Most active commenters
  • fkyoureadthedoc(4)

←back to thread

LLM Inevitabilism

(tomrenner.com)
1619 points SwoopsFromAbove | 11 comments | | HN request time: 0.964s | source | bottom
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 #
alonsonic ◴[] No.44570711[source]
I'm confused with your second point. LLM companies are not making any money from current models? Openai generates 10b USD ARR and has 100M MAUs. Yes they are running at a loss right now but that's because they are racing to improve models. If they stopped today to focus on optimization of their current models to minimize operating cost and monetizing their massive user base you think they don't have a successful business model? People use this tools daily, this is inevitable.
replies(11): >>44570725 #>>44570756 #>>44570760 #>>44570772 #>>44570780 #>>44570853 #>>44570896 #>>44570964 #>>44571007 #>>44571541 #>>44571655 #
lordnacho ◴[] No.44570853[source]
Are you saying they'd be profitable if they didn't pour all the winnings into research?

From where I'm standing, the models are useful as is. If Claude stopped improving today, I would still find use for it. Well worth 4 figures a year IMO.

replies(5): >>44570918 #>>44570925 #>>44570962 #>>44571742 #>>44572421 #
apwell23 ◴[] No.44570925[source]
> Well worth 4 figures a year IMO

only because software engineering pay hasn't adjusted down for the new reality . You don't know what its worth yet.

replies(2): >>44571084 #>>44574128 #
1. fkyoureadthedoc ◴[] No.44571084[source]
Can you explain this in more detail? The idiot bottom rate contractors that come through my team on the regular have not been helped at all by LLMs. The competent people do get a productivity boost though.

The only way I see compensation "adjusting" because of LLMs would need them to become significantly more competent and autonomous.

replies(2): >>44571579 #>>44573279 #
2. lelanthran ◴[] No.44571579[source]
> Can you explain this in more detail?

Not sure what GP meant specifically, but to me, if $200/m gets you a decent programmer, then $200/m is the new going rate for a programmer.

Sure, now it's all fun and games as the market hasn't adjusted yet, but if it really is true that for $200/m you can 10x your revenue, it's still only going to be true until the market adjusts!

> The competent people do get a productivity boost though.

And they are not likely to remain competent if they are all doing 80% review, 15% prompting and 5% coding. If they keep the ratios at, for example, 25% review, 5% prompting and the rest coding, then sure, they'll remain productive.

OTOH, the pipeline for juniors now seems to be irrevocably broken: the only way forward is to improve the LLM coding capabilities to the point that, when the current crop of knowledgeable people have retired, programmers are not required.

Otherwise, when the current crop of coders who have the experience retires, there'll be no experience in the pipeline to take their place.

If the new norm is "$200/m gets you a programmer", then that is exactly the labour rate for programming: $200/m. These were previously (at least) $5k/m jobs. They are now $200/m jobs.

replies(2): >>44571897 #>>44573056 #
3. fkyoureadthedoc ◴[] No.44571897[source]
$200 does not get you a decent programmer though. It needs constant prompting, babysitting, feedback, iteration. It's just a tool. It massively boosts productivity in many cases, yes. But it doesn't do your job for you. And I'm very bullish on LLM assisted coding when compared to most of HN.

High level languages also massively boosted productivity, but we didn't see salaries collapse from that.

> And they are not likely to remain competent if they are all doing 80% review, 15% prompting and 5% coding.

I've been doing 80% review and design for years, it's called not being a mid or junior level developer.

> OTOH, the pipeline for juniors now seems to be irrevocably broken

I constantly get junior developers handed to me from "strategic partners", they are just disguised as senior developers. I'm telling you brother, the LLMs aren't helping these guys do the job. I've let go 3 of them in July alone.

replies(3): >>44572544 #>>44572766 #>>44574269 #
4. handfuloflight ◴[] No.44572544{3}[source]
> It needs constant prompting, babysitting, feedback, iteration.

What do you think a product manager is doing?

replies(1): >>44572886 #
5. lelanthran ◴[] No.44572766{3}[source]
> It needs constant prompting, babysitting, feedback, iteration. It's just a tool. It massively boosts productivity in many cases, yes.

It doesn't sound like you are disagreeing with me: that role you described is one of manager, not of programmer.

> High level languages also massively boosted productivity, but we didn't see salaries collapse from that.

