Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Replies to comments on my "LLMs are eroding my career" post

human-in-the-loop.bearblog.dev

Post reply on HN
Print view View original

231–240 of 272 posts

Re: Replies to comments on my "LLMs are eroding my career" post

#231

Post by Denzel »

Ukv wrote: post #94
coffeefirst wrote: post #62

Earlier quoted context omitted.

> The cost for the same quality of output is going to drop at least 10x over the next 18-24 months. How do you know that? In 2026 the prices have been spiking. It now costs orders of magnitude more than it did in November.

Price of the current frontier may vary, but price for a given level of capability tends to drop pretty fast. April of last year you'd get 1431 ELO[0] from o3-2025-04-16 for $8.00 per million output tokens. April of this year you can get 1436 ELO from deepseek-v4-flash for $0.2 per million output tokens. [0]: https://huggingface.co/spaces/lmarena-ai/arena-leaderboard

Aren’t DeepSeek models deliberately priced lower than the cost to deliver? They’re subsidized which means the true cost is more than $0.2/Mtok.

Re: Replies to comments on my "LLMs are eroding my career" post

#232

Post by Ukv »

Denzel wrote: post #231
Ukv wrote: post #94

Earlier quoted context omitted.

Price of the current frontier may vary, but price for a given level of capability tends to drop pretty fast. April of last year you'd get 1431 ELO[0] from o3-2025-04-16 for $8.00 per million output tokens. April of this year you can get 1436 ELO from deepseek-v4-flash for $0.2 per million output tokens. [0]: https://huggingface.co/spaces/lmarena-ai/arena-leaderboard

Aren’t DeepSeek models deliberately priced lower than the cost to deliver? They’re subsidized which means the true cost is more than $0.2/Mtok.

DeepSeek models are open-source so there are a bunch of third-party providers offering similar prices. Factoring in that DeepSeek have to train the model (whereas third parties can make a small profit over just the inference costs) I'd assume that on net they're spending investor money, but I wouldn't think that's any less true of OpenAI.

Re: Replies to comments on my "LLMs are eroding my career" post

#233

Post by saltcured »

vrighter wrote: post #223
somebodythere wrote: post #164

Earlier quoted context omitted.

That is not true. You can tell you are on the latter part of the S-Curve you are on, if the rate of change of capabilities has decreased compared to before. That is not what we are seeing right now. The rate of change is increasing, or is at best, stable.

you compare the derivative yesterday, you measure the derivative today. difference between them is the jerk. It tells you where you are on the S-curve

It's only in retrospect that you know the level of filtering required to find this S curve trend among the noisy perturbations.

Just like people trying to time the market with "technical analysis". It is extremely easy to find whatever pattern you are looking for. A lot harder to accurately distinguish predictive power from fantasy.

Re: Replies to comments on my "LLMs are eroding my career" post

#234

Post by matheusmoreira »

Planktonne wrote: post #224
matheusmoreira wrote: post #207

Earlier quoted context omitted.

Have you seen the arrogance of artists? They acted as though they were above replacement, above automation. They acted as though they were superior . We're all facing very hard times ahead of us, but I would be lying if I said it wasn't at least a little cathartic to watch this unfold. Programmers, too, were just as arrogant until only a few years ago. As were doctors, lawyers... The list goes on. How the mighty have…

Do you see how this sort of hate-filled malevolence, as a pro-AI position, might make people less excited about AI? A lot of people are looking at AI now and seeing that its proponents sound like cartoon villains. That sends a message.

I certainly do. The only point I'm making is those people are sending plenty of messages of their own. You say that AI proponents sound like cartoon villains. To me the AI detractors sound a lot like elite lords being forcibly deposed from their titles. People who thought they were superior, but were proven wrong.

The only crime here would be stopping the AI onslaught just short of replacing the really powerful people. Let it happen.

Re: Replies to comments on my "LLMs are eroding my career" post

#235

Post by jason_oster »

sarchertech wrote: post #192
jason_oster wrote: post #174

Earlier quoted context omitted.

> people would have stopped learning how to write html and CSS around 2005. They did. Now it's all JSX or htmx or some other favored template or DSL monstrosity. Most people do not write HTML or CSS, and haven't in decades. You're spot on. This says nothing about quality, however. Quality of HTML/CSS is purely subjective. A website's presentation layer cannot meet any technical standard metric for quality in engineer…

I’m not going define away blocks of HTML inside of php scripts as not writing HTML by hand, but if you want to do that then sure most people were never writing HTML and CSS by hand.

