infinitesunrise

@infinitesunrise@slrpnk.net

• ⁨Comment⁩ on ⁨AI crawlers cause Wikimedia Commons bandwidth demands to surge 50%.⁩ ⁨⁨1⁩ ⁨day⁩ ago⁩:

  This is a cool use case. Just make sure you retain your own voice! If you read an AI-generated sentence out loud and think “I’d have said it this way instead”, you should absolutely then change it to be that way.

• ⁨Comment⁩ on ⁨AI crawlers cause Wikimedia Commons bandwidth demands to surge 50%.⁩ ⁨⁨1⁩ ⁨day⁩ ago⁩:

  AI has niches but they’re exactly that: Niches. Small duct tape tasks for fudging over “hard problems” where manual code would result in a worse outcome and take far more time. Little esoteric problem spaces, which notably don’t actually require you to use several states worth of electrical power training on a 50PB dataset of anime titties.

  An example: I have a name generator in my game that strings together several consonant+vowel phoneme pairs into a name. This means that the names are always pronounceable, but often the spelling looks really unintuitive. Eg Joosiffe, which the player would likely pronounce as Joseph. However, the leap we do in our head between those two spellings is a process of declassifying phonemes and then re-classifying phonemes, and is actually a “hard problem” from a coding perspective due to the unintituive, multifarious complexities of written, spoken, and conceptualized human language. Adding this step to my name generator in code would be a project of it’s own, larger than the game itself, and wouldn’t ever work nearly as well as it needed to. But relatively small (30MB) AI models that do this with something like 99.8% satisfaction already exist. They didn’t require a data center’s worth of resources to train, and since they’re academic projects they have licenses that allow them to be used for free in a game.

• ⁨Comment⁩ on ⁨Mozilla Thunderbird Challenges Gmail With Its Own Email Service⁩ ⁨⁨1⁩ ⁨day⁩ ago⁩:

  You can’t know that with absolute certainty. Sorry, but if you’re using someone elses server for your communications and they’re not end to end encrypted, you should just assume that they can and do read your emails.

• ⁨Comment⁩ on ⁨Real⁩ ⁨⁨1⁩ ⁨week⁩ ago⁩:

  Make a Windows install USB but instead of putting the Windows installer on it put a Linux installer on it /s

• ⁨Comment⁩ on ⁨Millennials are watching monthly student loan payments skyrocket from $500 to $5000 under Trump: This will ‘crash the economy’⁩ ⁨⁨2⁩ ⁨weeks⁩ ago⁩:

  I used to have apparently atrocious credit due to delaying payments on my student loans for years. But with proof of income it didn’t stop me from getting apartments in NYC. In the last place I asked athe broker what my credit score was and while he wasn’t at liberty to tell me he did say “Not good. But it’s all student debt related which we don’t consider relevant”. Still seems weird to me today but I guess landlords often don’t consider student debt to be a reflection of a tenant’s ability to pay rent. Probably because most people prioritize paying for shelter over paying for the classes in their past.

• ⁨Comment⁩ on ⁨‘The tyranny of apps’: those without smartphones are unfairly penalised, say campaigners⁩ ⁨⁨5⁩ ⁨weeks⁩ ago⁩:

  I missed that phase because I was using Netscape.

• ⁨Comment⁩ on ⁨Bad UX is keeping the majority of people away from Lemmy⁩ ⁨⁨1⁩ ⁨month⁩ ago⁩:

  Although, I think the answer to the barrier to entry is to be less concerned with making federated services feel like centralized apps, more concerned with rebranding server select as the advantage that it actually is. Educate those people.

• ⁨Comment⁩ on ⁨Amazon Will Spend Nearly A Year Of AWS Revenue On AI Investments⁩ ⁨⁨1⁩ ⁨month⁩ ago⁩:

  I love subsidizing these jackoff’s energy bills in my own power bill /s. 17c/kwh for me, 3c/kwh for Bezos…