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

Compare Google, Bing, Marginalia, Kagi, Mwmbl, and ChatGPT

danluu.com

Post reply on HN
Print view View original

81–90 of 493 posts

Re: Compare Google, Bing, Marginalia, Kagi, Mwmbl, and ChatGPT

#81

Post by execat »

teeray wrote: post #73
endisneigh wrote: post #5

I reckon these days search is pretty difficult and everyone knows how to game it. I recommend using a search engine that lets you effectively change which sites are shown. You can do this with Kagi, or with Google's Programmable Search Engines - I'm sure there are more too. In particular I block Youtube, not because they aren't sometimes correct, but because I don't want videos polluting the regular results - it just…

> it just takes too long to get info from videos. I can’t wait until video transcripts get fed into LLMs just to eliminate the whole “This video is sponsored by something-completely-unrelated, more about them later. What’s up Youtube, remember to like, share, subscribe… 5 entire minutes pass on similar drivel … the actual thing you want, but stretched out to an agonizing length ”

You need SponsorBlock.

Usually people leave a "highlight" marker which tells you where you're supposed to jump to. Along with the regular "This video was brought to you by VPN".

Re: Compare Google, Bing, Marginalia, Kagi, Mwmbl, and ChatGPT

#82

Post by boomboomsubban »

teeray wrote: post #79
boomboomsubban wrote: post #69

Can someone tell me why Bing, and thus DDG, has switched to prioritizing local results? I'll search the most inane things, like lyrics to a song, and get results for local businesses containing maybe one word in common. It's most frustrating with phone numbers. I picked up the habit of searching the random numbers that called me, to try and find out if they were possibly important. I used to get a bunch of spam sites…

> I'll search the most inane things, like lyrics to a song, and get results for local businesses Query: “I’m coming out of my cage…” Result (Ad): “You’ll be doing just fine with these amazing year-end closeout prices at Al’s Discount Car Barn. Gotta come down—you’ll want it all!”

Ads would make sense, but there's no way my local city council is paying Bing and they are the most frequently listed result.

Re: Compare Google, Bing, Marginalia, Kagi, Mwmbl, and ChatGPT

#83

Post by Gualdrapo »

boomboomsubban wrote: post #69

Can someone tell me why Bing, and thus DDG, has switched to prioritizing local results? I'll search the most inane things, like lyrics to a song, and get results for local businesses containing maybe one word in common. It's most frustrating with phone numbers. I picked up the habit of searching the random numbers that called me, to try and find out if they were possibly important. I used to get a bunch of spam sites…

Maybe it was an attempt to make better their results for local results?

When searching for results from my country in DDG (picking the country in the drop-down below the search box) still returned results from the USA or other countries. Even when searching stuff in the local language. Maybe they tried to fix that because it really sucked, so much I never used it again for searching into local websites.

Re: Compare Google, Bing, Marginalia, Kagi, Mwmbl, and ChatGPT

#84

Post by tentacleuno »

robrenaud wrote: post #54
ysavir wrote: post #38

Earlier quoted context omitted.

> In particular I block Youtube, not because they aren't sometimes correct, but because I don't want videos polluting the regular results - it just takes too long to get info from videos. Funnily enough, lately I've been prioritizing YT videos more when searching. So many sites now are just regurgitated SEO farms with minimal quality, and easy to see why: it's minimal effort to produce and cheap to host. But making a…

Would a browser feature that skipped to the relevant parts of the video based on closed captioning and understanding search intent be useful? It seems like this would be a good way for Google to fight to stay relevant in UX vs having the chat bots just quickly spitting out a readable answer. Hunting through ad laden webpages is annoying. Seeking to the relevant section of the video is a solvable problem, especially f…

> Seeking to the relevant section of the video is a solvable problem

...and it has already been solved, though partially: SponsorBlock allows people to add a "Highlight" section to a video, which denotes the part of the video which the user most likely wanted to see (sans the "what's up guys", "like and subscribe", etc.)

Of course, it's not perfect: it relies upon humans doing the work, though some may see that as a positive over something more computerized.

Re: Compare Google, Bing, Marginalia, Kagi, Mwmbl, and ChatGPT

#85

Post by skygazer »

boomboomsubban wrote: post #69

Can someone tell me why Bing, and thus DDG, has switched to prioritizing local results? I'll search the most inane things, like lyrics to a song, and get results for local businesses containing maybe one word in common. It's most frustrating with phone numbers. I picked up the habit of searching the random numbers that called me, to try and find out if they were possibly important. I used to get a bunch of spam sites…

If you’re going to search for phone numbers you’ll want to ensure you enable verbatim searching under tools on Google, and put the number in quotes, perhaps in “xxx-xxx-xxxx” OR “(xxx) xxx-xxxx” forms. Many of the sites you mention are fake sites with fake contacts just for ad serving, and I’ve read in some few cases the scammers seeded the spoofed numbers they appear to call from on to the sites they control to see who googles their phone numbers.