Those high level languages still needed actual programmers. If the LLM is able to 10x the output of a single programmer because that programmer is spending all their time managing, you don't really need a programmer anymore, do you?

> I've been doing 80% review and design for years, it's called not being a mid or junior level developer.

Maybe it differs from place to place. I was a senior and a staff engineer, at various places including a FAANG. My observations were that even staff engineer level was still spending around 2 - 3 hours a day writing code. If you're 10x'ing your productivity, you almost certainly aren't spending 2 - 3 hours a day writing code.

> I constantly get junior developers handed to me from "strategic partners", they are just disguised as senior developers. I'm telling you brother, the LLMs aren't helping these guys do the job. I've let go 3 of them in July alone.

This is a bit of a non-sequitor; what does that have to do with breaking the pipeline for actual juniors?

Without juniors, we don't get seniors. Without seniors and above, who will double-check the output of the LLM?[1]

If no one is hiring juniors anymore, then the pipeline is broken. And since the market price of a programmer is going to be set at $200/m, where will you find new entrants for this market?

Hell, even mid-level programmers will exit, because when a 10-programmer team can be replaced by a 1-person manager and a $200/m coding agent, those 9 people aren't quietly going to starve while the industry needs them again. They're going to go off and find something else to do, and their skills will atrophy (just like the 1-person LLM manager skills will atrophy eventually as well).

----------------------------

[1] Recall that my first post in this thread was to say that the LLM coding agents have to get so good that programmers aren't needed anymore because we won't have programmers anymore. If they aren't that good when the current crop starts retiring then we're in for some trouble, aren't we?

replies(1): >>44573132 #
6. fkyoureadthedoc ◴[] No.44572886{4}[source]
Not writing and committing code with GitHub Copilot, I'll tell you that. These things need to come a _long_ way before that's a reality.
7. sheiyei ◴[] No.44573056[source]
Your argument requires "Claude can replace a programme" to be true. Thus, your argument is false for the foreseeable future.
8. fkyoureadthedoc ◴[] No.44573132{4}[source]
> And since the market price of a programmer is going to be set at $200/m

You keep saying this, but I don't see it. The current tools just can't replace developers. They can't even be used in the same way you'd use a junior developer or intern. It's more akin to going from hand tools to power tools than it is getting an apprentice. The job has not been automated and hasn't been outsourced to LLMs.

Will it be? Who knows, but in my personal opinion, it's not looking like it will any time soon. There would need to be more improvement than we've seen from day 1 of ChatGPT until now before we could even be seriously considering this.

> Those high level languages still needed actual programmers.

So does the LLM from day one until now, and for the foreseeable future.

> This is a bit of a non-sequitor; what does that have to do with breaking the pipeline for actual juniors?

Who says the pipeline is even broken by LLMs? The job market went to shit with rising interest rates before LLMs hit the scene. Nobody was hiring them anyway.

replies(1): >>44578167 #
9. cgh ◴[] No.44573279[source]
There's another specific class of person that seems helped by them: the paralysis by analysis programmer. I work with someone really smart who simply cannot get started when given ordinary coding tasks. She researches, reads and understands the problem inside and out but cannot start actually writing code. LLMs have pushed her past this paralysis problem and given her the inertia to continue.

On the other end, I know a guy who writes deeply proprietary embedded code that lives in EV battery controllers and he's found LLMs useless.

10. nyarlathotep_ ◴[] No.44574269{3}[source]
> I constantly get junior developers handed to me from "strategic partners", they are just disguised as senior developers. I'm telling you brother, the LLMs aren't helping these guys do the job. I've let go 3 of them in July alone.

I find this surprising. I figured the opposite: that the quality of body shop type places would improve and the productivity increases would decrease as you went "up" the skill ladder.

I've worked on/inherited a few projects from the Big Name body shops and, frankly, I'd take some "vibe coded" LLM mess any day of the week. I really figured there was nowhere to go but "up" for those kinds of projects.

11. bscphil ◴[] No.44578167{5}[source]
> The current tools just can't replace developers. They can't even be used in the same way you'd use a junior developer or intern. It's more akin to going from hand tools to power tools than it is getting an apprentice.

In that case it seems to depend on what you mean by "replacing", doesn't it? It doesn't mean a non-developer can do a developers job, but it does mean that one developer can do two developer's jobs. That leads to a lot more competition for the remaining jobs and presumably many competent developers will accept lower salaries in exchange for having a job at all.