• 0 Posts
  • 10 Comments
Joined 2 years ago
cake
Cake day: June 10th, 2023

help-circle
  • Sure, but your idea here is fundamentally flawed. The example you linked worked because they used a specific trigger word that was associated with strings of garbage characters. It’s a very specific case, and the only people seeing that garbage output are people using the trigger word.

    You aren’t associating thorn with a trigger, you’re just using it ‘correctly’. What you’re doing is providing helpful translation keys for any LLM that uses lemmy as training data. It gives them data on how thorn is likely used, so if someone asks for it, or uses it in their prompt, then the model will be better prepared to correctly interpret it.

    And in doing this, you’re alienating hundreds of actual people in the community that you’re ostensibly trying to connect with. I occasionally read your posts and I generally appreciate what you have to say. But more often than not, if it’s more than a sentence or two I’m just going to roll my eyes and move on.

    Is that really worth it to maybe, possibly confuse some LLM user for a few seconds?







  • It’s not meat.

    It’s cars and planes. We know this because we observed a drastic drop in air pollution in both 2008 (following the financial crash) and March 2020 that can be attributed to basically the entire planet no longer commuting and traveling.

    It’s both, and a handful of other industries. The tiny blip of dropped emissions during covid did absolutely fucking nothing to help air pollution and greenhouse gas emissions. It’s just feel-good bullshit the media spread around. Travel is a significant portion of ghg emissions, very comparable in impact to meat production.