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

The librarian immediately attempts to sell you a vuvuzela

kaveland.no

Post reply on HN
Print view View original

61–70 of 346 posts

Re: The librarian immediately attempts to sell you a vuvuzela

#61

Post by sph »

reubenmorais wrote: post #43
boothby wrote: post #2

> These days, I find that I am using multiple search engines and often resort to using an LLM to help me find content. For a few months, I've been wondering: how long until advertisers get their grubby meathooks into the training data? It's trivial to add prompts encouraging product placement, but I would be completely shocked if the big players don't sell out within a year or two, and start biasing the models themse…

Google has been working on auctioning token-level influence during LLM generation for years now: https://research.google/blog/mechanism-design-for-large-lang...

Google: ruining their core product for that sweet ad money.

Re: The librarian immediately attempts to sell you a vuvuzela

#62

Post by DavidPiper »

> The LLM providers [are] businesses that have to find a way to earn a ridiculous amount of money for a huge number of big investors, and capitalism does not have builtin morals. What externalities will the broader public end up paying?

I actually have a theory about this. I hate it, but I can absolutely imagine this future.

I'm going to specifically talk about the software engineering industry, but let's assume that LLMs progress to the stage where "vibe coding" can be applied to other areas ("vibe writing", "vibe research", "vibe security", "vibe art", "vibe doctors", "vibe management", "vibe CEOs", etc.)

It only takes a few years of "vibe coding graduates" to be successful in their work to create a new class of software engineer - this is in fact what AI companies are actively encouraging / envisioning as the future. Assuming this happens in the next few years, we're still in the phase where AI companies are burning money acquiring as many of these users as possible.

In about 5 years, some of those vibe coders will become vibe managers, and executives will no doubt be even more invested in LLMs as the solution to their problems.

At a certain tipping point, a large part of the industry can't actually function effectively without LLMs. I don't know when this point will be, but vibe coders (or other vibe ers) don't have to be a majority, they just have to be a large enough group.

Suddenly AI companies have all their losses called in and they have to pay back their VCs.

LLM usage prices skyrocket.

----

Four things happen across 2 axes:

- [A] Companies that can afford to pay skyrocketing LLM costs, vs. [B] those that can't

- [C] Companies that have reached a critical mass of vibe coders, vs. [D] those that haven't

----

[BC] These companies collapse. They don't have talent and they can't afford to outsource it to LLMs anymore.

[BD] These companies lay off all their vibe coders because they can't afford LLMs anymore. They survive on the talent they retain, but this looks very different if you're a large or small business. Small businesses probably fail too.

[AC] These companies see an enormous increase in costs that they cannot avoid. Large layoffs likely, but widespread vibe coding continues.

[AD] These companies have a decision to make: lay off all their vibe coders, or foot the LLM bill. The action they take will depend on their exact circumstances. Again, most small business in this situation probably fail.

---

The real question is, for the surviving vibe companies [AC, AD]. Will they be able to sustain such high costs in the long-run, and even if they can, will enough be able to sustain them to successfully pay back all the AI companies' losses to that date?

Interesting times ahead, maybe.

Re: The librarian immediately attempts to sell you a vuvuzela

#63

Post by chgs »

bdangubic wrote: post #22
iamkeithmccoy wrote: post #20

> AIs are much better at responding to my intent, and they rarely attempt to sell me anything Yet. It's only a matter of time before AI becomes ad-riddled and enshittified.

they won’t as long as you are paying $200+/month

If they can make an extra $20 from adverts on to they will.

Sky TV makes 10 times as much from subscription as adverts but spends 30% of the time showing you adverts. London underground revenue is a similar ratio - for every £9 tickets they make £1 in adverts. If I go to the cinema they spend 20 minutes showing adverts to people who spent £50 on tickets and popcorn.

Companies shave very little incentive not to make things shit with adverts. The measurable cost to them is tiny, the cost to the rest of the world is massive. Odeon won’t attribute lost revenue from my reduced visits to their adverts, but will measure the 50p or whatever they get.

Re: The librarian immediately attempts to sell you a vuvuzela

#64

Post by willvarfar »

cbsmith wrote: post #12

Some curious misconceptions here about how the search industry works. There's a very strong financial incentive for ad-powered search engines to keep SEO spam out of search results, because that makes advertisers more willing to pay for search placement. A publicly run search engine would not have those incentives and would be at if anything graver risk of a "tragedy of the commons" type scenario where the engine is…

This stopped being true when Google brought Doubleclick?

