I'm working on a new search engine and directory project, where I'm making a really funny noodle joke.
Actually, the whole site is funny. The idea is that it brings you together with really amazing free content you can't see anywhere else.
The slogan is going to be ‘open your noodle.’ Think Yahoo, but without the impossibility of keeping a billion human links curated.
The site's going to be called Lokshen, which translates into ‘noodle’ in English.
Editorial standards instead of SEO
I'm building a search engine for indie creators like nothing the internet has seen for a minute. It spiders websites by following rules, and limits the amount of intake it takes in. It's polite and responsible.
The system has editorial standards based on quality and entertainment value, and it actively downranks content if it detects onsite SEO.
Boring websites, corporate websites, ecommerce websites with no editorial bent... banned.
Your uncle's fishing blog where he goes on about the technical details of finding and processing crawfish? A+ thrown to the top of the queue.
Got a stupid MIDI playing in the background of an ‘I will be an emo goth tyrant until I'm 50, which is like nine weeks from now’ web page on Neocities? BOOSTED!
Every website gets a grade letter from F+ to A+ and the ranking is fair.
The idea is to fuck with and remove the normal incentives that get people to cheat with traditional SEO.
Not to say the algo can't be gamed. You can do that by building more engaging content, because the spider does not care about your products unless you have something interesting to say about them.
You can see why your website was rated the way it was on a special bots page, where you can see the feedback and read recommended changes.
The news section will only feature good news and inspiring articles. Comfort food recipes will come in 10 a day, and be displayed on the home page based on freshness.
There's a trivia pirate who loses his mind by Friday, and a spiritual weather report based on site activity.
It's the place to ungoogle your noodle.
The classifier and the Professor
Last night, I set up the classifier and the reviewer on the spider piece of the search engine. The classifier is using a System1-style model, while the reviewer, who we're calling ‘The Professor,’ is using an LLM.
And they're arguing with each other! This is perfect.