It’s referring to both. The recompiler links to the Zelda project and basically tells you “if you want to haven an example how to.proceed/what to implement yourself after the recompilation finished, you can use the Zelda project as an example”.

Well, usually those re-compilers or transpilers just translate the binary to some sort of intermediate language and then any backend should be able to compile it for your target system. So, in theory those handheld could be targeted. Problem with this project is that it’s not just “start transpiler, load rom, click go and your port is ready”. It’s more like "ok, here’s your game logic. Now implement the rest (or use several other projects and duct tape their libraries together to get what you want).

Nah, it just marks your question as duplicate.

I mean, the hosting company would be the likely target then and they’d probably lock your account and switch off the server. Depending on your nationality and that of the hoster, at least.

Do you know the term “trust thermocline”?

Basically it described a problem with the boiling the frog technique. There’s a point for every user at which they’re fed up with the bullshit, lose all trust in you(r company) and are hard to impossible to get back as a customer. Every customer leaving has a little unnoticeable effect on you, but with time there will be so many people that you lost that all your tactics to lock your users in will fail.

I read it as “no, we won’t use your data for advertising, but collect it anyways. If you ever dare to stop paying, we’ll retroactively process this data, too”

Yeah, basically you have three options:

1. Create and train your own LLM. This is hard and needs a huge amount of training data, hardware,…
2. Use one of the available models, e.g. GPT-4. Give it a special prompt with instructions and a pile of data to get fine tuned with. That’s way easier, but you need good training data and it’s still a medium to hard task.
3. Do variant 2, but don’t fine-tune the model and just provide a system prompt.

Aperture Science.
We do what we must
Because we can.
For the good of all of us.
Except the ones who are dead.

When I was a student, I did an internship in a chemistry lab. On one of the days, someone brought in some samples of skunk secretion for an analysis.

Everyone was like Not again i dont want that, let the intern do this!

I thought how bad could it be?. Turns out really bad. It days to stop that smell. And I mostly handled the sealed phials and only opened a single one for a gas chromatography without spilling something…

Windows doesn’t have sudo (not yet, at least) and privileges work a bit different as even as an administrator, you may not have full rights.

To overcome that obstacle, you’d need to run a shell as an administrator (hold CTRL+Shift, then use the start menu entry or right-click it and select run as administrator).

Next obstacle: We have a separate drive for each partition, but no root folder.

If we assume we’re running on a laptop or PC with a single drive and a single partition*, then it’s just

In cmd.exe:

del /F /S C:\

In Powershell:

Remove-Item -Recurse -Force -Path C:\

When you want to delete all (mounted) partitions/drives, you need to iterate over them. (Note that’s from the top of my head, didn’t check the script if it works).

In cmd.exe:

REM Not gonna do that, I'm no masochist

In Powershell:

Get-PSDrive -PSProvider FileSystem | Foreach-Object {
    Remove-Item -Recurse -Force -Path "$($_.Name):\"
}

Done. Mounting additional partitions before that is left as an exercise for the reader.

*note that even a standard installation of windows creates 3 partitions. One for the bootloader, one for the recovery system and then the system drive. Only the latter is mounted and will be deleted by this. The other two will still be intact.

The EU at least tries to regulate AI with the AI Act

It’s not a catch all, but some things are banned from using AI and others will need an assessment before going live. Some other will not need an assessment, but serious incidents need to be reported

…but it’s got electrolytes!

Uhmmm based on my behavior before I left, the engagement is probably “click the three dots, hit report, select spam and block user”. That worked at least for a short while before they got rid of that feature…

Relevant xkcd

Not quite. If they had overlooked a few accounts, they’d probably not even implemented that function. They’d just said "well, if you forget your password - or need to change it - you need to use the forgot password workflow that sends an email. Everyone without an email Adresse associated with their account would be SOL.

Since they implemented it, they are aware of such accounts. But since “providing freely any email address for a password reset” makes absolutely no sense, this should only work for this special case - accounts without an associated email address.

Whether it’s only done for unlocking accounts, whether this would have also worked when clicking on “Forgot Password” or whether this account lock and unlocking workflow might even be intentional to associate an email address to such accounts, is unknown (to me)

I think newer accounts can only be created, when providing an email address. There may be some old accounts that don’t have an email address associated. So, in most cases, you’d just be able to restore the account if you have access to both the account password and the email address. This breaks apart, if there’s no email address associated so I think they provided this way of recovery although it doesn’t improve security since it only applies to very few accounts?

My newest vps runs with Caddy. Works like a charm. The downside was, that I didn’t think of the automatic certificate deployment when I set everything up and it wouldn’t come up a first when I only wanted to connect locally to it, as it tried to get a certificate but the challenge failed because I hadn’t the firewall open yet. But besides that it was very smooth so far.

Amazon Deep Glacier is a lot cheaper for storage (but expensive for retrieval).

I use Archive Storage in Oracle Cloud S3 for my dr backups which is their equivalent of AWS deep glacier archive. It’s quite cheap, no restore fees, inbound traffic is free and outbound traffic is only paid, when you’re using more than 10TB per month. (Also first 10 GB of S3 storage is free)

Remember <marquee>? And maybe add some dancing hamsters?

Remember that JS file that rendered a text besides your mouse pointer and when you moved your mouse, the text would follow it letter by letter?