conradludgate

I don't think I'll ever retire, but I'll take jobs knowing that the salary won't matter for me which unlocks a lot of options for me again

> This is a fun standalone project / proposal

Rely on it? Not that I can remember.

However, Rust makes use of it for fast safety checks. Because rust strings must be valid utf8, if you want to take a substring at some range, eg "Hello, World!"[7..12] then it's very simple to just check bytes 7 and 12 and see if they are the start of a codepoint, no other scanning or parsing is required.

It's usually the same site that keeps asking in my experience - they don't bother to remember your preferences

This is what I use to log into our remote desktop software at work. The passkey is on my phone and I scan a QR code presented by the remote desktop software (in this case it's the Windows App on my macbook and Microsoft Authenticator on my android).

The password integration breaks because it's trying to detect password fields rather and failing because of weird HTML/JS code.

The passkey integration goes the other way, which is much more reliable.

That's literally not what happened.

Yes, these are the only two options. That is definitely true.

It's of course impossible to have any nuance or middle ground here where you use LLMs to assist while you still focus on the engineering design decisions and the quality.

Likewise, I've found many bugs using LLMs in the small software libraries I hand wrote and considered complete and correct - even though it was my sole focus at the time and I thought I had perfected it

she*

For personal use, Luna-xhigh on the Plus subscription. I only ever hit my limits if I decide to use Sol or Astra for a bit more quality.

For work, given we pay API pricing anyway, I've been happy experimenting with K3-medium as my default, and now I'm planning on trying GLM 5.3 as well. Sol-medium is my fallback for my work purposes but I occasionally use Opus 4.8 as an additional reviewer.

I like the open weight models because much of my time at work is spent on security hardening (specifically, hardening my own service), and Sol/Opus keep snitching on me and blocking my prompts.

The slop is what upsets me the most right now.

Some people feel the need to attach an image to almost everything the post - even in chat rooms (including slack at work). These images serve no purpose, but the poster feels it is useful - though it was often reaction gifs in the past I'm seeing a lot more AI content than I ever saw reaction gifs, maybe because of novelty or maybe because it's ultra personalised.

Sometimes images help to visualise something or to get a point across, but I see so many people who think it's necessary to reply to a discord message with a cat with human limbs doing a dance, or a photo of "themselves" climbing a mountain with the Rust logo to show them mastering Rust... Ok?

Image models are useful and I'm thankful for much better visual reasoning but it really really frustrates me the constant need to burn money for all of this slop.

[flagged]

Nowadays it's the anonymous accounts that are generally most likely real, and the "real name" accounts with a human profile picture are most likely mass generated. You just use a name generator and pull profile pictures from Google images and you have a real user profile

I think a key difference is that most programming careers, the artistic outlet (code) is not what is being sold to customers. The customer might feel something if the app works fast, but they won't feel anything about the fun bit packing and struct-of-array design you did for personal enjoyment