I've been chewing on this for a while, but I don't think the thought is fully formed. I'll rough it out here anyways, it certainly isn't a novel or unique one.
I think my main issue with vibe coded apps is that the signal to noise ratio got flipped upside down.
Things that may have been positive signals in the past aren't really great indicators anymore. Maybe they shouldn't have been positive in the past, but they certainly aren't now. Broadly, things like polished website design, punchy language, clean branding, slick landing page, etc were all indicators. At a minimum, they at least meant that time and effort was put into it. It didn't really mean anything about the product, but it would display at least some level of time and effort.
Now, the outward appearance of a project no longer conveys anything about effort or durability, I can easily spin up a home page that would have taken me weeks pre-AI, then never touch it again.
The site states that it took roughly a hundred hours to develop. I have no real way of verifying that and I don't even think that it matters. The most evident signal that I feel I can still get a pulse on is when something is clearly, heavily AI written. All other signals that would supersede it require additional time to evaluate. Things like reading the docs, skimming the codebase, testing the product etc.
With a project like this, the ask made to the end user is to spend the time to, at a minimum, read the docs to see if it's a good fit. If they think it's a good fit, the next ask is to spend the time to integrate it with their current inventory. Then learn the workflow. Then learn the idiosyncrasies. All of these things involve, real human, time.
When an end user is asked to invest this time, there are only so many signals available to them to help decide if it's a good use of their time, energy and resources. For better or for worse, it seems like that list of signals is shrinking and right now, something that appears vibecoded is still one of the clearest visible signals.
I think it's hard to say one way or another. My guess is the latter, or perhaps the audience isn't even aware it's vibe coded.
I (like many other software folks) are aware of AI-isms in writing, so when I see something that it so obviously AI generated in a project, it's equivalent to a giant red waving flag.
For someone who doesn't interact with LLMs regularly? I have no idea, maybe it sounds great to them. Most non-tech folks I interact with to don't seem to have as well-tuned AI sense, whether it be image, video or text
Effective altruism. A lot of the folks working on AI at large tech companies are disproportionately represented in the movement. There's a lot of overlap between EA and the rationalist community as well. The wikipedia page is a good place to start https://en.wikipedia.org/wiki/Effective_altruism
On a recent trip to Japan, my partner and I visited the Kyoto Railway Museum (https://www.kyotorailwaymuseum.jp/en/
), which features many of the trains and predecessors listed in the OP.
Seeing it at the end of the trip after riding so many different trains made it especially impactful. Two highlights were the massive 100ft-wide railway diorama and the working vintage steam trains, which they drive (run?) out onto a turntable before rotating and returning to their stalls.
I think my main issue with vibe coded apps is that the signal to noise ratio got flipped upside down.
Things that may have been positive signals in the past aren't really great indicators anymore. Maybe they shouldn't have been positive in the past, but they certainly aren't now. Broadly, things like polished website design, punchy language, clean branding, slick landing page, etc were all indicators. At a minimum, they at least meant that time and effort was put into it. It didn't really mean anything about the product, but it would display at least some level of time and effort.
Now, the outward appearance of a project no longer conveys anything about effort or durability, I can easily spin up a home page that would have taken me weeks pre-AI, then never touch it again.
The site states that it took roughly a hundred hours to develop. I have no real way of verifying that and I don't even think that it matters. The most evident signal that I feel I can still get a pulse on is when something is clearly, heavily AI written. All other signals that would supersede it require additional time to evaluate. Things like reading the docs, skimming the codebase, testing the product etc.
With a project like this, the ask made to the end user is to spend the time to, at a minimum, read the docs to see if it's a good fit. If they think it's a good fit, the next ask is to spend the time to integrate it with their current inventory. Then learn the workflow. Then learn the idiosyncrasies. All of these things involve, real human, time.
When an end user is asked to invest this time, there are only so many signals available to them to help decide if it's a good use of their time, energy and resources. For better or for worse, it seems like that list of signals is shrinking and right now, something that appears vibecoded is still one of the clearest visible signals.