[atotic]

I hope Goggles can be used as a ranking signal to help default search engine deliver better results.

Google has never had the appetite for letting users explicitly improve its results. Their reasoning was that pure algorithm is the only possible solution that scales. Even the minor features, such as "I never want to see this site", were removed.

I am glad Brave is trying out something different. Crossing my fingers that it works, the current state of search is so meh..

[spaniard89277]

I hope forums gain some weight after google decided to f** them over in the results. IDK in English but for spanish language forums are still valuable, but you need to know the domain name so that hurts discoverability.

[aliasxneo]

I just used Brave for the first time and was absolutely delighted to see they have a dedicated "discussions" section when you search for something. It's literally what has sold me to trying it as my default engine for a few weeks.

[Agamus]

We users are a funny lot. I just used Brave for the first time, and the very first thing I did was look for how to turn off "discussions".

[danuker]

At least you know they're there, and you could turn them off.

[Nemi]

I want this but can’t find it. I want to find discussions on something (toyota sienna) but when I search I don’t see any discussions section. Even putting “discussions” in the search terms does not bring it up. Looks like something they should make toggleable.

EDIT: I take this back. If I put the term “discussion” in the search along with “toyota sienna”, I see a “discussions” section in the search results if I am on the “goggles” tab. I don’t see the discussions section if I omit the “discussion” term though.

[spaniard89277]

Mmm, gonna try, thanks.

Hmmm, can't find that discussions tab.

[lrae]

It's not a tab, it's a section of the default search results page. And it also is just Reddit. At least for the couple of searches I just tried it for.

[GeorgeWBasic]

Same, I've only ever seen reddit in that section. Helping reddit's hegemony is counterproductive, not helpful.

[throwaway81523]

> I hope Goggles can be used as a ranking signal to help default search engine deliver better results.

Ranking is already treated as predicting which of the search matches the user is most likely to want to see. Clickbait gets highly ranked because users do in fact click on it. Goggles (which sound like a great idea) apparently let you somewhat control the prior probability distribution as an input to the ranking algorithm.

I wonder what features it lets you select on. E.g. we had a thread here recently about finding useful product reviews. It would be great to have a goggle that downranks any site containing amazon affiliate links, since those are almost always basically shill sites despite being "reviews".

[oDot]

That is incorrect. Many high quality reviewers make money through affiliate links. For example, In 2020 LTT made 9% through Amazon Associates[0]

[0] https://linustechtips.com/topic/1270087-linus-media-group-ma...

[throwaway81523]

Yes there are some exceptions. Maybe I should have said "most of" instead of "almost all", but the thing is that it's common and frequent enough that I want to bypass that style of "review".

[hedora]

Google has had a large team that manually curates results for a very, very long time.

I have no idea why they aren't able to flag and block spam domains anymore.

[slig]

Seriously how hard must be to flag duplicated content from GitHub, Wikipedia and Stack Overflow? Seriously, last week I had a query that returned the exact same content three times on the same page: the first result was a GH issue, and the other two were spam websites with the same content from GH. Google used to be good at flagging and downranking duplicated content, but now they don't even care, it's a shame.

[davidatbu]

I'm always amazed at such comments that trivialise the challenges of fighting spam and SEO while being the most dominant search engine by far (which means spam and SEO targets you by far).

If it is so easy, why don't you build a dominant search engine that manages to complete avoid the "totally easy" spam you talk about?

If perhaps, you have extensive experience in the anti-spam section of a dominant search engine (or some similar position), I'd give your comment more weight. Do you have such experience?

[GeorgeWBasic]

What exactly is difficult about detecting direct copies of just those three sites? Because that's what's out there: exact copies. If Google isn't filtering those out, it's because they don't want to.

[davidatbu]

Google Search is most likely an incredibly complex engineering feat with many moving parts.

Unless you provide me with your qualifications and/or previous experience that would lend credence to your claim, I will view your claim as one from an armchair critic/engineer.

[slig]

They used to be good at it, now they seem bad. How can that be? How hard is to not show three exact copies of the same result for the same query on the first page?

And yes, I'm being "an armchair critic/engineer", like everyone else criticizing Google or any other big company that seems to be getting worse at what they excelled.

[davidatbu]

To be clear, I am not labeling your commentary armchair criticism because you are supposing that spam on Google is more prevalent in recent times, but because you are so confidently asserting that it is easy to fight spam at the scale and quality Google is operating at.

Since you haven't provided any qualifications/past experience, I am not convinced at all that it's as easy as you say it is.

[slig]

Fair enough.

[shepherdjerred]

It’s not easy, but it’s hard to believe that a company with Googles resources and dominance can’t build a good solution.

[davidatbu]

The only constant in my journey in comp sci is my underestimation of how much time and effort things take, so I guess I may be trying to compensate for that.

Incidentally, if anyone has insight into how to improve that aspect of my job, that would be much appreciated!

[esperent]

I don't think so, can you provide a source for this?

[not2b]

The web is much too big for that to be practical.

[raxxorraxor]

Google has problem with generated "marketing" sites whose content gets filled by bots. They also use the latest SEO tricks to get ahead. This is something any small private internet presence will never have while click-farms will.

Additionally we hide information behind proprietary platforms like Discord that cannot be crawled.

Ironically Googles research of text AI undermines their search algorithm because now spammers use said AIs to get good page ratings. A lot of sites have ever the same content with a few different words here and there. Especially new media content get immediately swamped with such sites since there are no established sites under certain key words. A domain filter sadly won't work as well since there a so many scamming sites.

[niutech]

> Google has never had the appetite for letting users explicitly improve its results

They used to ask some users for feedback back in the good days: https://searchengineland.com/google-feedback-experiment-whic...

[asiachick]

I feel like Google would do better by letting users register filters. For example I'd filter out ranker, vulture, and similar sites. That would seem like more signal for Google than not having that user preference. More signal = better searches = more searches = more money.

[niutech]

You can easily add that filters by yourself: just append "-site:ranker.com -site:vulture.com" to your search query, e.g. by setting the search URL to: "https://www.google.com/search?q=%s+-site%3Aranker.com+-site%..."