• 0 Posts
  • 19 Comments
Joined 3 years ago
cake
Cake day: June 30th, 2023

help-circle

  • I’m sure it’s all but officially confirmed, that all the people upload to it goes to training data unsanitised. And Microslop for sure uploaded entirety of their codebase to all the bots. This is at least part of the reason why claude is so surprisingly OK at reverse-engineering all the big stuff.


  • There is a saying in Russian that sounds something like “one little claw is stuck means the whole bird gets eaten”. In other words, slop, not even once.
    I have an old collegue that I meet about once a year. A year ago he was all about using it responsibly and stuff, this year we started arguing about merits of slop, and in the middle of conversation he opened a phone, asked a bot for more arguments, and read the response verbatim.



  • Now picture a different, more computer-curious plumber. He doesn’t hate computers, so he reads two pages on a wiki and now can at least spot when the url has changed or when the diff added some new step that wasn’t there before. You don’t need computer science degree to see that there is a new command that weren’t previously, and google that command.





  • I’m referring to the letter of his when he said that he’s not going to ban generated code, not gonna “wage a war” on it or whatever. It might sound like “devs are still accountable for the code they commit” and all is a good position, but like with a lot of the other shit, if you’re not explicitly ban 100% of it, you’re now in a position of arguing with bots about nonsense. You still can be tough and all, but you’re alone and the bots are numerous and nonsense is neverending.
    That’s what I mean not-anti-slop, and right now it looks like being all fence-sitting about this wouldn’t actually work. I predict that he’ll either will rethink the policy, or limit the amount of patches merged significantly, or lose his mind and will write a big rant and possibly does something stupid.





  • Google was at the beginning of the market to sell the data, it didn’t exist before. Obviously they didn’t try to scrape every ounce of it before they made (not them specifically, but they played a big role in it) a market to sell it on. The second it became profitable, they started doing it.
    You can theoretically be for-profit and don’t do anti-human shit, but that will be going directly against the only reason for-profit companies exist: maximising shareholder value. So those companies either don’t survive or change for the worse.






  • I fear that this all-or-nothing attitude brings more harm than good. If leads to the thinking that if nothing is perfect nothing should be done anyway. More than once people were saying to me that they’ve read that they can get your data anyway so there is no difference between a firefox with adblock on linux and google chrome on windows 11 with all the trackers active.
    And this is exactly the attitude that those who are selling our data want us to have.


  • To be honest, this is one of those times where I would say that it’s not that bad. Can’t expect a company to outright violate the laws of the country they’re registered in, you need to be friends with Trump to do that. If they keep their shit away from all the corpo vultures and only comply with direct subpoenas, it’s probably the best we can hope for.