Yeah, welcome to what the software engineering discipline has been grappling with, as this started for us a bit earlier. I've been coding for 22 years now, and in the current situation, I'm still quite employable as long as I can use the tools well; but, there just is no junior software engineer role in many, many companies now. Even our mid-level or early senior people aren't getting the same practice that I did when I was at the same point in my career.

It's unclear how this will turn out. Eventually people my age will retire, and it's the same story you wrote.

I do think it's pretty funny that Andreas is working on this project and Andreas is also leading Ladybird. That said, I would support Ladybird over this one at this point in time.

It definitely has. I've been using TOTP in the US as a second factor for years (maybe over a decade?) across websites and apps and the like. They almost never call it TOTP, though. It's usually "authenticator app" or something like that.

I store my TOTP secrets in Bitwarden for less-sensitive stuff or on my Yubikey directly for more-sensitive stuff.

If you set up your OS to use a compose key, then Compose + --- should be —. Back when I was on Windows, I used a nice app called WinCompose, but now I'm on Linux across the board and it's just built in. I believe it's built in to MacOS, but I'm not sure.

Yeah, that's correct; so, it's kind of a balancing act between wanting to preserve context for a new feature or refactor as it's being built out and wanting to minimize context window usage and therefore token usage. Other stuff like CLAUDE.md or AGENTS.md will also be ingested into the context automatically, so those are things to consider. Claude, for example, also has a /context command you can use to visualize your context usage.

Be careful. Someone got shot and killed for throwing popcorn at another moviegoer, and then the shooter was acquitted...

https://abcnews.com/US/cop-acq...

This, so much. I'm so tired of people complaining about the 467th CoD game while buying all 466 before it and then buying the 67th as well.

Such a strange question. It's like asking, "Is it okay to have red hair?" Skin color comes from genetics and may correlate with behaviors but does not cause them.

https://rufus.ie/

Mumble

Maybe you could get Winamp (or Wacup) working on Linux? :D

I think this is an example of something that I see far too often and which took me years to realize was even a thing: a person reading a question and interpreting it as some kind of attack or implication. It's exhausting.

Back when I got my most-recent car (2015), after I let the trial lapse, they called me every month for like 6 months, ignoring anything I said to them, until eventually I used a more-expressive tone of voice and improved my word choice. It was only after that that they stopped calling me. I can't stand that company.

I was going to comment this on another thread, but I'll just reinforce this here. As far as my definitions go:

Vibe Coding means you're interacting with the LLM completely in text and have very little or zero knowledge of the code it's actually building outside of whatever can be described or displayed in the conversation with the LLM.

LLM/AI-Assisted Coding means you're going back and forth with the LLM, sharing the responsibility for the code and its organization. One of you writes code, maybe the other refines it -- the human retains knowledge, understanding, and shared authorship of the system being built.

But, I do acknowledge that maybe some don't look at it like that. This is just how I differentiate those two terms and types of interaction.

How many points of failure do you have without a password manager?