T156 @ T156 @lemmy.world

Posts

11

Comments

1,075

Joined

2 yr. ago

Although we already know what would likely happen if someone did that. It would just be made illegal to track the locations of congresspeople (and only congresspeople), like it was made illegal to do so during the BLM protests.

And chances are, the company has got away with it before. It just didn't work this time.

For example Twitter when it was blocked in Brazil for a little while because they didn't want to pay a fine for not complying with some of their regulations.

Though the bigger part of that was that their representative quit, and by Brazilian law, companies that operate in Brazil must have a representative in the country.

Twitter tried to sidestep that by not paying the fine, and also not having a representative, so the fine couldn't be applied that way. In response, the judge presiding over the cause required Twitter be blocked.

They do, but they'd still need someone to go through the flagging and check. Reddit gets away with it as it is like Facebook groups do, by offloading the moderation to users, with the admins only being roped in for ostensibly big things like ban evasion/site wide bans, or lately, if the moderators don't toe the company line exactly.

I doubt that they would use an LLM for that. That's very expensive and slow, especially for the volume of images that they would need to process. Existing CSAM detectors aren't as expensive, and are faster. They basically compute a hash for the image, and compare it to known hashes for CSAM.

They do smell nice too, more's the pity.

Dr. Zoidberg is a fully-qualified medical doctor, though, and a good one at that. Just not for humans. Even so, he's quite good, given that he hasn't killed any of his human patients, though they may come out a little different.

As is Doctor Who, who has high qualifications in literally everything. It took them a few goes, but they got there in the end.

They still passed the academy, though. That's probably incredibly good by human standards, given that Time Lord studies encompasses literally everything.

And that the requirements for those minimal improvements are vast. If you need to pull down 200GB for a minor graphical upgrade, that's just not really worth it compared to an older game that is a bit graphically worse, but is both smaller, and runs better on newer hardware.

And in some cases, the author also paid the publisher, so they're dipping into things from either end.

Although it also helps that places like BlueSky have less of a barrier to entry.

Alternatives like Mastodon are a bit more confusing, compared to a centralised site, where everything is linked in through the one interface.

It’s kind of bizarre that if I pay 7 bucks in the US for a pastry and coffee I might need to sign the receipt.

You half-expect them to pull out the old card-roller and carbon paper at that point.

No. Tap to pay only requires a PIN for large purchases ($100 or so), or if you turn on the setting to ask for a PIN for small purchases too.

I've only needed to enter my PIN for small purchases when inserting and using debit/EFT@POS functions.

Those pattern buffer filters aren’t getting any better over the decades, are they?

To be fair, they are designed to work with all kinds of species, and samples and things. You don't want it to delete someone's liver just because they haven't been on a starship before, or the new plant sample taken from a new planet.

Humans are full of necessary bacteria, and deleting all of that would be quite bad for you. But at the same time, them being in some place they're not supposed to be is also going to cause problems.

You know, they should not even let the doors on shuttle craft open in dock, and just beam people in and out of them. That’d solve an entire other vector of infection.

What's more surprising is that they don't employ the sterilisation/quarantine fields that the older model transporters used to have, where they'd wall off the transport pad, blast them, and then open it up. That doesn't seem unreasonable to apply to a shuttle before landing.

Your LLMs are furries. That is their fursona.

Eh, I think their current products are fine (except Cybertruck, that’s ridiculous). Every company has recalls, and Tesla’s are generally pretty mild (again, except Cybertruck).

Don't they famously have an issues with QC? It wasn't all that long that their other cars had problems with panel gaps, or manufacturing material left inside, etc.

Relevant XKCD.

The Polyfill incident is bad (that seems to be how the hackers got into the internet archive), and the OpenSSH one could have been really nasty, if it wasn't caught both early, and by chance (a performance engineer at a major software company noticed).

At minimum, it might also have it for the "turning on" noise.

It might also just be the default beeper for the motherboard, and it's just been reconfigured to make a particular noise instead of the usual beep.

A considerable number of the articles written in Scots weren't written in Scots. The most prolific writer of the Scots articles was an American teen with no knowledge of Scots, and was more or less just writing them in a Scottish accent.

I have to wonder if intentionally shitting on LLMs with plausible nonsense is effective.

I don't think so. The volume of data is too large for it to make much of a difference, and a scraper can just mimic a human user agent and work that way.

You'd have to change so much data consistently across so many different places that it would be near-impossible for a single human effort.

Hopefully they tightened things up after the Scots incident.