←back to thread

LLM Inevitabilism

(tomrenner.com)
1613 points SwoopsFromAbove | 4 comments | 15 Jul 25 04:35 UTC | HN request time: 1.016s | source
Show context
mg ◴[15 Jul 25 05:41 UTC] No.44568158[source]▶
In the 90s a friend told me about the internet. And that he knows someone who is in a university and has access to it and can show us. An hour later, we were sitting in front of a computer in that university and watched his friend surfing the web. Clicking on links, receiving pages of text. Faster than one could read. In a nice layout. Even with images. And links to other pages. We were shocked. No printing, no shipping, no waiting. This was the future. It was inevitable.

Yesterday I wanted to rewrite a program to use a large library that would have required me to dive deep down into the documentation or read its code to tackle my use case. As a first try, I just copy+pasted the whole library and my whole program into GPT 4.1 and told it to rewrite it using the library. It succeeded at the first attempt. The rewrite itself was small enough that I could read all code changes in 15 minutes and make a few stylistic changes. Done. Hours of time saved. This is the future. It is inevitable.

PS: Most replies seem to compare my experience to experiences that the responders have with agentic coding, where the developer is iteratively changing the code by chatting with an LLM. I am not doing that. I use a "One prompt one file. No code edits." approach, which I describe here:

https://www.gibney.org/prompt_coding

replies(58): >>44568182 #>>44568188 #>>44568190 #>>44568192 #>>44568320 #>>44568350 #>>44568360 #>>44568380 #>>44568449 #>>44568468 #>>44568473 #>>44568515 #>>44568537 #>>44568578 #>>44568699 #>>44568746 #>>44568760 #>>44568767 #>>44568791 #>>44568805 #>>44568823 #>>44568844 #>>44568871 #>>44568887 #>>44568901 #>>44568927 #>>44569007 #>>44569010 #>>44569128 #>>44569134 #>>44569145 #>>44569203 #>>44569303 #>>44569320 #>>44569347 #>>44569391 #>>44569396 #>>44569574 #>>44569581 #>>44569584 #>>44569621 #>>44569732 #>>44569761 #>>44569803 #>>44569903 #>>44570005 #>>44570024 #>>44570069 #>>44570120 #>>44570129 #>>44570365 #>>44570482 #>>44570537 #>>44570585 #>>44570642 #>>44570674 #>>44572113 #>>44574176 #
pavlov ◴[15 Jul 25 09:20 UTC] No.44569391[source]▶
Compare these positive introductory experiences with two technologies that were pushed extremely hard by commercial interests in the past decade: crypto/web3 and VR/metaverse.

Neither was ever able to offer this kind of instant usefulness. With crypto, it’s still the case that you create a wallet and then… there’s nothing to do on the platform. You’re expected to send real money to someone so they’ll give you some of the funny money that lets you play the game. (At this point, a lot of people reasonably start thinking of pyramid schemes and multi-level marketing which have the same kind of joining experience.)

With the “metaverse”, you clear out a space around you, strap a heavy thing on your head, and shut yourself into an artificial environment. After the first oohs and aahs, you enter a VR chat room… And realize the thing on your head adds absolutely nothing to the interaction.

replies(10): >>44569596 #>>44569648 #>>44569660 #>>44569706 #>>44569711 #>>44569739 #>>44569855 #>>44569906 #>>44570576 #>>44570592 #
1. pjc50 ◴[15 Jul 25 12:47 UTC] No.44570576[source]▶
I think AI is inevitable in the way that bitcoin is now inevitable: it's not going to go away, it consumes a huge amount of energy, has various negative externalities, but a massive fanbase.

It doesn't really matter whether crypto is "useful", it has billions of dollars worth of fans. Similarly the LLM fans are not going to go away. However, there will probably be curated little oases for human-made works. We're also going to see a technique adapted from self-crashing cars: the liability human. A giant codebase is launched and a single human "takes responsibility" (whatever that ends up meaning) for the failures.

replies(1): >>44570747 #
2. DebtDeflation ◴[15 Jul 25 13:09 UTC] No.44570747[source]▶
It does matter whether something is useful and I really wish people would stop making comparisons to crypto because it's an absolutely terrible comparison.

AI is certainly in a bubble right now, as with dotcoms in 1999. But AI is also delivering a lot of value right now and advancing at an incredible pace. It will become ubiquitous and at a faster pace than the Internet ultimately did.

Meanwhile, Bitcoin has been around for 17 years and there still are no non-criminal use cases apart from buying it and hoping it will be worth more in the future.

replies(1): >>44575146 #
3. bigstrat2003 ◴[15 Jul 25 19:52 UTC] No.44575146[source]▶
> Meanwhile, Bitcoin has been around for 17 years and there still are no non-criminal use cases apart from buying it and hoping it will be worth more in the future.

That is plain and simply false. It works just fine as a currency, and some legitimate businesses even accept it. I think it's true that Bitcoin is not particularly useful, but that's not the same as there being no non-criminal use cases.

replies(1): >>44581020 #
4. DebtDeflation ◴[16 Jul 25 11:33 UTC] No.44581020{3}[source]▶
It doesn't though. Go to your local Wal Mart, Costco, Home Depot, Target, etc. and try to make a purchase with it. Go to a local supermarket or restaurant and try to buy food with it. Stop at a gas station and try to buy fuel with it. Go to Amazon or eBay and try to make a purchase with it. Go to an airline or hotel website and try to book travel with it. Go into your Uber app and try to pay with it. Call your landlord and try to pay your rent with it. Objectively, it doesn't "work just fine as a currency" or we'd have empirical evidence of it actually being used as such in real life.