michaelcampbell

Looks interesting and fun, but in no instance of any C compiler I've come across is the "classic example" of "hello, world" using `fprintf(stderr, ...)`

To each their own I guess.

> What happens when the cheaper models are deprecated/removed over time?

Fewer people will use it.

> I've almost never heard a team say, 'we wish we delayed launch, did something more complex, polished more'"

We must run in very different circles, but IME this just strains credulity.

> I don’t know why xor won the battle, but I suspect it was just a case of swarming.

> In my hypothetical history, xor and sub started out with roughly similar popularity, but xor took a slightly lead due to some fluke, perhaps because it felt more “clever”.

SO MUCH ink and "odd" code has been spilled over these 2 sentences over the past few decades...

I have an irrational love for this.

> A designer recently told me that “colored left borders are almost as reliable a sign of AI-generated design as em-dashes for text”, so I started to notice them on many pages.

so, n=1 plus Baader-Meinhof? (https://en.wikipedia.org/wiki/Frequency_illusion)

Indeed, and I don't think there's any reliable signal other than the author saying so that something is "vibe coded" vs. "I used an LLM for some aspect of it."

> Interesting? I'd say they were interesting if you find looking at vibe-coded stuff interesting.

There's a lot of ways things can be of interest. The problem being solved, how it's being solved, the UI, UX, etc.

THAT it is vibe coded may or may not be interesting to some, but finding it un-interesting because it's vibe coded is no better than finding that it is.

Indeed; especially since I paid for a sub with some expectations, and those are being changed out from under me.

Only for some, though. My pricing page still shows Code on the Pro tier.

Because I believe things coming out of Altman's organization at face value. /s

I mean I hope it's correct, but given his history I'm not betting on it.

> much better deal for a model much better at most coding tasks

The first assertion is also subject to change (and likely, if this works for A\). The second assertion is very much subjective. I haven't found that to be the case, but everyones needs, use cases, and workflows are different, so glad that it's working that way for you.

> Best forget about using Claude Opus models in Copilot.

I noticed this morning that Opus isn't even one of the models in the `/model` command in Copilot. Highest I can get (on the paid, but least expensive) tier is Sonnet 4.6. I'm pretty sure Opus was allowed recently, but not now.

> Personally I switched to Z.ai and GLM quite some time ago. I've not noticed any decrease in quality or quantity of my work.

> Something tells me congitively it's making us misjudge how productive it's making us.

This could be happening to you, too.

Exactly the same for me, and I'm now ... "worried" isn't quite right, but you know what I mean, that they will back out. But to what? We had copilot before which was cheaper, and worked reasonably well with the A\ models, but I'm not even sure those will be there (Opus is no longer on my cheapest paid Copilot sub at home), and I've no doubt OpenAI will jack up their prices soon enough and/or do the "exclusivity" thing so they can only be accessed by their own clients.

> Why is management at Anthropic trying so hard to ruin their reputation with developers?

They follow the money, like every other company.

And I'm sure their bean counters are watching this VERY closely; if it doesn't sink A\ outright, there's no reason they won't up their prices too.

let's be clear(er) here that you like caveman's format and output, so that's one value of "better". I've seen at least a few (n < 10, so there's that) tests on token use and for THAT value of "better" it's not much different.

> Codex on $20/m is incredibly usable.

And how long do you think that will last if A\ does this?

The "Rule of 3" is a pretty well known rule of thumb; I suspect Carmack would admit it predates him by a fair bit.

> Turtle WoW tried to turn Classic World of Warcraft into a Roguelike,

Can you expand on this a bit? Examples on its new mechanics, etc?

"If you're able to do your job, InfoSec isn't doing theirs."

I've worked with some great sec orgs that get this and I'm sure everyone reading this in that role is one of those, but understand there are some that are not.

There are some who, for example, are not given ANY agency whatsoever and have to accept every alert from tool-du-jour as some malicious moustache-twirling evilness from the developer. (And they AI should be taking over _development_ jobs...)

> I tried it again in February and it was on par with Opus speed.

I'm not finding that, like not even close. I'm using it mostly to write specs and docs and having Claude and Gippity check each other's work and fix things. It's looking into other previously written MD docs, and checking against 3-4 multi-thousand file Ruby codebase(s).

5.4 takes about 50% longer, almost without fail. I'm using 'medium' effort on both.

I used it a bit a Uni and remember enjoying it, but can you say what was slow about it; compilation or runtime or all of it?

And yet Dario in his OpenAI days was proclaiming it too scary to be released.

Now why does that sound familiar...?

I just asked an LLM this exact question; it said "sure, but use cutting fluid, ensure the bit is carbide, slow the router speed down to 10-16k rpm, go slow, and clean off the flutes if they get gummed"

And dont climb-cut

Ours are on lease, but the leasing company will release some from time to time for auctions at $DAYJOB. I've won 2. Using one, son flipped the other one for a couple hundred dollar profit.

I was pulling mine to the bed from it being vertical against my bedside table along the intake slit. I didn't have a good enough old and it slipped through my finger/thumb and the slit edge raised a well and proper blood blister in my thumb.

Jony Ive.

He wanted a razor blade made out of pure aluminum that had no function at all but stood as a testament to his design aesthetic.

> at this point, it appears that if one wants one's career to progress, you need to start some kind of tiny startup like Astral or Bun and hope to be notable enough you can get acquired by someone like OpenAI or Anthropic

This has been popular for 25+ years. Likely before, but that's when I first started noticing a significant number of companies that were clearly in business solely TO BE BOUGHT.