Spyke

Syndicated from the fediverse. Read and engage on the original instance.

View original on feddit.online

13 replies

MurrayL
lemmy.world

How much AI was used to generate this website? The text certainly seems like it came straight from an LLM, which is ironic for a site celebrating the hand-made web.

7
hansoloreply
lemmy.today

Look at the source and look for explanations behind a # tag, that's usually how Claude and CGPT write html.

Edit, within < and >. I was thinking of bash with the #

6
MurrayLreply
lemmy.world

Yep, page source is littered with verbose and unnecessary comments that no human would bother making.

6
MobiusFlightreply
feddit.online

Fair. I leaned on AI, for the code and for picking and drafting a lot of the recent additions. The comments in the source were its working notes; I've stripped them from the build. I overstated "by hand" in the post: the rules and the final say are mine (no algorithm, no tracking).

1
Prataireply
piefed.ca

There’s so little text, if it’s done with AI- that’s the epitome of lazy.

1
Coelacanthreply
feddit.nu

The only thing rivalling mankind's capacity for evil is its capacity for laziness.

3
hansolo
lemmy.today

OP, please clean the LLM comments out of your html. You're gonna get flamed for that.

Case in point, staring on line 16:

  The one place a domain is hardcoded, and it has to be.

  Two hosts serve identical content - culdesac.site and the vercel.app
  subdomain, which stays working alongside on purpose. To a search engine
  that is duplicate content, and without this it picks a winner itself and
  may well pick the wrong one. A canonical URL is absolute by definition,
  so this cannot be relative the way the internal navigation is.

  If the primary domain ever changes, it changes here and in
  public/sitemap.xml and public/robots.txt together. check-canonical fails
  the build if those three disagree.
-->
4

You reached the end