Post

Annoying, but also pretty sure it's on purpose for optical alignment because our brains expect the side of "B" to be slightly tapered

yeah - it's like the O being larger on purpose, right? thing I don't get here is that the B alignment is so minuscule that it could almost be believably unintentional 😅

Yeah, same principle. It very well could be unintentional, but at the same time anything larger would stick out too much + they've had however many decades to refine their logo. Just my 2¢ tho :)

oh 100% same. I wish I had good design chops, since most of the stuff I end up "designing" looks like this lmao

Personally I'm not a big fan of AI tools for design or coding (and I say this as someone who's tried a *lot* of AI tools for coding, but not many for design). I think this post describes how I feel... (1/4) bsky.app/profile/did:...

As far as coding goes, the current crop of LLMs remind me of very savvy junior devs — the type of people who are very smart and capable of following established patterns but lack the experience and perspective to make the right architectural and engineering decisions about a greenfield project.

(2/4)... And as someone who's building a polished app for users that we want to keep around for a long time, I feel like starting from a rocky base (especially for something that I may not be as skilled at!) it setting our team up for future failures rather than getting a human to design things...

(3/4) ... from the ground up from a far more experienced and nuanced perspective rather than something that looks good on the surface but doesn't take into account small idiosyncrasies that AI can't catch (yet) -- and those add up in the long term. For short projects I may consider it, but that's..

(5/4) [addendum] I'm very grateful to have people on my team as well who are far more experienced in design than I am now -- if I were still doing this solo, there's a good chance I would've reconsidered my stance on this.

think we have the same sentiments - with AI, it's great for assisting me on things I can do but want to save time (data manipulation scripts, etc.) where there are binary right/wrong results but when there's nuance, I'm not keen

100% agreed on both above comments! I think that the potential is there, just not for my workflow or use case. If I was starting on a new project from scratch, I could see myself bouncing ideas off of AI -- I still rubber-duck debug with Claude sometimes if I get stuck on something :P

one current thing i like to do is rubber-duck a11y implementation ideas back and forth to see what it comes up with in terms of nuances etc. was having this conversation earlier with someone about how it is still very much on the driver the danger of AI is it can amplify habits, good or bad

1 repost1 like

No replies yet.