Google is wrong to put AI search features behind paywall, says HPC leader You may have heard that Google is considering putting its latest AI search innovations behind a paywall, something that doesn't sit well with Rosanne Kincaid-Smith, COO at German HPC firm Northern Data Group. Kincaid-Smith told us in an interview that Google is in a different position than firms like OpenAI and others that …
I guess that the problem for Google is that if the AI scours google to construct an answer to a query then the user isn't seeing the ads which would accompany the searches and until AI has its own pocket money then Google won't get any ad-revenue. Unless, of course, it stuffs the AI's response full of ads, or includes product placement, or hides the output under the ketchup in a local Wimpy.
I recall that each query costs a few cents, and Google has billions of users. It's all very well to speak of democratization of access, and Google says they want to organize the world's information and make it universally accessible and useful... But these days, they also really insist on not losing money.
Depends on how it’s accounted for.
Normally, if you are treating R&D as investment for tax purposes, it needs to come out of gross margin, although some costs will get treated as normal opex. This is where a good accountant is necessary, so you are maximising your tax kick backs.
Can we cram a few more meaningless buzzwords in there?
I don't think they're meaningless buzzwords, they're kind of the point.
Google has already allowed us access to their transformative AI technology. So now, when we type in exact phrases we're searching for, we get a bunch of loosely related search results based on whatever their algorithms have been trained to promote. I can copy & paste from a paper sitting in front of me, and Google either won't find it at all, or it'll be buried waay down the results. Google (and Bing) have transformed search for the worse. And I suspect that'll be what happens with paywalled versions, only you'll get less adds and promoted links with the paywalled version.
I also detect some optimism. Sure, AlphaGoo could behave ethically, yet it's displayed no interests in doing so to date, and if anything, it's been steadily getting worse. It certainly moderates, but perhaps not in the way it should. I liked the example of Google promoting malicious links though. That is certainly something it should be able to prevent and moderate out. Especially given the way Google's built around link crawling and trawling. This is possibly something governments, regulators or just lawyers could fix by making Google liable for serving up those links. It's also a warning sign for all the regulations being proposed around 'fake news' and 'misinformation'. If AlphaGoo's giant silicon brains can't spot a malicious link, what chance do they have in detecting 'fake news' or 'misinformation'? Could SEO tricks be used to convince AIs that the Earth really is flat?
So many thumbs! I guess all the layoffs from Google and other 'big tech' is creating plenty of time to get that thumb exercise in.
Curious what people are objecting to though. Is it the idea that Google should be able to detect and filter malicious links? Isn't that a GoodThing(tm)? I suspect it's the idea that Google et al should be prevented from censoring non-official misinformation though.
This post has been deleted by its author
Interviewer missed asking the obvious follow on question: are Northern Data Group also going to democratise access by providing public (for free) access to their “ state-of-the-art generative AI cloud platform”…
What isn’t entirely clear is how Kincaid thinks Google giving AI stuff away for free will generate the revenues and market demand necessary to purchase the hundreds of HPCs she is hoping to sell.
While we'd all love everything to be free, the current surveillance capitalism, advertising based business model of the internet should have taught us that you pay for free stuff in other ways.
With search, you pay for free search with inaccurate results that are littered with sponsored content, sometimes disguised so carefully as to fool the user.
I think a lot of people would be happy to pay a small amount per month to know that the answers to their questions are the best, most accurate answers, and not the answers that sponsors paid Google the most to display.
In a generic sense paying for a productive AI service seems entirely reasonable. The problem is the thoroughly unethical and greedy Google elbwoing their way into the lead. And, that's not taking in account their very cozy relationship government mass surveillance opertations. Who are the good guys in the AI gold rush?
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
