As others have mentioned, adblockers seem to break it somehow. It’s probably not intentional. The page doesn’t serve any (actual) ads, so it’s safe to try disabling your ad blocker. For me, refreshing the page a couple of times also fixed the issue even with uBO enabled.

Happened to me too, had to reload a couple of times. Not sure if it was intentional, but realistic anyhow xD The only thing missing is random loading times between each step.

I might be guilty of using an argument like that sometimes. What I really mean is that I currently value the convenience more than my privacy. “I’m using service X because that seems the most convenient choice for me and I don’t really care whether someone I won’t ever meet is reading my data. But I could also use service Y if it’s just as good and convenient as service X.”

Point being, I’m always willing to start using a different thing as long as it works for my use case as well as (or preferably better than) the thing I’m currently using.

So, instead of trying to scare me about someone spying on me, try convincing me how the other service works better than the one I’m currently using.

I think Vim is great if you use it everyday and know it inside out. For those of us who don’t, Nano is just easier, not least because it tells you in the UI how to exit it xD

I think it’s written that way because the instructions are meant to work on Windows too. But it still doesn’t make much sense. I think just saying “Install CMake” and letting the agent figure it out for the current environment would be a much better approach.

Also, I think the model they were using was working with outdated data. Why specifically require Java 21 when Java 25 is the latest LTS? And furthermore, why even link to a specific JDK implementation when it surely works with almost any implementation? Probably just makes the agent install the Temurin implementation even if the system already had a working java.

My point being if the author’s idea is to let AI handle the installation to avoid writing environment-specific instructions, it shouldn’t then have these hardcoded assumptions about the environment. So yeah, it seems they (and the AI that wrote it) didn’t really know what they were doing.

Yeah, but I would never let an AI agent run commands on my machine without verifying each command. And at that point I would just rather copy and paste commands by hand, takes about the same time.

If the installation for example required creating some complicated host specific config files, then sure, I could see how that kind of a “runbook” could maybe be useful. But that doesn’t seem to be the case here.

It’s real

https://github.com/0xeb/ghidrasql#install-with-an-ai-agent-recommended

Haha, reading some of the responses in this thread actually reminded me of that xkcd but I couldn’t immediately find it. Thanks!

Oh wow, things really have gotten bad over there. For me personally, the much greater risk is that I forget my phone somewhere or someone steals it and in that scenario a weak password is the larger issue.

It seems there isn’t a single correct answer here. The threat model is different for everyone.

Yes, but that’s besides the point. If the convenient options for a normie user are

• Having a weak password
• Having a strong password and a fingerprint

Out of those the fingerprint with a strong password is way better option, imo.

In the USA, they can legally force you to unlock a device using biometrics.

Also, how does that work? Can’t they legally force you to enter your password too? Or can you claim you don’t remember it? If that works, can’t you just have a band-aid on your finger or something? Surely they cannot force you to take it off and risk getting an infection on the large wound you just happened to get yesterday…?

I think it depends. Some people might be inclined to use too short or simple password because they don’t wanna constantly be typing a long password. It’s much better to have a strong password and fingerprint/face rec for convince than just a insecure password.

I run Debian on my homelab and manage Ubuntu Servers at work. I haven’t seen any practical differences between the two. Sometimes you might get a bit newer version of something from the Ubuntu repositories than from the Debian ones but not even once has that been an issue and most of stuff I use come from 3rd party repositories anyway.

I think the whole thing is a bit clickbaity. The only reason it works is that they specifically configured the AI workflow to have access to both the public and the private repositories. The solution? Don’t do that. The only thing GitHub could fix here is disallowing such an obviously unsafe configuration.

I’m from Finland and usually the subtitles for English content are good.

One challenging example that comes to mind is Baby Shower. Since this concept is mainly from the US, there isn’t an equivalent term in many languages. People here usually just use its English term, but I’ve seen subtitlers sometimes translate it as vauvasuihku, which literally means “baby shower”, but no one is going to understand that term.

In one series, there was the phrase “What is this, my shower?”, when a pregnant woman was asked if she had already chosen a name for the baby. The subtitler had translated this literally as Mikä on tämä suihkuni? (≈What is this shower of mine?), completely losing the context.

I think most newbies would rather find the answer themselves in the wiki or old forum threads, but that’s often difficult if you don’t know what exactly you’re looking for.

Running with your example, let’s say I’m trying to find out the color of the sky but don’t know it’s called sky. And to make it worse, right now it’s covered by some kind of gray mass… Is that perhaps The Cloud I’ve heard people talking about? I would Google something like “huge thing above color” but unsurprisingly I wouldn’t end up on your wiki. So I end up asking the question on the forums instead.

I used to work in IT support, where 95% of the questions were about things that were already comprehensively documented if people would just read. Instead of yelling at customers for asking dumb questions, we had response templates we could send with just a single click.

I don’t understand why communities overwhelmed by repeat questions don’t do something similar. The next time someone asks about the sky color, it would take just one click to reply politely with a link to that wiki article and everyone would be happy.

It works great when you log in as a user like you normally would on a Windows pc. But on a server you want it to run as a service, starting automatically in the background and/or being managed by the failover cluster without requiring a user to log in.

The solution back then was to use netplwiz autologon. Obviously a hack and bad for security, but fine for a homelab. After googling it now, it looks like some other “solutions” are documented in the GitHub issue about this.

Asking out of curiosity: is there a specific reason you’re running Windows on your server? I used to do the same on my home servers because that’s what we used to have at work, and I wanted to learn and test some stuff. But it was a difficult road, to put it mildly. Simple things, like getting Docker autostart on boot seemed almost impossible. At some point I just gave up and switched to Linux.

Caddy is not mentioned, I take it it’s not affected?

Never ever add any users to the docker group. Rootless mode is cool tho (albeit with some caveats)

Other users cannot. Your instance admin can obviously see it if they want. Then, it depends on the settings of your instance if all media downloads from other instances are proxied through your instance (hiding your IP from other instance admins) or not. I’m not sure if it’s enabled by default or not.