alt Hacker News> the "taste-skill discrepancy." Your taste (your ability to recognize quality) develops faster than your skill (your ability to produce it). This creates what Ira Glass famously called "the gap," but I think of it as the thing that separates creators from consumers.
This resonated quite strongly with me. It puts into words something that I've been feeling when working with AI. If you're new to something and using AI for it, it automatically boosts the floor of your taste, but not your skill. And you end up never slowing down to make mistakes and learn, because you can just do it without friction.
Replies
I'm confused. I often say of every genAI I've seen of all types that it is totally lacking in taste and only has skill. And it drastically raises your skill floor immediately, perhaps all the way up to your taste, closing the gap.
Maybe that actually is what you were saying? But I'm confused because you used the opposite words.
This is Rick Rubin pretty much. He has 100/100 in taste, but almost 0/100 in skill.
He can't really play an instrument, but he knows exactly what works and what doesn't and can articulate it.
This is the disconnect between proponents and detractors of AI.
Detractors say it's the process and learning that builds depth.
Proponents say it doesn't matter because the tool exists and will always exist.
It's interesting seeing people argue about AI, because they're plainly not speaking about the same issue and simply talking past each other.
This is not what Ira Glass meant by taste gap. What he rather means is that taste is important. It’s what gets you into the field and what makes you stick around. Happy to be corrected on this.
I don't know much about Ira Glass and I'm not going to be a 5 minute wikipedia expert about it, so maybe I'm missing out on very relevant philosophy (I hope someone links the must read thing), but those would be very intentionally inverted meanings of the taste/skill dichotomy.
LLMs are good at things with a lot of quantity in the training set, you can signal boost stuff, but its not perfect (and its non-obvious that you want rare/special/advanced stuff to be the sweet spot as a vendor, that's a small part of your TAM by construction).
This has all kinds of interesting tells, for example Claude is better at Bazel than Gemini is, which is kind of extreme given Google has infinite perfect Bazel and Anthropic has open source (really bad) Bazel, so you know Gemini hasn't gotten the google4 pipeline decontamination thing dialed in.
All else equal you expect a homogenizing effect where over time everything is like NextJS, Golang, and Docker.
There are outlier events, like how Claude got trained on nixpkgs in a serious way recently, but idk, maybe they want to get into defense or something.
Skill is very rarely the problem for computers, if you're considering it as district from taste (sometimes you call them both together just skill).
If anything it's the opposite, except maybe at the very low end: AI boosts implementation skill (at least by increasing speed), but not {research, coding, writing} taste. Hence slop of all sorts.
There's no meaningful taste-skill gap in programming because programming doesn't involve tacit skills. If you know what you're supposed to do, it is trivial to type that into a keyboard.
The taste-skill gap emerges when you intellectually recognize what a quality creation would be, but are physically unable to produce that creation, and judge the creations you are physically capable of producing as low quality
The oft cited example is drawing a circle. Everyone knows what a perfectly round circle looks like, but drawing one takes practice.
It doesn't take practice to type code. If you know what code you're supposed to write, you write it. The problem is all in the taste step, to know what code to write in the first place.
This is exactly why I’m wary of ever attempting a developer-focused startup ever again.
What’s not mentioned is the utter frustration when you can see your own output is not up to your own expectations, but you can’t execute on any plan to resolve that discrepancy.
“I know what developers want, so I can build it for them” is a death knell proportionate to your own standards…
The most profitable business I built was something I hacked together in two weeks during college holiday break, when I barely knew how to code. There was no source control (I was googling “what is GitHub” at the time), it was my first time writing Python, I stored passwords in plaintext… but within a year it was generating $20k a month in revenue. It did eventually collapse under its own weight from technical debt, bugs and support cost… and I wasn’t equipped to solve those problems.
But meanwhile, as the years went on and I actually learned about quality, I lost the ability to ship because I gained the ability to recognize when it wasn’t ready… it’s not quite “perfectionism,” but it’s borne of the same pathology, of letting perfect be the enemy of good.