I agree, I wouldn't classify PHP as "not writing HTML by hand". But that's not what I'm talking about. Most websites in the last 20 years have heavily leaned into client-side rendering, picking up with jQuery circa 2006. It's only recently with htmx that the pendulum has started swinging back toward server-side rendering.

Re: Replies to comments on my "LLMs are eroding my career" post

#236

Post by Planktonne »

matheusmoreira wrote: post #234
Planktonne wrote: post #224

Earlier quoted context omitted.

Do you see how this sort of hate-filled malevolence, as a pro-AI position, might make people less excited about AI? A lot of people are looking at AI now and seeing that its proponents sound like cartoon villains. That sends a message.

I certainly do. The only point I'm making is those people are sending plenty of messages of their own. You say that AI proponents sound like cartoon villains. To me the AI detractors sound a lot like elite lords being forcibly deposed from their titles. People who thought they were superior, but were proven wrong. The only crime here would be stopping the AI onslaught just short of replacing the really powerful peopl…

I don't think I've seen anything that smacks of elite lordship. Artists don't generally believe that other people should have their livelihoods taken away for the crime of not being artists.

Re: Replies to comments on my "LLMs are eroding my career" post

#237

Post by Denzel »

Ukv wrote: post #232
Denzel wrote: post #231

Earlier quoted context omitted.

Aren’t DeepSeek models deliberately priced lower than the cost to deliver? They’re subsidized which means the true cost is more than $0.2/Mtok.

DeepSeek models are open-source so there are a bunch of third-party providers offering similar prices. Factoring in that DeepSeek have to train the model (whereas third parties can make a small profit over just the inference costs) I'd assume that on net they're spending investor money, but I wouldn't think that's any less true of OpenAI.

Yes, DeepSeek is open-weight, but these third-party providers offering similar prices are subsidized with VC money as well. And you can find a range of prices for deepseek-v4-flash going up to and over $1/Mtok.

Even that $1/Mtok provided by Together AI is heavily subsidized by more than $1B in VC money.

This makes it unclear how the true cost curve is progressing. It’s not possible to confidently comment one way or another on the rate that cost is coming down when the entire industry is so heavily subsidized.

Re: Replies to comments on my "LLMs are eroding my career" post

#238

Post by sarchertech »

jason_oster wrote: post #235
sarchertech wrote: post #192

Earlier quoted context omitted.

I’m not going define away blocks of HTML inside of php scripts as not writing HTML by hand, but if you want to do that then sure most people were never writing HTML and CSS by hand.

I agree, I wouldn't classify PHP as "not writing HTML by hand". But that's not what I'm talking about. Most websites in the last 20 years have heavily leaned into client-side rendering, picking up with jQuery circa 2006. It's only recently with htmx that the pendulum has started swinging back toward server-side rendering.

I was around for all of this. Websites using jQuery were almost all using hand written HTML or HTML generated by something like rails web templates or PHP.

SPAs didn’t go mainstream until almost 10 years after 2006. And even at their height they never represented the majority of all websites.

But also most SPAs aren’t any less HTML by hand than using PHP templates are.

Re: Replies to comments on my "LLMs are eroding my career" post

#239

Post by Ferret7446 »

oompydoompy74 wrote: post #70
lelanthran wrote: post #60

Earlier quoted context omitted.

> I think in the future, those who succeed will be equivalent to wayfinders. In the future, those who succeed will be the owners of capital.

Past, Present, and Future. If you control the means of production you win. Knowledge, skill, and experience are largely irrelevant to the conversation. I’ve held this opinion for quite some time and would be interested to hear alternative perspectives.

That's clearly wrong, because capital doesn't just appear out of thin air. You are ignoring that there's clearly rare skills involved that enable a few to become very successful. Your strawman only applies to the second generation that inherits wealth, and case in point inherited wealth tends to disappear in a couple of generations further proving that skill is required to build and maintain wealth.

Re: Replies to comments on my "LLMs are eroding my career" post

#240

Post by teleforce »

I think the replies were pre-mature since when I commented on HN the comments total still has not reached 1000, but he already linked the replies already in the blog. Perhaps should have waited the comments to settle down in the original HN post, so that he can chose the most relevant and the interesting comments before posting the replies.

Personally, I'd really appreciate it if the original OP article author can respond to my comment in the previous post [1].

[1] https://news.ycombinator.com/item?id=48442851

Post reply on HN