Hacker News Reader

Comment by dizzydes | Hacker News Reader

> Soon, you’ll see AI-powered features in Search

Despite this blog obviously being written for Wall St, this line is key. This is why anyone is going to struggle to compete with Google in AI.

Google Search is used so widely and its taxonomy of meaning so vast already that once they flip this on they’ll be unstoppable. (Assuming any rate of improvement whatsoever as we search.)

I’m not mentioning all other learning data like email, photos and drive, in part because other companies have similar, albeit in more specialised forms.

View on HN

I wonder how long till they open it up to your private data stored in Google? Eg.

User: What should I be doing today?

AI: You have a court appearance at 9am. Prioritize that, because failing to appear might result in an arrest warrant. Next, prioritize your mums birthday - all the other family will be there, and with your mums cancer results last week, this might be the last. If you travel between them by bus, then you can spend the time going through your teams messages that you should have replied to last week.

Me: What about the leaking faucet in my house?

AI: You don't have the time to fix it yourself, nor the money to pay a plumber to do it, so I suggest you leave it leaking for now. I watched the video - its leaking down the basin, so won't do damage, and the water cost is around 8 cents a day.

Yea this isn’t a huge reach.

At that point the line between commerce and giving the best advice is going to get REALLY messy.

I'd imagine it'll be something they do. Hooking it into things you've told assistant, documents you've stored, emails, calendar events, map data, etc.

Pretty creepy and a tad dystopian but thats Google for you.

Fun exercise for the reader: how much of this is actually possible with LLMs and how much is not.

> Fun exercise for the reader: how much of this is actually possible with LLMs and how much is not.

From what I understand about LLMs - not directly. But it may be possible to integrate LLMs with other services such that LLMs respond this way.

> Fun exercise for the reader: how much of this is actually possible with LLMs and how much is not.

I have no idea.

But it would be nice?* if this LLM stuff could be incorporated into the telephone system for my bank, pharmacy, etc.

It's obvious that these organizations don't want to connect me with a $killed human, so instead of having me interact with an infuriating "pretend" human, maybe CVS can fuse their phone system with ChatGPT to make the experience a little less maddening.

"Hey, I already gave you my birth date. No need to ask again. And I told you 30 seconds ago that I don't need to schedule a COVID vaccine. Just tell the pharmacist <xyz>."

* Be careful what you (I) wish for, I guess.

Reading through all your private documents and extracting the useful information to answer an open ended question like this is still a little way off.

Specifically, LLM's can generally only take into account ~8000 words of context when deciding on a response. Summarizing the question and all necessary information for the answer into 8000 words is hard when the user might have millions of words in their inbox.

Having said that, I don't think it's far off. There are already prototypes of LLM's which have information retrieval abilities (ie. it could do a keyword search of your inbox to find a few relevant documents to read to decide on a response). There are also promising efforts to make that 8000 word number far larger.

User: What should I be doing today?

AI: You have a court appearance today, I have booked a route for you on google maps. Next buy a gift for your mum’s birthday, here are some recommendations sorted by ad spend spend. For your travel, I recommend these options, sorted by ad spend.

Me: What about the leaking faucet in my house?

AI: You don't have the time to fix it yourself, here is a list of contractors that can fix the leak, sorted by google ad spend.

Is that any different from trying to source it "on your own"?

Nah pry not, since search results are ranked by ads, same thing just in a chat based format.

I recall reading a very interesting short story about exactly this - handing over the entirety of your daily minutiae to AI and the moral consequences. I'm unable to find the name of it, maybe something by Ted Chiang?

Google could do this now. They already have your calendar and your contacts. Do you think that a conversational interface was what was holding them back? Thousands of engineers, but if only they could figure out how to merge Gmail and Eliza!

Isn't this what the Google assistant basically does?

I don't think it's that clearcut. These AI queries are resource intensive. Google will have to cut a fine line between quality and profit margin. If it's no good people just won't bother and it won't gain traction. The searches we make are already analyzed and all models tend to use billions of parameters anyway, even self generated. Google also are limited by their risk aversion, as seen in this extremely sanitized letter by Sundar, devoid of substance. Their existing AI tool is severely limited. If this new tool comes out looking like that prior Google tool, it's going to be very difficult for them to gain traction.

1. ChatGPT have made a decent go of it on a much smaller budget

2. Wall St tends to be very forgiving when the investment is something with a strong a moat as AI. Metaverse is an example of what happens when they don’t believe…

3. With a few exceptions like weather, their current searches provide links to static pages others have built, so much of their understanding and analysis is going to waste. Giving custom results breaks through that.

Didn’t Altman respond to Elon on Twitter and say an interaction cost 1 cent?

https://twitter.com/sama/status/1599671496636780546?s=20&t=p...

If Google handles 10 billion searches per day (guessing), what does that work out to?

100 million dollars

Do we actually know the budget behind ChatGPT? I know Microsoft has thrown money at them and are providing a huge amount of cloud resources but beyond that I've missed any mention of the funding/costs.

ChatGPT is notorious for being damn sure about absolutely wrong things. If something like that powers the search…

All the book recommendations I got were for non-existent books. Makes you wonder if it creates summaries of the titles themselves.

> These AI queries are resource intensive.

But perhaps only for now. When youtube first started taking off, Google was bleeding cash on the resources needed to support it. Perhaps the same will be true for LLMs.

It's still unclear if YouTube is actually profitable.

