nchie

I've (more or less; I've read quite a bit of the code) vibecoded several houndred thousand lines of Rust and I've not seen this happen a single time. It sounds like something it'd do when you ask it to "write a linked list while satisfying the borrow checker". Are you sure you haven't (possibly unknowingly) been giving it instructions which ended up luring it into doing these things?

In my experience, when that happens it's most likely because you drew the wrong boundaries. Iterating on the boundaries also becomes quite cheap when developing this way though, you should never expect to get them right the first time unless it's a very common problem you're solving _or_ you've done it before.

For my side project where uptime doesn't really matter but the codebase is quite huge (gamedev of a multiplayer game with Bevy, so both server code, game code, content build pipeline, as well as a TON of editor and authoring code, probably 150k+ LoC), I've been moving towards just reviewing boundaries between modules, APIs, and schemas, and I'm pushing towards keeping things as modular as possible. If something ends up wrong, it going to be isolated and it doesn't infect the rest of the codebase. Every now and then I zoom into a module and tidy it up (or even rewrite it).

I wouldn't dare doing this "carelessly" for anything where people other than me actually depend on it, but you can move very fast while doing it and the risks with bad code is quite hedged. Different modules can mature at different rates when it becomes necessary.

Maybe not related to all of what you were saying, but I think this enables scaling without ending up with progress grinding down to a halt due to shitty code.

That honestly sounds like it's more about either the quality of the codebase, or the driver.

I understand that to be the "emergent abilities" which are spoken about. There are correlations in the dataset that are strong enough for it to seem to have an understanding which wasn't obvious it would have from simply "predicting the next word".

>The coding agents got good in November.

Maybe irrelevant to your point, but I'd argue they were really good already in May if one used the right workflow (planning etc.). They've become better, but they're not saving me significantly more time now than they did 12 months ago.

Agreed, async function coloring makes for better structured code because it incentivizes keeping IO code near the edges while having a synchronous core.

I doubt it, the difference between someone slightly inefficient and someone extremely efficient isn't big enough to matter compared to how much they cost in salary.

That doesn't make it not true though. Markdown generally supports HTML (though oftentimes only a subset), and is typically styled using CSS. Using a web view makes complete sense to me.

> But are we really saying that the primary motivation for async/await is performance?

Of course - what else would it be? The whole async trend started because moving away from each http request spawning (or being bound to) an OS thread gave quite extreme improvements in requests/second metrics, didn't it?

I haven't made a ESP32 design, but I recently learnt KiCad and PCB design enough to do a RP235x board with a non-reference design choice (1.8v VDDIO). I only used the official hardware guide + LLMs for questions, and had it work on the first try - it wasn't too hard!

But dependencies are part of a website? It literally says "Still here when the internet isn't." - but I can't go on there without an internet connection?

I use Claude for work and Codex for private use due to already having a Plus subscription.

I can't say that I have noticed that 5.3-Codex is much better, but it's definitely on par with Opus 4.6, and its limits for $25/months is comparable to Max x5 at 1/4th of the cost (not to mention pay-per-token which we use at work). Claude Code is generally a much better experience though.

Bandwidth doesn't tell us about how many clock cycles it's waiting for the data though (latency), does it?

For example with paging, I don't think those bandwidths make much difference and it's instead the latency which matters. Or am I wrong?

Not during office hours, no. Do you?

The interesting part is mostly comparing the nordic countries against each other. Sweden seemed to be doing much much worse early on due to the chosen strategy, but that no longer seems to be the case.

I won't be cancelling now as I seem to be averaging a bit more than 1000, meaning pricing will stay about the same since I should get the early adopter subscription.

However, the "average amount of searches" they're stating really can't be accurate for any tech savvy user who is using it professionally and privately. It really has to become cheaper, if it doesn't I sadly don't see myself using it long-term.

This is as accurate in Europe, and much of it has been for the last 15 years at the very least.

So, people should stop being humans? Are you also offended when you hear that the clothes you wear give off an impression?

I'm really surprised at how defensive some people in this thread are getting regarding this.

Not everyone follows the "netiquette" though. That doesn't mean it doesn't exist.

Most of these rules rang very true to me at least. For example, capitalizing letters in _informal_ chats does give off a formal tone. This doesn't mean no one does it, but it's definitely a thing.

So I posted this without finishing (probably managed to fat finger it somehow). I meant to say that I thought 2400 meant quite a bit more than 99.9%.

"People" on the planet? If that's accurate it's not super impressive, considering that probably 95-99% (or more) of people don't take chess very seriously. I used to be the top 99-99.5% of Counter-Strike players (_very_ roughly obviously), and that wasn't very impressive at all even though it's way past top 99.9% of "people on the planet".

(that being said,

I don't get what the point of that quiz is. All of the example countries in it instituted their mask mandates between the first and second wave. None of them did not?

He's doesn't seem very credible though. For example, there's a lot of very obvious reasons which could help explain this[1].

1: https://twitter.com/ianmSC/status/1481042778293358593

Is it really abusive if it's up to you to choose "Work at company X, make ~10% more but you have to be on-call for 1 week every 8 weeks" vs. "Work at company Y, make ~10% less, but never be on-call"?

Honestly, I'm not even sure how it's even remotely possible to draw a parallel between slavery and that. When negotiating a salary you'll have to take these things into account. Not worth it for the pay? Find another place to work. Jobs do grow on trees for people in our industry.

Most people use a laptop (or even tablet) for their everyday needs nowadays, and they likely don't want a heavy and hot gaming laptop for that.

>Much was made about the M1 Max taking the single thread performance crown; for less than a month.

This is quite misleading in multiple ways.

The M1 Max has the same single thread performance as the M1, which has been out for around a year.

You're also comparing Intel's latest desktop CPU to a laptop CPU. Before the M1, a laptop CPU competing with the highest-end consumer desktop CPU in multi-core OR single-core was unthinkable and had never happened. Yes, there have been leaks in which Alder Lake's mobile CPU beats the M1, but we don't know whether these benchmarks are representing the numbers we'll be seeing in consumer laptops with laptop cooling.

That being said, I hope there will be competition!

No, that's true, but it's definitely possible. I know of plenty of engineering students in my home country who came here from China (and India) to do their masters, and ended up staying for work.

But yeah, it definitely requires certain circumstances, and it's not realistic for everyone.

Well, you can..?