Agree with the efficiency framing — the aesthetic homogeneity is
downstream of the fact that these are side projects and LLMs are
faster than crafting a design system from scratch.
The more interesting question the post raises, at least for me,
is that distribution platforms like Show HN, Product Hunt, etc.
were designed for an era when launching something was costly
enough to be a signal. When a weekend project can ship a
production-looking landing page, upvotes on these platforms start
selecting for whatever catches the eye fastest, not whatever
actually solves a problem. The signal degrades.
I've been thinking about this a lot because I'm building a
directory where you have to rank 5 other projects before you can
post your own — trying to see if forced engagement produces
better signal than one-click upvotes. Too early to say if it
works, but I do think "how do we find the good stuff under the
slop" is the real problem and it probably isn't solved by
detecting AI design patterns.
The more interesting question the post raises, at least for me, is that distribution platforms like Show HN, Product Hunt, etc. were designed for an era when launching something was costly enough to be a signal. When a weekend project can ship a production-looking landing page, upvotes on these platforms start selecting for whatever catches the eye fastest, not whatever actually solves a problem. The signal degrades.
I've been thinking about this a lot because I'm building a directory where you have to rank 5 other projects before you can post your own — trying to see if forced engagement produces better signal than one-click upvotes. Too early to say if it works, but I do think "how do we find the good stuff under the slop" is the real problem and it probably isn't solved by detecting AI design patterns.