hachyderm.io is one of the many independent Mastodon servers you can use to participate in the fediverse.
Hachyderm is a safe space, LGBTQIA+ and BLM, primarily comprised of tech industry professionals world wide. Note that many non-user account types have restrictions - please see our About page.

Administered by:

Server stats:

8.9K
active users

Josiah Winslow

My friend @ConserveLetters told me he was experimenting with making generate . He got it to spit out things like "apple pie = peel a pip".

I then tried to use AutoGPT myself for the same task. I failed to get it to do almost anything on-task.

The closest gave me for SLOT MACHINES was SMELT CHAOS. Which is soooo close to an anagram...

But mostly, it either did nothing, got into loops, or went off-topic. At one point, it generated this image using while trying to brainstorm casino slogans.

What with all the hype around lately, I was kind of disappointed that I couldn't get some basic thing working.

Maybe I need to write better prompts. Or maybe isn't ready for this type of task.

But I at least expected to be more competent and able to learn.

But this begs the question: what would be needed to create a competent maker?

I'd imagine being able to give it SLOT MACHINES, and it'd spit out sensical results with relevant words, like CASH LOST IN ME, SLAM THE COINS, NO SLIM CHEATS, HALTS INCOMES, etc.

Whenever I make , I try to find related keywords or phrases first, and build a coherent phrase around that with the remaining letters.

I wonder how, or if, an can be trained to undergo a similar process, with a more specialized understanding of anagrams as a whole.

I'm almost tempted to use with or something, and a simple generator myself.

But it's hard to turn my human process for creating anagrams into an . There's so much subjectivity in making one, or even recognizing a good one when I see one.