Google gets money from sending users to ad-laden sites. They get to double-dip.

If Google _didn't_ own Doubleclick etc, then there would be an incentive for them to prioritise content over content farms etc.

Re: The librarian immediately attempts to sell you a vuvuzela

#65

Post by sph »

pjerem wrote: post #55
hug wrote: post #15

The author would do well to check out Kagi -- the search results for all of the suggested queries are, to my eye, better than the results they found. Given that Kagi's higher tier plans come with search-enabled LLM chat interfaces, and those searches use Kagi's results (which, again, appear to be superior) it seems to me that you get the best of both worlds: Better search, and better search results to feed into your…

Happy customer of Kagi here too. However if I’m being honest, I’m starting to get LLM generated content in Kagi results. Don’t think it’s their fault or that it happens more than everywhere else neither what they could do about it but it happens.

I am pretty sure 50% of my search results on Kagi (or Google or Bing) are LLM-generated. You can search anything and find a dozen websites that, surprise, have a page dedicated exactly to that topic, organized in three or four neat sections.

The internet is dead and is starting to smell.

Re: The librarian immediately attempts to sell you a vuvuzela

#66

Post by Kiyo-Lynn »

The current search engines really feel like a librarian who's always trying to sell you something. I just want to find a simple answer, but I keep getting led to all kinds of other pages. I believe if search engines were more like public libraries, focused on providing information rather than recommending things for commercial reasons, the experience would be so much better.

Re: The librarian immediately attempts to sell you a vuvuzela

#67

Post by willvarfar »

reubenmorais wrote: post #43
boothby wrote: post #2

> These days, I find that I am using multiple search engines and often resort to using an LLM to help me find content. For a few months, I've been wondering: how long until advertisers get their grubby meathooks into the training data? It's trivial to add prompts encouraging product placement, but I would be completely shocked if the big players don't sell out within a year or two, and start biasing the models themse…

Google has been working on auctioning token-level influence during LLM generation for years now: https://research.google/blog/mechanism-design-for-large-lang...

And just over a year ago now the OpenAI "preferred publisher program" pitch deck to investors leaked. https://news.ycombinator.com/item?id=40310228

Re: The librarian immediately attempts to sell you a vuvuzela

#68

Post by hnbad »

moron4hire wrote: post #7
boothby wrote: post #2

> These days, I find that I am using multiple search engines and often resort to using an LLM to help me find content. For a few months, I've been wondering: how long until advertisers get their grubby meathooks into the training data? It's trivial to add prompts encouraging product placement, but I would be completely shocked if the big players don't sell out within a year or two, and start biasing the models themse…

It's kind of already happening. For example, if you ask an LLM for advice on building an application, it's going to pigeon-hole you into using React.

Or for a more concerning example, GitHub is owned by Microsoft who want to sell cloud services so it stands to reason it would be in their interest to have GitHub Copilot steer developers towards building applications using architectural patterns that lend themselves more to using those cloud services, e.g. service-oriented architecture even when it is against the developer's interests.

This doesn't have to be as blunt as promoting specific libraries or services and it's a bias that could even be introduced "accidentally".

Re: The librarian immediately attempts to sell you a vuvuzela

#69

Post by isoprophlex »

collingreen wrote: post #49
jedbrooke wrote: post #45

Earlier quoted context omitted.

not quite ads in LLMS, but I had an interesting experience with google maps the other day. the directions voice said "in 100 feet, turn left at the ". Normally it would say "at the traffic light" or similar. And this wasn't some easy to miss hidden street, it was just a normal intersection. I can only hope they aren't changing the routes yet to make you drive by the highest bidder

This is devilish. I'm adding your idea to my torment nexus list.

"Continue driving on Thisandthat Avenue, and admire the happy, handsome people you see on your right, shopping at Vuvuzelas'R'Us, your place for anything airhorn!"

Re: The librarian immediately attempts to sell you a vuvuzela

#70

Post by hnbad »

magarnicle wrote: post #35

So why aren't real libraries like this? Can you fund the internet the same way?

Because real libraries are not for-profit enterprises and have a negative (or at least extremely unimpressive) ARR. Incidentally their leadership positions also don't provide anywhere near the levels of compensation you'd expect for a C-level position at an equally large company.

Asking for "the Internet" to be funded the same way as real libraries is quite the contrary to the dominant cultural narrative which asks for public services and entire governments to be operated "like a business", which usually means cutting funding, selling assets, doing layoffs and eventually scrapping it for parts once it has become predictably defunct.

Post reply on HN