froztbyte @ froztbyte @awful.systems

I'm @froztbyte more or less everywhere that matters

---

Read more

Posts

17

Comments

1,846

Joined

2 yr. ago

as a lot can be learned from trying to fix weird bugs

a truism, but not one I believe many of our esteemed promptfuckers could appreciate

citation

and here I was graciously giving the promptfuckers a choice

for anyone that finds this thread in the future: "check if vga@sopuli.xiz contributed to this codebase" is an easy hack for this test

since you’re married to not fucking getting it

.....tempting me to come up with a new word

thank you for your belief in me, that's so cool! you barely even know me! so cool to have people support people!

I'd still prefer concrete and precise technical detail though

many of the proponents of things in this field will propose/argue $x thing to be massively valuable for $x

thing is, that doesn't often work out

yes, there's some value in the tech for translation outcomes. to anyone even mildly online, "so are language teaching apps/sites using this?" is probably a very nearby question. and rightly so!

and then when you go digging into how that's going in practice, wow fuck damn doesn't that Glorious AI Future sheen just fall right off...

...wat

also, unless extraneous circumstance, please don't clip their display of abusive nonsense

I bet they'd just haaaate it to be on the internet, and yet it's exactly the kind of fingerprint of theirs that people should be able to find

on the one hand I feel for other people who'll maybe read this thread somewhen down the line

on the other, it's not exactly like I clipped words in my post

I moderately regret this post

because the counterposter in question went on to have some decidedly "fucking ugggggggh" posts

ah well. so we learn.

maybe train your model better! I know I know, they were already supposed to be taking over the world... alas...

jesus fuck how do you fail to understand any post of this kind this badly

RIP my hopes and dreams :<

like LLM like shithead

fuck, there's potential here, but a bit too specific for a t-shirt?

like llm like idiot

perhaps?

oh, I get it, you personally choose not to make these structurally-repeatable-by-foundation errors? you personally choose to be a Unique And Correct Snowflake?

wow shit damn, I sure want to read your eventual uni paper, see what kind of distinctly novel insight you've had to wrangle this domain!

(also I did my disclaimer at the start there, so, y'know (but also igwym))

woo! but still also check out pomsky, it's legit handy!

a'ight, sure bub, let's play

tell me what hw spec I need to deploy some kind of interactive user-facing prompt system backed by whatever favourite LLM/transformer-model you want to pick. idgaf if it's llama or qwen or some shit you've got brewing in your back shed - if it's on huggingface, fair game. here's the baselines:

• expected response latencies: human, or better
• expected topical coherence: mid-support capability or above
• expected correctness: at worst "I misunderstood $x" in the sense of "whoops, sorry, I thought you were asking about ${foo} but I answered about ${bar}"; i.e. actual, contextual, concrete contextual understanding

(so, basically, anything a competent L2 support engineer at some random ISP or whatever could do)

hit it, I'm waiting.

how dare I use someone's own words against them!