Hacker News new | ask | show | jobs
by not2b 834 days ago
The article seems to assume that "AI OS" will work for the user, rather than for the owners. In practice, OpenAI and its competitors would just extort advertisers: pay us and we'll let your ads through even if the user doesn't want the ads.

In theory a user with sufficient compute power could run their filter entirely locally and not rely on OpenAI, Google, Meta, or some other competitor, but that might require a home data center filled with GPUs and an enormous electric bill.

2 comments

> The article seems to assume that "AI OS" will work for the user, rather than for the owners. In practice, OpenAI and its competitors would just extort advertisers: pay us and we'll let your ads through even if the user doesn't want the ads.

In the real world though we've already seen how that shakes out: uBlock remains possibly the most popular browser extension on the planet, for good reason. Does everyone block ads? Of course not, but ads are also pretty easy to ignore if you're not the kind of technically inclined person who knows how to make them go away.

The up end of a product like a browser extension that removes ads, marketing fluff, surveys, etc. from your experience is much, much bigger than something like even uBlock can manage. Or hell, why even stop at a browser extension? What if you just had an entire browser that runs on an AI that answers to you and only you, or at the very least, only your fellow consumers and not big tech?

Like, AI is hard tech to scale, sure, but it's not impossible and I can easily envision a company making a sizable amount of bank on the idea of sanitizing the internet for their users.

> uBlock remains possibly the most popular browser extension on the planet, for good reason

And the most popular web browser is working hard on making it useless. Chipping away at what users are allowed to block and conditioning a new generation of users to accept the loss of freedom. Expect it to continue.

> I can easily envision a company making a sizable amount of bank on the idea of sanitizing the internet for their users.

I can easily that company being bought by an ad company and buried. There are times when it's more profitable to not give consumers what they want and very few companies are going to leave money on the table when it's right there for the taking.

And uBlock is just the vanguard, the forefront.

The next step is things like YouTube Premium - you pay to not see ads. But this can get corrupted which leads to things like ... Kagi - where the whole point is you're paying instead of ads.

YouTube Premium is an extortion racket. Agreeing to "Pay us to solve the problem we're causing you" is never a good idea. Kagi at least offers a solution to a problem caused by others.
> Pay us to solve the problem we're causing you

More like: "We offer a service funded by ads, pay us if you don't want ads"

You are welcome to pay for cable and go tough grass to get your other entertainment need fulfilled instead of watching YouTube with ads or paying for premium.

Generally people who don't want ads on YouTube nor want to pay for YouTube Premium also don't understand just how much money hosting video costs.

> You are welcome to pay for cable and go tough grass to get your other entertainment need fulfilled instead of watching YouTube with ads or paying for premium.

Alternately, I'm perfectly welcome to install an adblocker or use any number of other tools that ensure that I never see their ads at all. Personally, I've been paying Google with my data for as long as Google has existed and they've taken far more from me than I was willing to give so I won't cry over the costs that they, with their hundreds of billions, have to spend to keep collecting our data and manipulating our culture via youtube. It's been working out just fine for them so far without them being able to waste my time with ads, and I suspect that won't change any time soon.

This is fine but then don't complain about the adblocker randomly not working and Google implementing measures to prevent adblockers, or actively making you, a non-paying user's experience worse.

And arguably if it has been working just fine for them they wouldn't need to keep spending billions and implementing a paid tier.

Also your adblocker only prevents client side tracking, nothing is stopping Google (or any other services) from collecting "anonymous" usage data and storing it.

After all does it matter at all that user53728646288661442883 in the database was searching for a coffee machine on google and is now watching reviews on YouTube for coffee machines if there is no way to reference that user to who you are personally?

If your complaint is tracking then stop using google services, an adblocker just reduces the amount of tracking, it doesn't eliminate it.

Google offered a way out of watching ads for what is probably a quite reasonable amount of money for anyone browsing Hackernews and your reasonse is "I'm just going to keep using my adblocker and actively complain about it"

That makes it seem like YT's decision is market-driven, when the reality is it's been making losses since its inception, ads or no ads. Premium cost $9.99, then $12.99, now $14.99. Sounds like an operation that has no clue how to price for sustainability.
> when the reality is it's been making losses since its inception

They make billions in revenue off of youtube. Where did you see that they're losing money on it? I suspect that even if youtube was not profitable on its own it'd still be valuable enough for google to continue running it and that they can more than afford to keep the light on.

And yet this is your argument for not paying? Or just an observation because I am fine with just an observation.
You need a data centre if you want to run the bleeding edge SOTA models but the weights of these models are fixed which means it should be possible to instantiate them in hardware and once you have a model smart enough to do most things you’d be crazy not to build a factory to churn out little boxes that you can connect to literally anything to give it a natural language interface and useful levels of intelligence. Mark my words: your doorbell will have an LLM in it in less than 10 years.