Exameter

33, goat, seeking to become moth.

  • he/they/it

18+ but not explicit

It's been a lovely time on this website.


AD accounts
ask in my inbox or over discord
Website League
pleasetf.me/exameter
thanks for a great couple of years, everyone
crouton.net/

gnar
@gnar

So everyone remembers my mayonnaise post, right?

Well, in the third image of this tweet (reproduced below)

google bard conversation. Question: How many times does the letter "n" appear in the word "mayonnaise"? Response: The letter quot;nquot; appears 3 times in the word quot;mayonnaisequot;. One quot;nquot; is in the second position, another in the fourth position, and the last one in the seventh position.

The answer above is word for word identical to the response I have in the post. More over, checking out the user, @hypodronic, that google cites as the source of this information shows that @hypodronic did indeed rechost a share of the mayonnaise post, so I am somewhat convinced that Google Bard really is pulling from my post/@hypodronic’s rechost.

This is hilarious. I think we might've just permanently ruined this question for everyone who uses Google Bard and maybe Bing AI.

I'll note that my manual promptings of Google Bard to produce a good answer to the question weren't any better. Here's an example: The letter quot;nquot; does not appear in the word quot;mayonnaisequot;


atomicthumbs
@atomicthumbs

how the fuck is cohost's SEO as good as it is? any given thing on here tends to rocket to the top of Google results


Exameter
@Exameter

highly influential seclusive cabal of technoskeptic furries


You must log in to comment.

in reply to @gnar's post:

in reply to @atomicthumbs's post:

i assume another contributing factor is that lots of image based memes are often just normal text. everyone might be taking Screenshots of Funny AI Gives Wrong Answer but only cohost has Textpost of Funny AI Gives Wrong Answer

Cohost uses static pages and is pretty accessible (think a11y). Both of those things tend to give favor to SEO and search engines friggen LOVE fully compliant sites. Old scraping methods (like using <meta tags> or <meta description>) have either been obsoleted or have very strict rules in order to be used for SEO due to historical spam abuse. So simple is best!

Tumblr fails horribly with SEO because it is not very compliant and is in general just a mess of a codebase. I'm also speculating that because of their adult content restriction they add a layer of obfuscation to intentionally keep a lot of content out of SEO radar

(also I'm reading some people say that sharing chosts spins up a new page for them which is also SEO scrape-able, so that's a huge bonus!)