Personally I don't really care, because I like coding and learning myself and DeepSeek Flash is all I really care about. But it's really easy to have a ton of benchmarks where the top models can't get anywhere close - and I like to test them on these problems to see how good they are getting.
Fable 5 is def a little better than 4.8 btw.
Myth. Total myth! I recently had to beg for more RAM after continually hitting swap space which causes tools like dictation to stop working, failure to load certain websites without rebooting, and so on. Devs do in fact need powerful machines and the ~$500-1000 an employer saves upfront in machine costs is dwarfed by productivity losses.
Giving your engineering employees new machines in a 2-year cycle that are between the middle and high end is one of the cheapest ROI decisions that a tech org can make.
Fable 5 found quite a few issues Opus 4.8 missed on code review, even though the stupid cybersecurity nonsense downgraded it. I can't tell you more, I only get a single session per 5h window on Max 5x. Only ran two sessions so far.
A small portion of this effort is having a high quality Lua in Rust repo. I’m using mythos to fix some of the performance issues with my Lua interpreter that gpt 5.5/ opus 4.8 had stone walled on.
Not sure if Mythos will be able to crack this but it has been running for a couple hours now with some promising results.
Performance charts linked here if your curious https://github.com/ianm199/lua-rs
The other reason is that because mlua is just a wrapper around the C code, it has unsafe you can't really get around. So for example Lua is used in Redis, which has this critical CVE https://github.com/redis/redis/security/advisories/GHSA-4789... that a memory safe version of Lua wouldn't have to deal with.
Mlua is still fine or even better for many other cases though!
It just seems like a lot of hassle to write a lua interpreter, although it would be nice to see a high quality one in Rust :)
Hematita was promising, but looks abandoned.
And yes it seems like there has been many attempts to get a solid Rust Lua over the years and most never reached parity so hoping some people can find use case for it! This one is at full parity in terms of behavior and performance is getting to within striking distance.
On the margins, suppose the prompt is literally: "Build a feature complete, high polish Facebook clone". Facebook is complex but likely not super complicated tech, and still I would assume that (after having burned through a substantial amount of tokens) you would find substantial enough differences in the outcomes between different models on that prompt on various fronts.
The above ask is obviously not useful, but what's preventing you from taking on bigger chunks until you approach the limit? At some point you would hit a boundary, where the diff will be obvious.