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

The next Google

dkb.io

Post reply on HN
Print view View original

221–230 of 561 posts

Re: The next Google

#221

Post by thwayunion »

riazrizvi wrote: post #173
thwayunion wrote: post #104

Earlier quoted context omitted.

Sorry, a typo of LLM. gpt-3 et al.

Oh AI is the future of search? Maybe. I think the bigger reason why search isn’t getting much better is that incremental technical improvements are being offset by Google trading their dominance for greater shareholder value. A safe business can take away or buy-and-bury features that are better for the customer but worse for the business’s top line. Things like the Power Search API, or search that integrates with ot…

NLP generative models that you can interrogate with natural language prompts, and which generate coherent responses together with citations for sources, is definitely the future of search.

I don't know anything about AI or the relationship between market participation and innovation. All I know is that LLMs are a useful tool that is already displacing search in many parts of my life.

Re: The next Google

#223

Post by ezekg »

wormer wrote: post #218
egeozcan wrote: post #54

Earlier quoted context omitted.

> Give me relevant links to my term, with no SEO spam I think achieving this would be worth billions, at least? We really tend to underestimate the amount of spam on the internet. Google does badly here: it overcorrects and you get only super safe results. Yandex? Unless you get lucky, lots of spam.

I think it would probably be worthless. SEO means money spent to the search engine company, and why improve UX when you could buy your seventh super yacht?

Not every search engine needs to be an ads company, though.

I'd pay /mo for a good one. Why not? Currently trying Kagi.

Re: The next Google

#224

Post by Melatonic »

mostlysimilar wrote: post #74
Barrin92 wrote: post #45

I think 'more customization' which is a theme with a lot of these alternatives is a fundamental dead end. I don't know where this persistent myth comes from that people love choice and tinkering, because they clearly don't. There's a huge cost associated with having to make choices, and one feature of successful modern apps is that they're frictionless. That's why TikTok is so successful. There's no login, no user ch…

> a simple box where you put a question in and it gives you answers Google is very good at this, but this is exactly what I _don't_ want in the "Next Google". I want a _search engine_ for the web, not an answers engine that tries to know what I want better than I do. Search the web. Give me links to websites. This seems obvious to me, but everyone is trying to be like Google. I'm convinced the death of the web and in…

Google still searches forums pretty decently - I think what you are describing are two separate phenomena

1. Yes Google search has gone to shit - even putting stuff in quotes now does not do an exact search (there is another checkbox you ALSO need to use for that). It tries to be too smart even when no user is logged in.

2. Giant mega forums like Reddit have really taken over. Instead of a dedicated forum people just go to subreddits. Personally I think it is good and bad and I still try to actively participate in both.

Re: The next Google

#225

Post by baxtr »

idiotsecant wrote: post #100
mostlysimilar wrote: post #74

Earlier quoted context omitted.

> a simple box where you put a question in and it gives you answers Google is very good at this, but this is exactly what I _don't_ want in the "Next Google". I want a _search engine_ for the web, not an answers engine that tries to know what I want better than I do. Search the web. Give me links to websites. This seems obvious to me, but everyone is trying to be like Google. I'm convinced the death of the web and in…

The death of the web is because most people don't want what you want. They don't mind walled gardens, so long as they are easy to use and have the content and connections that they want to see. The audience of HN is extremely skewed towards preferring systems that allow tinkering but that's not what the market wants.

This. People don't realize that the early web was elitist. Now, the entire population is online. And, as you said, most people simply don't care about the stuff we care about.

That's also why "Google's search results are soo bad." They're not. For the bulk of Google's visitors, they're good enough.

Re: The next Google

#226

Post by mountainriver »

thwayunion wrote: post #41
hubraumhugo wrote: post #37

Google won because they started early and had a right algorithm. They then had the scale to grow with the internet. It would be very hard to start a new general-purpose search engine now (capital, monopoly, tech) Nobody seems to have The Next Big Idea for a better search engine yet.

LMMs are the next big idea for search.