It’s been clear since Google started breaking out YouTube in 2020 that YouTube is profitable.

do you have a source for YouTube profit numbers? all I've ever seen is revenue stats with zero details on cost.

Google’s inability execute is why they will lose the AI war. The fact the CEO has to write this article to wall street shows that they were caught offguard and have no real plans. If they had plans, why would Google management declare a code red over chatGPT?

Take the difference between Microsoft and Google. Microsoft just released tools that will make your life much easier. Intelligent Recap of teams meetings, ability to assign tasks based on what was discussed etc. Microsoft will announce bing with have chatgpt integration tomorrow. MSFT is eating Goog’s lunch and Sundar needs to go.

Sundar needs to go? And who will replace him? Larry and Sergey? They figured out the early web but the web of today is the whole another beast that is waiting to be understood.

Sundar needed to be kicked out years ago.

I would have kicked out Sundar after Stadia. Or failing to keep up Azure and Microsoft on cloud computing market share because he was late to game in building an enterprise sales team. Google needs to find a Satya Nadella to turn them around, i don't knfow if Larry or Sergey could that but anyone would be better than Sundar.

Google is on a collision course with regulators over search. If they are seen to be taking yet more content and knowledge from websites and "republishing" it they are going to find themselves with large fines and sanctions.

I think they are too "big" to really push AI, it needs smaller companies willing to take a higher risk.

Fair. Moreover, what are search or even content websites for if Google gobble up all that knowledge and distil it out.

But with China, Russia and India lurking, perhaps Congress might drag their feet on reigning AI in. Google Cloud are now poised to be going for some pretty decent govt contracts.

I know this wasn’t your point, but now I’m wondering if countries like India can afford GPU intensive search results.

It's an interesting conundrum because facts aren't copyrightable, but in a world with intelligent language models that can ingest facts and spit them back out in whatever phrasing or format is most useful to the user, facts are also the only valuable part of such websites.

On the flip side, trying to make facts copyrightable seems like a terrible idea for all sorts of reasons. For example, if facts were copyrightable, that would make online discussion of factual stories illegal, since it's hard to discuss a fact without revealing what it is. Also, it's not always clear who should get "credit" for a fact, since facts are by their very nature true independent of who first reported them.

Maybe journalism will eventually become similar to academia, where journalists are funded by governments and large corporations who have an interest in learning, and the resulting discoveries are (ideally) made freely available for everyone to access?

In case of Google it’s not as much about facts as about antitrust.

They can break antitrust laws if they go too much in a direction of delivering outright replies instead of redirecting to websites.

After all the buzz of decentralization, AI is going to be the new era of centralization. One AI interface to rule them all.

Thats a very interesting point, are all the web3 folks looking the wrong way?

There are a number of decentralized AI projects in that space. However, based on how effective web3 has been competitively against other big tech competitors for search and social media, they will need to be an order of magnitude more successful than on previous efforts.

Most projects are focused on privacy, user control and being censorship resistant. Which are all important, just not that important to the majority of everyday consumers who will take features and convenience over those other benefits. If decentralized AI is going to be competitive, it must actually be better along the lines of its actual features that enhance productivity.

Hey look, a decentralized chatbot running BLOOMZ-176B (an open source LLM about the size of GPT-3)

http://chat.petals.ml

I'm contributing to the project by running a node in my garage with a single RTX 3060ti in it, and you can too: https://github.com/bigscience-workshop/petals

It's early days, but the tech is super promising.

Thanks, I'm aware there are actually numerous decentralized AI projects, but until proven otherwise, their success is unlikely. Outside of cryptocurrency, there just hasn't been much success challenging the big tech institutions.

Decentralized search and social media never gained traction. It turns out that the majority of people just aren't that interested in privacy and freedom. Features and convenience still win. If decentralized AI is going to be successful, it will have to compete head to head for features and convenience. I do hope that it is successful.

Nonetheless, decentralization is just one major hurdle. I have numerous concerns about the application of the tech overall. Too much to put into this post, but if you are interested I've written much more here.

https://dakara.substack.com/p/ai-and-the-end-to-all-things

> Google Search is used so widely and its taxonomy of meaning so vast already that once they flip this on they’ll be unstoppable

Yeah, remember Google+? they tried very hard and resorted to very intrusive UX patterns, but they ultimately failed.

IMO this is similar, ChatGPT has a huge head start, and Google needs to do A LOT of work.

Yes, but will their revenue survive flipping that switch or will that kill the golden Adsense goose?

Fair question.

Depending on how slimy they’re willing to be I’d argue there’s even more scope to push ads when “advising” user through a grey area vs straight up keyword search.

Think going several steps ahead of initial q, increased trust and reliance on responses, multiple follow up questions with more ads etc.

I’m just envisaging this based on my time with ChatGPT these last few weeks.

Doesn't seem like too much of a stretch from where they are right now, which is showing ads at the top of search results formatted to look just like search results. They do have a small sponsored tag, but nothing stops them from doing the same on ai results.

This is a huge revenue opportunity. Google would not be investing in it if it wasn't.

> that once they flip this on they’ll be unstoppable. (Assuming any rate of improvement whatsoever as we search.)

Unstoppable as measured by what? More ad revenue?

They're already the #1 search engine due to two decades of general goodwill and a decade of monopoly power in mobile.

>its taxonomy of meaning so vast already

Can this be integrated for training though?

I think the new NLP technology made all their search technology obsolete. This company is in trouble.

Reply on news.ycombinator.com