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

The force-feeding of AI features on an unwilling public

honest-broker.com

Post reply on HN
Print view View original

101–110 of 421 posts

Re: The force-feeding of AI features on an unwilling public

#101

Post by pu_pe »

kgeist wrote: post #36
einrealist wrote: post #26

The major AI gatekeepers, with their powerful models, are already experiencing capacity and scale issues. This won't change unless the underlying technology (LLMs) undergoes a fundamental shift. As more and more things become AI-enabled, how dependent will we be on these gatekeepers and their computing capacity? And how much will they charge us for prioritised access to these resources? And we haven't really gotten t…

We run our own LLM server at the office for a month now, as an experiment (for privacy/infosec reasons), and a single RTX 5090 is enough to serve 50 people for occasional use. We run Qwen3 32b which in some benchmarks is equivalent to GPT 4.1-mini or Gemini 2.5 Flash. The GPU allows 2 concurrent requests at the same time with 32k context each and 60 tok/s. At first I was skeptical a single GPU would be enough, but it…

That's really great performance! Could you share more details about the implementation (ie which quantized version of the model, how much RAM, etc.)?

Re: The force-feeding of AI features on an unwilling public

#102

Post by brookst »

cs702 wrote: post #95

Your may agree or disagree with the OP, but this passage is spot-on: "I don’t want AI customer service—but I don’t get a choice. I don’t want AI responses to my Google searches—but I don’t get a choice. I don’t want AI integrated into my software—but I don’t get a choice. I don’t want AI sending me emails—but I don’t get a choice. I don’t want AI music on Spotify—but I don’t get a choice. I don’t want AI books on Ama…

It’s not spot on. Buying and using all of these products is a choice.

The last is especially egregious. I don’t want poorly-written (by my standards) books cluttering up bookstores, but all my life I’ve walked into bookstores and found my favorite genres have lots of books I’m not interested in. Do I have some kind of right to have stores only stock products that I want?

The whole thing is just so damn entitled. If you don’t like something, don’t buy it. If you find the presence of some products offensive in a marketplace, don’t shop there. Spotify is not a human right.

Re: The force-feeding of AI features on an unwilling public

#103

Post by amelius »

cs702 wrote: post #95

Your may agree or disagree with the OP, but this passage is spot-on: "I don’t want AI customer service—but I don’t get a choice. I don’t want AI responses to my Google searches—but I don’t get a choice. I don’t want AI integrated into my software—but I don’t get a choice. I don’t want AI sending me emails—but I don’t get a choice. I don’t want AI music on Spotify—but I don’t get a choice. I don’t want AI books on Ama…

There are plenty of non AI books on Amazon.

Re: The force-feeding of AI features on an unwilling public

#104

Post by brookst »

einrealist wrote: post #44
kgeist wrote: post #36

Earlier quoted context omitted.

We run our own LLM server at the office for a month now, as an experiment (for privacy/infosec reasons), and a single RTX 5090 is enough to serve 50 people for occasional use. We run Qwen3 32b which in some benchmarks is equivalent to GPT 4.1-mini or Gemini 2.5 Flash. The GPU allows 2 concurrent requests at the same time with 32k context each and 60 tok/s. At first I was skeptical a single GPU would be enough, but it…

If those smaller models are sufficient for your use cases, go for it. But for how much longer will companies release smaller models for free? They invested so much. They have to recoup that money. Much will depend on investor pressure and the financial environment (tax deductions etc). Open Source endeavors will have a hard time to bear the resources to train models that are competitive. Maybe we will see larger coop…

Pricing for commodities does not allow for “recouping costs”. All it takes is one company seeing models as a complementary good to their core product, worth losing money on, and nobody else can charge more.

I’d support an Apache for ML but I suspect it’s unnecessary. Look at all of the money companies spend developing Linux; it will likely be the same story.

Re: The force-feeding of AI features on an unwilling public

#105

Post by brookst »

relaxing wrote: post #100
mrob wrote: post #25

>Everybody wanted the Internet. I don't think this is true. A lot of people had no interest until smartphones arrived. Doing anything on a smartphone is a miserable experience compared to using a desktop computer, but it's more convenient. "Worse but more convenient" is the same sales pitch as for AI, so I can only assume that AI will be accepted by the masses too.

Yes, everyone wanted the internet. It was massively hyped and the uptake was widespread and rapid. Obviously saying “everyone” is hyperbole. There were luddites and skeptics about it just like with electricity and telephones. Nevertheless the dotcom boom is what every new industry hopes to be.

I was there. There was massive skepticism, endless jokes about internet-enabled toasters and the uselessness and undesirability of connecting everything to the internet, people bemoaning the loss of critical skills like using library card catalogs, all the same stuff we see today.

In 20 years AI will be pervasive and nobody will remember being one of the luddites.

