51–60 of 412 posts
Many repeat the point of “$165k is cheaper than team of multiple engineers working on the rewrite for a year”, which I think is flawed — the team of engineers would have produced idiomatic rust, and it would take probably 100k+ of tokens more to make the bun in rust idiomatic rust.
Earlier quoted context omitted.
Author kept "going" in title because there has been no minor/major release in last two months which unusual for this project if you review the older releases.
A major rewrite in another language is also unusual for this project. It took TypeScript a year to go from announcing tsgo to releasing TypeScript 7.0. That work was done in parallel; here the work is being done serially — and it’s likely to take much less than a year for a new release.
Many repeat the point of “$165k is cheaper than team of multiple engineers working on the rewrite for a year”, which I think is flawed — the team of engineers would have produced idiomatic rust, and it would take probably 100k+ of tokens more to make the bun in rust idiomatic rust.
I feel your estimate of tokens is a few orders of magnitude off, it’s on the low side. I use more (albeit cached) when centering a div.
I will have to steal it for an upcoming AI tools meeting I have at work.
Also, pretty clever as centering a div w/ CSS has been notoriously difficult to achieve.
I still believe it was the good decision, but I also know that I wouldn't be the first person to run the release in prod.
I think Jarred should start making release candidates instead of releases to take some of the pressure off.
I really do not understand how software developer think anymore. Using a LLM to translate a project in a short time, is by itself incredible. Just like one-shot whatever office clone. But what makes software is not the fast creation of a "product" but that actual development of its features. Figuring out how everything needs to work together, fixing the bugs, and the o so boring UI work. I have used LLMs to create st…
> I have used LLMs to create stuff like word clones just for fun. It was a disaster. Sure, it had the basic functionality. But the moment you started with page structure (harder then one non-stop scrolling page), tables, images, rotating, and so many details that make up just the basics of word. Not even the extended functionality. You see every LLM just fall on its face. What do you think of https://scramblequest.ap…
In case of ccc (claudes C compiler) they just did not use any tests, though they are many existing C testsuite. They just stopped, when it finished compiling the kernel, still failing hundreds of tests.
If you're being pedantic, no it's not "done," but nothing ever is.
Earlier quoted context omitted.
> I have used LLMs to create stuff like word clones just for fun. It was a disaster. Sure, it had the basic functionality. But the moment you started with page structure (harder then one non-stop scrolling page), tables, images, rotating, and so many details that make up just the basics of word. Not even the extended functionality. You see every LLM just fall on its face. What do you think of https://scramblequest.ap…
Nonsense. Provide/create proper tests for all features, and let the LLM test and fix everything. In case of ccc (claudes C compiler) they just did not use any tests, though they are many existing C testsuite. They just stopped, when it finished compiling the kernel, still failing hundreds of tests.
I don't get it. Because they haven't cut a new arbitrary version number, we're denying it got done (in their perspective) and being used in a massive product already? If you're being pedantic, no it's not "done," but nothing ever is.
One single product, AFAIK, and one ladden with bugs and issue too.