Skip Navigation

User banner

just another dev

@ admin @lemmy.my-box.dev

Posts
2
Comments
415
Joined
2 yr. ago

  • They could, but adding random zero width characters into words would also destroy ever spell checker, giving it away immediately and making sure that even unaware people would filter it. Doing it outside the words would leave them with too few spots to use for proper watermarking.

    I think it's far more likely they'll use some kind of pattern in the tokens - that way the watermark will remain even when you don't copypaste it.

    But yeah, as said, they will never tell how it's implemented, but it can still be simply subverted.

  • Yeah, no chance they'd rely on something that would be so easy to defeat. Watermarking by using word patterns is far more likely.

    Still easy to defeat by just using another LLM to rephrase it though.

  • By that logic every news website is spam, because those also contain ads.

    I agree the article is without much merit. But calling it spam because it also appears in a book and it mentions that source, is just diluting the term.

  • This article could do with a Bottom Line Up Front. I got halfway through the page and I still had no idea what problem it was trying to solve by adding new problems.

  • Looked up her name on Twitter to see what people were saying about this

    I'm seriously wondering what your intentions were when you did that.

  • Better yet, just spin up your own instance, subscribe to all major communities, and have the servers push the comments to yours. No scraping required, and nobody will ever find out it was you.

    Statistically it's likely to have happened already.

  • My brother worked for such a Dutch company (ASM) and often got sent overseas to supervise the setting up of the production lines with these machines.

    He mentioned when he'd get sent to Asia, the workers would make sure to get it done over a weekend, while implementing the same setup would take 2 to 3 weeks in the US. In part that was due to the working conditions mentioned, but also simple lack of planning in case of the latter (things would grind down to a haalt because certain changes would need to be made, and the person responsible for the decision wouldn't respond for hours or days, etc).

    Side note: while 36 hour work weeks are common in the Netherlands, 40 hours is still the norm in my experience.

  • Legally? No idea. What might be adequate protection in the country your instance is hosted, is probably unenforceable in another country where a federated instance might be.

    Technically, you could try by using your own, self hosted instance, and not federating with others, so they won't be able to scrape your content as easily.

    But realistically speaking, your comments are possibly more likely to be scraped on Lemmy, since it's so much more open for bots, and your content is replicated to much more servers, not all of which may have noble intents.

  • I stopped using twitter a couple of years ago, so I fully agree that one is better off without it

    But when you reduce it to a nazi echochamber, don't you feel at least a teensy sense of irony?

  • I wonder which others he has tried to get to that conclusion, and how recently.

  • Yups. Report and move on.

  • They already are. They put all nsfw content behind a privacy paywall (pay with email and browsing habits). Luckily it can still be subverted through old.reddit.com - but the question is for how long.

  • Report these articles with "Business news, not tech news", and there's a high chance it'll be removed.

  • I'm probably a minority in this (although probably not so much here on Lemmy), but if anything, I'd want my TV to be less smart, and less personalised. I don't want Google to know what my favourite TV shows and movies are. I don't want "suggestions" on which streaming platforms I could also install (often before the content I would actually want to see). And I most definitely don't want my TV to be monitoring the rest of my "smart" home.

    For the people who are part of this articles titular "we", I seriously wonder: why would you have been waiting for this?

  • It's not a competition :(

  • One could also say that building a camera from first principles is a lot more work than entering a prompt in DALL-E, but using false equivalents isn't going up get us very far.

  • Also policing training would be completely unenforcable

    That's where laws would come in. Obviously it would have civil law, not criminal law, but making sure it would be enforceable would have to be part of such laws. For example, forcing model makers to disclose their training dataset in one way or another.

  • "yearly"?

  • Can you elaborate on that claim? I couldn't find anything substantial in the article.