[chubot]

I feel like one of the hardest problems in tech is making search engine that credibly competes with Google.

It's not just because the tech is hard (and it is), but also because they have some of the strongest network effects, user lock-in, and distribution deals.

It sounds almost dumb/crazy for people to try, but there is a newer effort from Europe that I saw on HN recently (i.e. they've been working for ~5 years and are just starting to publicize). And I think PG has mentioned it a few times -- i.e. encouraged people to take on Google :)

BTW I think "schleps" are different than "hard startups". I think schleps are about raw effort where you kind of "know" you'll succeed at the end. It seems like most companies are hard because you don't know if what you're building will succeed, even if you build it.

edit: "new" search company is cliqz: https://news.ycombinator.com/item?id=21724191

[chris_f]

"It sounds almost dumb/crazy for people to try..."

Timely comment. I hesitantly shared a "Show HN" this morning for an early search engine I have been working on.

I almost feel silly when I tell people about it because I realize it is such a crazy idea. It's obviously nowhere even close to being a competitor to anyone today, but I figure why not.

I have been lucky with two prior highly vertical startups, so I decided that my next project would involve asymmetric risk.

[jacquesm]

You should definitely connect with Chuck McManis.

https://news.ycombinator.com/user?id=ChuckMcM

[erikpukinskis]

My guess is that the company that beats Google will be the one that lets users create and monetize their own multifarious search engines that each has an editorial slant and an aptitude in a specific area.

Essentially, Google is CBS/NBC/ABC and their value is based on their name being the destination. They’ll be taken down a peg by a YouSearch (née YouTube).

Although Google, like CBS/NBC/ABC, will still exist they’ll just lose half the market (the middle tail).

What I’m describing is a technical challenge, but not like Nuclear 2.0 is a technical challenge. You could create an MVP for a niche market in a year or so. I think it’s mostly just hard work, doing marketing, understanding early adopters, doing the tech support, having a good feature-request-to-patch pipeline, managing investor expectations, etc... all the normal grunt work of a startup.

The search doesn’t even have to be that good if you can search psychological spaces Google could never prioritize.

[tim333]

>dumb/crazy for people to try

I thought so but Duck Duck Go did ok going for a niche with a privacy focus. I know they don't really compete with Google but it shows you can start niche wise and potentially build.