I could be misreading this, but looking at the numbers, it seems that--despite the claim "we didn't estimate this"--basically everything in this is an estimate. It's a little hard to tell though, because the generated prose seems to occasionally contradict or repeat itself.
it's really hard to measure some times. for example, once of the things we have seen is that the specific `workflow-triage` schedule we have have been running much faster and cheaper since we moved some parts of it to use scripts. From our analytics we saw times from ~5min before to ~1 or 2 min now. And in terms of costs we saw +50% reduction.
Why not 90%? because even if it uses the script, that's the "repeatable" part. we still have a reasoning around the output of it, and then deciding how to escalate, which removes the 40% reduction. Still 50% is nice for something we run daily.
There's definitely a group of people who trumpet continually that AI is going to replace everything soon and anyone who doesn't use it maximally will be left behind.
This is a genuine comment, rather than snark, though I appreciate it might initially look that way.
Your comment is an example of something that seems to accrue around AI users: casual, even gleeful, spite. You don't get this with every technology; there is no significant strand of the pro-3D printing community, for example, that goes around taunting people who don't use printers themselves.
It's significant, I think, that this sort of attitude is common around AI, but not other technologies, and it does have a meaningful impact on public attitudes around it. I'm not sure whether it's a result of using AI, or a characteristic that leads to a propensity for AI, but there's something here.
I think GP is aware that they're waging a culture war, on the side that stands for brutal pragmatism and embracing the race to the bottom, and they know that casual displays of cynicism are the simplest way to prop up their side.
The gleeful spite and disdain for building community norms isn't an accident, it's an attitude they want to normalize.
If you're really passionate about serving the best quality coffee, you'll be passionate about the experience around that, and you'll put effort in even if it doesn't pay off.
AI-generated art signals that you didn't put effort in, and so aren't passionate about the experience, and so aren't really passionate about the central thing.
"At least you tried" is not a high standard, but it's the one that AI art fails.
I think generally the opposite is true. A place with bad coffee but great decor can charge the same as a place with good coffee but crappy everything else. And so to maximize profitability, it does just that, meaning that at any price level, the place with the worst non-coffee has the best coffee. If you judge them by their posters, then what you’re paying for is the posters.
I think people complaining about AI art at businesses are making an aesthetic argument and then laundering it into ones with more heft. I mostly agree with the aesthetic argument, but it’s only that, not some wider indictment.
While I agree with the general sentiment (people not putting effort is a bad signal), I also think people are sometimes being too harsh. Everybody has limited time and resources and knowledge. Making something (a business, a book, a website, etc.) means making some trade-offs with the resources you got.
When I can easily (without lots of effort/resources) check the quality of something, I prefer not to be that influenced by this type of "bad signals". Think of buying a coffee or trying a sandwich place. When I make a big investment, like a laptop or a TV, yes then it is a different story.
I disagree, but it's also academic. Even if every single one of my assumptions is unfair, they're still being made by me and others. Restaurants need to consider the reputational risk caused by customer perception. There's a whole film about it [1].
I think there's a difference between passively and actively putting no thought into something; generating a menu with AI involves deciding that it's worth having but that you don't care about the quality.
>If you're really passionate about serving the best quality coffee, you'll be passionate about the experience around that
The degree to which the menu is part of "the experience around that" is exactly the debate here; you can't just claim it as axiomatic.
>AI-generated art signals that you didn't put effort in,
or it signals you spend your time sourcing the best beans, understanding the climate of the various bean producing regions, reading trade news for the coffee industry, ...
>"At least you tried" is not a high standard, but it's the one that AI art fails.
I've thought about the specific example of distorted menu photos before as well and to me it also signals intelligence level in the proprietor. I can't imagine myself in that position, looking at that and saying, "yup, that's good." And I also have a hard time imagining not understanding why it looks bad, what one could have done to not make it bad. It feels like a distinct decision to be lazy.
And like you said, if they are that careless and unadept at one thing, how are they going to work the intricacies of an espresso machine and food safety standards? Not that an espresso machine is that hard to operate, but surely it's more difficult than maintaining the aspect ratio of a photo. You don't even need to know the term "aspect ratio" to get it right.
But I also have to admit that my expectations seem to be very off on this specific issue. My wife is quite intelligent (actual, literal genius) and even quite artsy, and yet I had to teach her how to not resize photos incorrectly, both for aspect ratio and quality by resizing too large. It's seems like such an easy thing to see and figure out to me. This must be how my 10 year old sees my chess play.
I think my issue with this project--and so many other similar ones--is that the provenance of the code does undermine the intention. If a project purports to be about quality, then knowing that the creator abdicated some of the responsibility for creating the thing they ostensibly care about makes it harder to put faith in them as having high standards elsewhere.
Perhaps I am just old-fashioned, and vibe-coding is something that can be done with full focus and care for high quality, but I remain unconvinced. This is a project that needs to be cared about sincerely to be trustable/meaningful/useful, and the approach taken casts doubt on that.
> Programming, even systems engineering, is a surprisingly shallow field.
I agree, and I don't think we should be at all ashamed of this.
The beauty of programming is that we can produce incredibly complex and powerful things by manipulating a small set of simple constructs together. There are only a few core tools--iteration, conditionals, etc.--but they can be snapped together into much more capable configurations.
Good programming is the art of deconstructing complex behaviour into these few constructs, and that's really fascinating.
I suspect you are right, but it is getting really difficult to tell the difference between AI and someone who simply don't know what they are writing about.
The US defines its objectives on a large scale (e.g. regime change, or peace in the Middle East), but doesn't then use tactics that can achieve that.
What they're doing in Iran, for example, is probably best described as 'raiding'--quick, damaging strikes--but that doesn't lead to the final result they want. It hurts the attacked country, but it doesn't control it.
reply