Re: The force-feeding of AI features on an unwilling public

#106

Post by dimitri-vs »

oceanplexian wrote: post #91
einrealist wrote: post #86

Earlier quoted context omitted.

Isn't "Engineering" is based on predictability, on repeatability? LLMs are not very predictable. And that's not just true for the output. Each change to the model impacts how it parses and computes the input. For someone claiming to be a "Prompt Engineer", this cannot work. There are so many variables that are simply unknown to the casual user: training methods, the training set, biases, ... If I get the feeling I am…

> LLMs are not very predictable. And that's not just true for the output. If you run an open source model from the same seed on the same hardware they are completely deterministic. It will spit out the same answer every time. So it’s not an issue with the technology and there’s nothing stopping you from writing repeatable prompts and promoting techniques.

Realistically, how many people do you think have the time, skills and hardware required to do this?

Re: The force-feeding of AI features on an unwilling public

#107

Post by brookst »

iLoveOncall wrote: post #46
m000 wrote: post #3

I mostly agree with TFA, with one glaring exception: The quality of Google search results has regressed so badly in the past years (played by SEO experts), that AI was actually a welcome improvement.

User issue. Every single time this comes up. People don't know how to search, that's it. Even the HN population. Every time this gets posted, I ask for one example of thing you tried to find and what keywords you used. So I'm giving you the same offer, give me for one thing you couldn't find easily on Google and the keywords you used, and I'll show you Google search is just fine.

Google is nearly useless for recipes. Try finding a recipe for beef bourguignon. They exist, but with huge prefaces and elaboration that mean endless scrolling on a phone, all in the name of maximizing time spent on page (which is a search ranking criteria).

Re: The force-feeding of AI features on an unwilling public

#108

Post by fnordpiglet »

brookst wrote: post #102
cs702 wrote: post #95

Your may agree or disagree with the OP, but this passage is spot-on: "I don’t want AI customer service—but I don’t get a choice. I don’t want AI responses to my Google searches—but I don’t get a choice. I don’t want AI integrated into my software—but I don’t get a choice. I don’t want AI sending me emails—but I don’t get a choice. I don’t want AI music on Spotify—but I don’t get a choice. I don’t want AI books on Ama…

It’s not spot on. Buying and using all of these products is a choice. The last is especially egregious. I don’t want poorly-written (by my standards) books cluttering up bookstores, but all my life I’ve walked into bookstores and found my favorite genres have lots of books I’m not interested in. Do I have some kind of right to have stores only stock products that I want? The whole thing is just so damn entitled . If…

I actually use the AI books that litter kindle unlimited to teach my daughter how to differentiate and be more sophisticated. I think a feature of all this is it inculcates a lot of people to AI spew. If it were isolated to the elite and the unscrupulous alone people would be a lot more vulnerable. By saturating the world with it, people get a true choice by being able to recognize it when they see it and avoid the output. It’s not like all our surfaces are not covered in enshittification as it is, another dose of it won’t make it meaningfully worse. And I know a lot of non English speakers that really appreciate the AI writing assistants built into email, the ai summaries built into search. Assuming no one finds them beneficial because it litters an already littered experience is a bit close minded. Many people otherwise challenged in some way. Summaries help dyslexics get through otherwise intractable walls of text, multi modal glasses help the vision impaired, witting assistants help bilingual workers level the playing field. Just because these don’t apply to you doesn’t mean it’s bothersome. (Now should you be able to disable it? Maybe, but as the author points out that’s a product choice made for financial reasons and there’s a market of products that make a different choice - don’t like google? Don’t feel so entitled that every service be free and pay for kagi)

Probably no one enjoys AI books though. I did my best at devils advocate on that above.

Re: The force-feeding of AI features on an unwilling public

#109

Post by gchamonlive »

But that's exactly the problem with proprietary software. It's not force-feeding you anything, it's working exactly as intended.

Software is loyal to the owner. If you don't own your software, software won't be loyal to you. It can be convenient for you, but as time passes and interest changes, if you don't own software it can turn against you. And you shouldn't blame Microsoft or it's utilities. It doesn't owe you anything just because you put effort in it and invested time in it. It'll work according to who it's loyal to, who owns it.

If it bothers you, choose software you can own. If you can't choose software you own now, change your life so you can in the future. And if you just can't, you have to accept the consequences.

Re: The force-feeding of AI features on an unwilling public

#110

Post by brookst »

ethan_smith wrote: post #94
daft_pink wrote: post #2

The issue really is that the AI isn’t good enough that people actually want it and are willing to pay for it. It’s like IPV6, if it really was a huge benefit to the end user, we’d have adopted it already.

IPv6 adoption is actually limited by network effect and infrastructure transition costs, not lack of end-user benefits - unlike AI, which faces a value perception problem.

ChatGPT has more than 500m DAU, three years after creation. Is that really a value perception problem?
Post reply on HN