LeoPanthera

Is anyone still making new optical drives?

Many TVs are advertised with features (even local features) that don't work until the first firmware update.

So plug it in, and update it once.

Then you can do all the things these other comments suggest.

Wristwatches are sort of unique in that they are technology that you literally wear, so it's important that they look and work exactly how you want.

TNG is a fine example of "competency porn", where everyone knows what they're doing and usually makes the best available choices. In today's world, that's even more compelling than it used to be.

This is partly why I'm not too worried about AI slopcode.

The human bar is not that high.

I would much rather have rigorous testing than a filter based on the origin of the code.

A script to automate this. https://gist.github.com/DavidCWGA/a5d09105371f734ccc7b0693a7...

Tailscale SSH is not enabled by default.

I'm stunned that "mapping out" blocks of bad RAM is a thing you can do. It reminds me of the old days when hard disks had to be "low-level formatted" first (ask your parents, kids) and it would ask you for a list of bad sectors to map out. Sometimes hard disks would even come with a list of known bad ones printed on the top.

This is a cool idea. I'm actually quite glad that AI generated stuff is still so very obvious. The world will be a darker place when it isn't.

You have to accept the cookie notice. If your user agent hides the notice there's probably no way to do that.

But that's not a bug. It's a design choice.

This sounds more like a story than a real issue. Do you have specific examples you can point to?

If this is real, it's going to kill people.

Mail does support it, but Google refuses to allow it. They used to, in the past.

There used to be a hologram art gallery in San Francisco where you could get a hologram of yourself created.

https://computerchronicles.tv/#episode/1610 (Video, skip to 11:47)

neverssl.com is slightly better for that, since it resists caching.

It seems they simply disabled a bunch of power limits. Probably just settings, rather than an actual code update.

I remember when the early GPTs started to appear people worried that the text it generated would be indistinguishable from human-written text.

I suppose it's quite pleasing that not only is AI text still obvious, but AI websites also jump out and are immediately recognizable.

Apple Wallet doesn't exist on non-Apple devices, so why should this?

That sentence has changed meaning twice, depending on what generation you are, it's either "snow", solid blue, or a screensaver.

n=1 is useless. The output is not deterministic.

It's one thing being willing to fund people who are trying to destroy the life of LGBT folks, but it's quite to another to advertise that fact proudly by saying that having a better web browser is worth it.

What a disgusting person this is.

A few months ago Claude nuked one of our databases. We keep backups, so, "no big deal", but since then we just put "Never delete files, use the the trash or rename aside" (or a wordier version of that) in the global CLAUDE.md.

This works slightly too well, in fact. Our Claude is absolutely paranoid of deleting files, including writing code that has the capability to delete things.

Frankly I'm OK with this.

I don't know why this was linked through archive.org. Although the original post was deleted, you can still read Reddit posts if you know the URL.

https://www.reddit.com/r/ClaudeAI/comments/1wl5cgo/code_just...

This is exactly what I do. If it's asking a lot of questions your spec isn't specified enough. It's very useful.

Wait, you can turn that back on? Where's the flag?

Early versions of plan mode would erase the chat history and present the plan as the start of a new session when you approved it.

I miss that. It worked really well, and it kept the context clean.

A lot of people are abandoning Mullvad because their CEO is directly funding a far-right political party.

As they should, IMHO.