@333software "the real limit is your hardware".
I don't get it. Isn't that true for every software ever made? Every limitation we put on software is because we don't have infinite computing power.
@adampatricknc@Chris34259480@FrankRundatz@GaryMarcus They have no vision capabilities. The agents/products that use LLMs process the images utilizing other tools, transforming the images into tokens the LLMs can understand. The LLMs themselves are functions that receive and spill out tokens, the "vision" comes from another tool.
@Davidmdrpi@AlexanderKalian Or they could also have some infrastructure problem running models so big. One thing I can assure you: they do NOT care about "impacts" of text generators. Or at least do not care about that more than they care about making money.
@Davidmdrpi@AlexanderKalian Bullshit. Do you really think the companies deeply care about the negative impacts of powerful LLMs? My bet is that they didn't release those more powerful models yet to generate hype while competitors play catch up with current market-available models.
@Oreen_77@BernardLekaj@JamianGerard Yes. Wasn't that the whole point? Art does not need to be judged "good" to be called art. Art can be annoying, unskilled, displeasing, ugly and evil and it's still art. The robot, in my view, cannot make art.
@profclaudia_ap@marinahelenabr Sim. Escravidão já foi lei e nem por isso era certo. O que a lei (ou um juiz à sua própria vontade, sem precisar de base legal) diz não é métrica pra dizer o que é certo ou errado.
@JLima11111 @Antonysccpssj @leiatheinvestor Ao contrário, eles precisam ter 20% do dinheiro deles guardado. Então se um banco tem 8 reais emprestado por aí, ele precisa ter 2 reais guardados.