Anneke

Yesterday I saw someone boast that as a designer he can just feed his figma file to an AI and “instantly get a working product”. “It’s not vibecoding”, he wrote on LinkedIn, “I am guiding the software”.
This saves him so much time he now has time left for freelance work! …

What is it with these people who think their expertise is the one thing AI can’t do, while shouting from the rooftops it can do everything and replace everyone else? 🤪 is this how some people mentally deal with this stuff?

Mar 10 at 8:46amWeb
Show threadHein RagasMar 10
@Anneke It is my experience that people who use LLMs a lot get some kind of intellectual laziness, which causes them to stop inspecting their own thought processes and/or their biases. For them, it's just vibes all the way down. For us, we see someone who has lost the plot.
Show threadfnordMar 10
@Anneke I suspect it's because LinkedIn is part of the Microsoft circus that they enhance the reach of voices like this. I once reported a wild claim of "AGI in our times" that seemed from a real person only to get the LI service get back to me that it was an ad(!) that totally fell within the posting guidelines. #ThePurposeOfASystemIsWhatItDoes
Show threadAnnekeMar 10
@fnrd no this was actually a dude I personally knew as an intern for a company I was employed with 14 years ago. People are actually like this 🫠
Show threadfnordMar 10
@Anneke Ouch! Painful to hear actual people do that 🤦‍♀️
Show threadgabriele renziMar 10

@Anneke tbh it does seem that LLMs suck more at design then they do at writing code :)

IMO there is a kind of cross-competency-incompetence: it's easy for a designer to see that GenAIs can't design for shit, but you need a bit of skill to tell the code is shit.

Likewise, programmers can tell that code is shit, but have a hard time telling that a document is written terribly, cause they (we, I am one) are not trained on how to write well. Cue a million ai-slop blog posts. Repeat with any skill.

Show threadbovazMar 10
@Anneke them, probably:
Show threadAnnekeMar 10
@bovaz Wow! That describes how I feel very well :)
Show thread✅ Evert AlbersMar 10

@Anneke

Good question and discussion. I am trying to come up with a theory:

3 possible LLM outcome categories:
1) I could have done this myself but AI did it faster
2) It seems to work but I am not sure how
3) It sucks

The hard part is to be honest about whether your LLM outcome goes into 1) or 2).