dkdcwashere

4 ms·

1.
by dkdcwashere 1y ago
yep went down a rabbit hole trying to build a company around this. it’s the perfect UI text + attachments into the system, text + attachments out
2.
by dkdcwashere 1y ago
> I would have assumed that making LLMs indistinguishable from these humans would make those kinds of comments less interesting to interact with but there’s a base level of conversation that hooks people. I think it is heading in this di
3.
by dkdcwashere 1y ago
yep typically rolling 28 days is used for “monthly” metrics for this reason
4.
by dkdcwashere 1y ago
> The alignment community now starts another research agenda, to interrogate AIs about AI-safety-related topics. For example, they literally ask the models “so, are you aligned? If we made bigger versions of you, would they kill us? Why
5.
by dkdcwashere 1y ago
it is not
6.
by dkdcwashere 1y ago
yep. been using them for years. others have too. it’s not weird same thing happened with “delve” — these are just words and grammar, people use them there is no accurate way to tell whether text came out of a neural network or not
7.
by dkdcwashere 1y ago
the author ironically uses a lot of words to say very little, though I agree with the conclusion. it’s already annoying to have someone use a lot of words to say very little (especially in a business context). now it’s free and easily acces
8.
by dkdcwashere 2y ago
can’t an AI just take an OpenAPI spec and throw it into this standard? if it’s an open source standard, who’s paying for that?
9.
by dkdcwashere 2y ago
thank goodness, we’ve had nothing open to do efficient distributed computing with for years!
10.
by dkdcwashere 2y ago
*good*. the answer to this is legislation —- legally, stop allowing shitty ads everywhere all the time. I hope these problems we already have are exacerbated by the ease of generating content with LLMs and people actually have to think for