←back to thread

LLM Inevitabilism

(tomrenner.com)
1664 points SwoopsFromAbove | 3 comments | | HN request time: 0.684s | source
Show context
lsy ◴[] No.44568114[source]
I think two things can be true simultaneously:

1. LLMs are a new technology and it's hard to put the genie back in the bottle with that. It's difficult to imagine a future where they don't continue to exist in some form, with all the timesaving benefits and social issues that come with them.

2. Almost three years in, companies investing in LLMs have not yet discovered a business model that justifies the massive expenditure of training and hosting them, the majority of consumer usage is at the free tier, the industry is seeing the first signs of pulling back investments, and model capabilities are plateauing at a level where most people agree that the output is trite and unpleasant to consume.

There are many technologies that have seemed inevitable and seen retreats under the lack of commensurate business return (the supersonic jetliner), and several that seemed poised to displace both old tech and labor but have settled into specific use cases (the microwave oven). Given the lack of a sufficiently profitable business model, it feels as likely as not that LLMs settle somewhere a little less remarkable, and hopefully less annoying, than today's almost universally disliked attempts to cram it everywhere.

replies(26): >>44568145 #>>44568416 #>>44568799 #>>44569151 #>>44569734 #>>44570520 #>>44570663 #>>44570711 #>>44570870 #>>44571050 #>>44571189 #>>44571513 #>>44571570 #>>44572142 #>>44572326 #>>44572360 #>>44572627 #>>44572898 #>>44573137 #>>44573370 #>>44573406 #>>44574774 #>>44575820 #>>44577486 #>>44577751 #>>44577911 #
brokencode ◴[] No.44572627[source]
> “most people agree that the output is trite and unpleasant to consume”

That is a such a wild claim. People like the output of LLMs so much that ChatGPT is the fastest growing app ever. It and other AI apps like Perplexity are now beginning to challenge Google’s search dominance.

Sure, probably not a lot of people would go out and buy a novel or collection of poetry written by ChatGPT. But that doesn’t mean the output is unpleasant to consume. It pretty undeniably produces clear and readable summaries and explanations.

replies(13): >>44572785 #>>44572809 #>>44572887 #>>44573175 #>>44573204 #>>44573208 #>>44573215 #>>44573682 #>>44573868 #>>44574638 #>>44574872 #>>44575425 #>>44578731 #
JohnMakin ◴[] No.44575425[source]
> That is a such a wild claim. People like the output of LLMs so much that ChatGPT is the fastest growing app ever.

And this kind of meaningless factoid was immediately usurped by the Threads app release, which IMO is kind of a pointless app. Maybe let's find a more meaningful metric before saying someone else's claim is wild.

replies(1): >>44576935 #
og_kalu ◴[] No.44576935[source]
Asking your Instagram Users to hop on to your ready made TikTok Clone is hardly in the same sphere as spinning up that much users from nothing.

And while Threads growth and usage stalled, ChatGPT is very much still growing and has *far* more monthly visits than threads.

There's really nothing meaningless about ChatGPT being the 5th most visited site on the planet, not even 3 years after release. Threads doesn't make the top 50.

replies(1): >>44577923 #
JohnMakin ◴[] No.44577923[source]
I think you just precisely explained why MAU / DAU growth is a meaningless metric in such discussions.
replies(1): >>44583580 #
1. og_kalu ◴[] No.44583580[source]
Seems like it's only meaningless if you ignore basic context.
replies(1): >>44585340 #
2. JohnMakin ◴[] No.44585340[source]
What basic context is being ignored? Here's how the thread has gone:

"chatGPT has the fastest growing userbase in history which shows users really like the output!"

This unsourced (and wrong) claim was offered in rebuttal to another post saying people don't like the output of LLM's. This rebuttal offers DAU/MAU as a metric of how much people like the app, I presume, and thus the output of the app. Besides that being a wild jump on its own, it's incorrect. As I pointed out - threads almost immediately beat that DAU/MAU record, and I'd offer a claim it hasn't exactly been a tremendous success either in popularity or monetarily. Pointing out that they got that DAU/MAU by registering their own users to it is precisely the point that is being made - this metric is a meaningless gauge of how popular an app is, and especially when viewed from the context of this argument, which is whether the popularity of the app (as it relates to DAU/MAU growth) also suggests people love consuming the output of it.

No offense, but are you sure you're following this conversation?

replies(1): >>44590265 #
3. og_kalu ◴[] No.44590265[source]
>Pointing out that they got that DAU/MAU by registering their own users to it is precisely the point that is being made - this metric is a meaningless gauge of how popular an app is, and especially when viewed from the context of this argument.

How does that make DAU/MAU growth meaningless ? Threads has special context. That's it. Almost all the other software applications that orbited that record are staples of internet life today. So because one entry had some special circumstances to take into account (that users weren't gained from scratch), the growth as a concept or comparison (for uses gained from scratch) is meaningless ? How does that make any sense ?

Also, yeah strong adoption (which is the real point here beyond just the growth) is a strong signal for satisfaction. It's very strange to claim most people don't like the output of what has half a billion weekly active users and is one of the most visited sites on the planet.

>Besides that being a wild jump on its own, it's incorrect.

It's not incorrect. Threads was the fastest to hit some early milestones (like 100M) sure but since growth stalled, ChatGPT is still the software application with the fastest adoption because it reached further milestones threads hasn't and may not reach.