an NLP lib). It's made specifically for German, and the reason why I wanted to make it is that no existing software managed to handle separable verbs properly - for example learning "Wir fangen jetzt an." is just wrong if you learn it as 'fangen' and 'an' separately, you actually care about 'anfangen', dictionary-wise.
It unfortunately does have false-positives (a complete solution would require LLMs, I believe over the much less complicated NLP algorithms - I just don't want to send whole books to ChatGPT, as that would quickly become expensive), but I found it usable, so I made it public now: https://github.com/tenaf0/lwt
I don't want to "advertise" it even more, as the NLP lib is run by academia as a free service, and I don't want to overburden it (I have been planning on hosting it myself, but didn't yet get there).