Yup ML is definitely the future of search and google is on the cutting edge of that so I expect they will continue to dominate

Re: The next Google

#227

Post by mullr »

Barrin92 wrote: post #45

I think 'more customization' which is a theme with a lot of these alternatives is a fundamental dead end. I don't know where this persistent myth comes from that people love choice and tinkering, because they clearly don't. There's a huge cost associated with having to make choices, and one feature of successful modern apps is that they're frictionless. That's why TikTok is so successful. There's no login, no user ch…

For Kagi, at least, there's a very well integrated search customization method that they didn't bother to show here. For any search result, you can add a ranking adjustment for the site it came from. This is directly in the results, so it's very accessible, and quite easy. One of the choices is 'pin', which is fantastic for technical work: 'sqlite.org' is now boosted over everything else, for me, and it's exactly what I want. I could just as easily take it out, if it becomes a problem.

Re: The next Google

#228

Post by Nextgrid »

marginalia_nu wrote: post #86
sdoering wrote: post #77

Earlier quoted context omitted.

Really? Or are you the one I should have refrained from feeding. But if you must know: First you need to collect a lot of content from the internet. From many different sites. With very different types of code structure. Broken html. More often than not behind some SPA JS code. Behind robots.txt files and bot protection efforts. So the first problem to solve would be building a crawler at scale. That is able to crawl…

All of this is a long series of solvable problems. I should know, I've dabbled in solving most of them. This is why I suggest actually taking a stab at it before you dismiss it as impossible. There are some problems that aren't as big as they seem. Parts of an SPA can't be reliably linked to anyway even if you find interesting text there, so you can just leave them out of the index. Likewise, there isn't as great of…

Out of curiosity, how much disk space does your index currently use, and what's the storage hardware (SSD or spinning rust)?

Re: The next Google

#229

Post by b8 »

Kagi requires for an account to be registered and will charge a monthly rate when it's out of beta, so I don't really think that it'll be a threat to Google.

Re: The next Google

#230

Post by notriddle »

simion314 wrote: post #131
Barrin92 wrote: post #45

I think 'more customization' which is a theme with a lot of these alternatives is a fundamental dead end. I don't know where this persistent myth comes from that people love choice and tinkering, because they clearly don't. There's a huge cost associated with having to make choices, and one feature of successful modern apps is that they're frictionless. That's why TikTok is so successful. There's no login, no user ch…

>I don't know where this persistent myth comes from that people love choice and tinkering, because they clearly don't. People want to accomplish stuff, this means they need tools to do stuff and if a tool can be customized to do the stuff faster or better people want the customization. At my job we have paying users that requests features that indeed are work related(not moving shit around). I know GNOME-minded peopl…

The biggest problem with GNOME, and with Mozilla, and with almost everyone who's commenting on their choices, is that all of them are shuffling deck chairs around on the Titanic.

The Titanic didn't sink because of the arrangement of deck chairs, and Mozilla didn't sink because of any features they did or didn't provide (and GNOME didn't fail to achieve significant market share in the first place because of features either). The actual problem doesn't have anything to do with the stuff on the deck at all.

It's the one-two combination of vendor-lock in and bottomless marketing budgets. Since most of the value of the Windows platform and the Web is the immense amount of stuff that's built on top of it, there's a huge lock-in effect that prevents you from even reaching parity, much less exceeding it. And in order to overcome the marketing budget and pure inertia, you need to be ten times better, not just on par.

If GNOME becomes as usable as Windows, it won't have anything to do with what they actually do in the desktop environment itself one way or the other, whether it's continuing on the road they take now, or reverting everything back to the way GNOME 2 was, it's totally irrelevant. GNOME becoming usable will be entirely because of Valve investing in Wine, combined with a whole bunch of other apps moving to Electron and shipping Linux versions because heck why not?

Unfortunately, while they are probably already on par with Windows, they aren't ten times better than the Mac:

* The Mac has a bottomless marketing budget. Good luck competing with that, GNOME.

* They've shown a lot more restraint than Microsoft has, probably because macOS is considered a niche product to round out their catalog rather than being their one and only operating system like Windows is for Microsoft. They have even reversed course on a few anti-features, like adding back USB-A ports to the Macbook Pro even though it made the laptop slightly thicker. And unlike Windows RT, they didn't lock down the ARM Macs.

* Those tectonic shifts I mentioned that made Linux usable? They also make the Mac usable, because Wine is open source and Electron is basically its own operating system. Anything truly good that GNOME does, Apple can copy it just like Chrome copied all the really good stuff from Firefox.

Post reply on HN