Re: Compare Google, Bing, Marginalia, Kagi, Mwmbl, and ChatGPT

#86

Post by bombcar »

marginalia_nu wrote: post #11

While I've made huge improvements to the algo recently, I do think Marginalia Search got a bit lucky with the sample queries, as it is still IMO far more hit and miss than many alternatives, but that also speaks for how hard evaluating search quality is. Its efficacy is also strongly dependent on understanding that it's a keyword search engine with no semantic understanding.

I notice you completely avoid the question on how a single developer can do so well ;)

I do think that search has gotten much worse but my ability to know the magic words like “ublock origin” instead of “Adblock” and “yt-dlp” instead of “download YouTube” and phrase my search has gotten better.

We’ve all been doing prompt engineering against the Internet-wide LLM that is the spam houses.

Re: Compare Google, Bing, Marginalia, Kagi, Mwmbl, and ChatGPT

#87

Post by tentacleuno »

marginalia_nu wrote: post #11

While I've made huge improvements to the algo recently, I do think Marginalia Search got a bit lucky with the sample queries, as it is still IMO far more hit and miss than many alternatives, but that also speaks for how hard evaluating search quality is. Its efficacy is also strongly dependent on understanding that it's a keyword search engine with no semantic understanding.

> [...] but that also speaks for how hard evaluating search quality is.

Would you be able to share some of your personal highlights regarding this?

I've partially kept up-to-date with the DIY, non-corporate search space (YaCY and friends). I'd love to understand a bit more behind the engineering decisions made when creating a search engine; it seems like a very hard problem to solve.

P.S. Marginalia is a very impressive piece of work, overall -- I've heard nothing but positive remarks from users on here. I've been meaning to try it for a while, but time constraints have... well, constrained, thus far.

Re: Compare Google, Bing, Marginalia, Kagi, Mwmbl, and ChatGPT

#88

Post by boomboomsubban »

Gualdrapo wrote: post #83
boomboomsubban wrote: post #69

Can someone tell me why Bing, and thus DDG, has switched to prioritizing local results? I'll search the most inane things, like lyrics to a song, and get results for local businesses containing maybe one word in common. It's most frustrating with phone numbers. I picked up the habit of searching the random numbers that called me, to try and find out if they were possibly important. I used to get a bunch of spam sites…

Maybe it was an attempt to make better their results for local results? When searching for results from my country in DDG (picking the country in the drop-down below the search box) still returned results from the USA or other countries. Even when searching stuff in the local language. Maybe they tried to fix that because it really sucked, so much I never used it again for searching into local websites.

This is the one area it still ignores my location. I live in a town named after a UK city, there's several bigger towns in the US with the same name. I just searched "McDonalds city name." I got results for the locations at least half the US away from me, as well as Uber Eats GB.

Re: Compare Google, Bing, Marginalia, Kagi, Mwmbl, and ChatGPT

#89

Post by imiric »

ysavir wrote: post #38
endisneigh wrote: post #5

I reckon these days search is pretty difficult and everyone knows how to game it. I recommend using a search engine that lets you effectively change which sites are shown. You can do this with Kagi, or with Google's Programmable Search Engines - I'm sure there are more too. In particular I block Youtube, not because they aren't sometimes correct, but because I don't want videos polluting the regular results - it just…

> In particular I block Youtube, not because they aren't sometimes correct, but because I don't want videos polluting the regular results - it just takes too long to get info from videos. Funnily enough, lately I've been prioritizing YT videos more when searching. So many sites now are just regurgitated SEO farms with minimal quality, and easy to see why: it's minimal effort to produce and cheap to host. But making a…

> But making a video takes time and effort, so has a much higher barrier to use as a click farm.

> The mainstream channels aren't all that useful, but there's a ton of "old web" style videos (some even recent) passionately providing details for almost anything you'd think to search. And they're a gold mine.

This won't be the case for long. YT is already starting to be polluted with spam and AI generated content, which will get more and more common. The same thing that happened to the web in text form, will happen to videos.

I think the only solutions are using allowlists for specific domains, and ironically enough more AI to filter specific results. Or just straight up LLMs instead of web search, assuming they're not trained on spam data themselves.

Re: Compare Google, Bing, Marginalia, Kagi, Mwmbl, and ChatGPT

#90

Post by bombcar »

taberiand wrote: post #61
johnfn wrote: post #17

I noticed that the author uses ChatGPT3.5 rather than 4, which is a rather large difference. I don't have the knowledge to rerank all questions the author asked, but I will say that a test of ChatGPT 4 leads me directly to youtube-dl, which is better than every other search engine listed.

That was the first thing I checked reading the article. Although the argument would be 3.5 is free - any comparison of systems against ChatGPT that isn't using ChatGPT 4 can be dismissed almost out of hand; there is not much point talking about ChatGPT if it's not using ChatGPT 4 and making proper use of its capabilities. That is not to say that there aren't valid criticisms of and shortcomings in ChatGPT 4 - just th…

He gives the full queries - do you have chat 4.0 that you ran run it against?
Post reply on HN