Hacker News Reader

Comment by causality0 | Hacker News Reader

GPT-3-generated writing is a combination of comically absurd and incredibly deep. The weak infants of the woman next door who had to be picked up with the tongs and thrown into the dustbin is quite silly but then there's He wanted to crawl away from it, but there was no place to go. which is a surprisingly deep statement about the revulsion he feels towards himself. That a computer can generate that says very uncomfortable things about the skill of humans in recognizing genuine creative significance.

View on HN

> which is a surprisingly deep statement

People used to say the same thing about Markov chains, and then the hype died down it took 20 years for another hypetrain to pull into the station. The same will happen with GPT-n.

> says very uncomfortable things about the skill of humans in recognizing genuine creative significance.

Indeed. People see pictures in clouds and omnipotent beings in the stars. Humans are uncomfortably good at finding meaning in nothings.

What applications are Markov chains used in now?

Kubernetes uses them to come up with random generated cluster names if you don't provide a name of your own..I bet many other applications use them in a similar manner

Crypto currencies and decentralized protocols like the libp2p (used in IPFS) like Markov chains iirc

Sequence analysis in Bioinformatics, for example to align sequences and detect homology and as in HMMER: http://hmmer.org/

And other people reward those findings with money. Good writing is hard. Writing that evokes emotions in others is even harder.

It is extraordinary rare to ear money by writing. It may be hard, but people dont reward it with cash.

Exactly. Getting people to see patterns is easy. Getting people to fork over money based on those perceptions is much more difficult. Ask any church plant.

The limitations of Markov chains are obvious within hours of starting to use them. GPT-3, less-so.

My biggest problem with people the world round is invariably rooted in this sense of ownership of what other people find meaningful. This kind of top-down dictation of ultimately subjective meaning is what is uncomfortable (because meaning is a form of tacit knowledge and therefore incommunicable) and is the path to inquisitions, woke / oracular authoritarian thought-policing regimes with reeducation camps.

Deep statements are deep because the communicate feeling, if the machine never felt anything but tripped over meaning like this (and it's pretty obvious it's just gluing things together) it really dulls the effect.

Who’s to say that our greatest writers didn’t “trip over meaning”? A lot of poetic processes involve finding, curating and re-contextualizing interesting phrases (see found poetry), and there’s potentially a lot of value in applying the same process to computer-generated writing.

That's why no one is truly moved by a single phrase. We often quote memorable bits from larger works ("to be, or not to be"), but that quote is only memorable because the entire work makes us believe that it is not an accident.

I was by a famous 6 word poem: "For sale: Baby shoes. Never worn."

Sure, there are exceptions. And for those exceptions it may be fair to say that perhaps they were just a lucky stumble.

Even here I do wonder how well this poem would be remembered if it weren't associated with Hemingway.

Yeah. But the thing is, after you have own baby, you may find yourself with like 5 pairs of never worn baby shoes.

So while it somewhat suggest death, it is only because of the literally context, which makes you expect something deep. If you seen that in an ad, you would not feel the same.

I am wondering if we call some writers great because of what they did not, could not or cannot express (anymore). By their limited and graspable existence they serve as targets for our projections, whereas an AI is less attractive for that purpose, at least in my view.

I'd rather think about the authors of AIs, their assumptions and their worldview.

Today at least all the gpt3 we read is still hand selected by another human, so if it communicated feeling in the editor, we get to read it.

Is it "surprisingly deep" though? Looks like two standard cliches of human writing often used in close proximity and the wider context suggests literal motion (which is also present in the silly line about the weak infants!)

I did like "fuzzy belly" tbf, and Kafka himself might have appreciated the purely serendipitous irony of a line about "the bugs his father used to bring him as a little boy"

I was actually taken in by the infants-in-the-dustbin image. I thought it was gruesome, then I reflected on how infant mortality has decreased and what a better time we're living in. This article was uncanny valley for me.

Writers use all kinds of tools to "cheat", up to and including not actually writing. Mining computer-generated nonsense or coherent-but-empty prose for the odd accidentally-excellent phrase or sentence might end up being just another. Hell, it may already be.

Reply on news.ycombinator.com