[WoodenChair]

Ironically this post comes across to me as written by an LLM. The em-dashes, the prepositions, the "not this, that" lines. As a college instructor, I can usually tell. I put it through GPTZero and it said it's 96% LLM written. GPTZero is not full-proof but I think it's likely right on this one and I find it very ironic.

[bjacobel]

> Upgrading your CPU wasn’t a spec sheet exercise — it was transformative.

[raincole]

It uses even more em dashes than typical ChatGPT generated articles. At this point I think it's parody.

[nunez]

Pangram says 87%, and they are the gold standard right now.

[fabianholzer]

I find this knee-jerk reaction, that everything that shows certain stylistic choices is put under the suspicion of "might be generated", to have become a tedious cliché.

Unless the author had live-streamed the writing process, how could we know? Humans have been exposed to LLM generated texts on nearly all channels for more than three years, so by now it would be a surprise if there had not been a reciprocal reaction. Writers imitate what they read, the tools we shaped might have started to shape us.

[stagg]

I feel like if you converse with an LLM all day, you end up writing like one.

[shminge]

fwiw, it's 'foolproof' not 'full-proof'

[grodriguez100]

That’s to make it clear it was written by a human and not by an AI :)

[wdporter]

I think he might mean it does not fully prove anything. Is the dash an awkward attempt at a pun?

[Eliezer]

Came here to say the same.

[mirrorlake]

I feel like I'm being rickrolled over and over again by infomercial-grade slop.