If you have high standards, its regurgitating info in an ill disciplined fashion. Because its input isn't really of the highest standard upon which the model is trained on.
It touches surface level stuff - which makes sense, most stuff on the internet is surface level. Good enough to pass an exam (since exams are essentially memorisation and regurigation of that nowadays) but not enough depth of understanding to be able to apply it in a wide range of contexts.
The application is where all the value is in the real world.
Generating profits is about the most tasteless thing one could do, yet it underpins all of our professional efforts.
The paradox is baked in, and some of us do our best to navigate it.
Absolutely not. Profit simply means other people find it valuable enough to compensate you to use whatever you’ve made.
Art is rarely profitable for its own sake, but that doesn’t mean everything that is profitable is intrinsically devoid of taste.
Not a single user finds advertising valuable, and yet it’s the focal point of profit maximization nowadays. Welcome to late-stage capitalism.
By looking them up when they need them
It's tracking, micro targeting, retargeting, and trying to sell me a fridge that I literally just bought while I'm off reading about sailboats that's intrusive.
Advertise shoes, cleats, sails, and charters in the Bahamas while I'm doing that, not singles near me and bicycles because I posted in a Facebook group.
Presumably the advertisers do.
There were times were advertising was useful and desirable, e.g. Small Ads pages.
There was also a time when ads were a single unintrusive scrolling line, curated by the website owner so as to be relevant to their audience. Those were fine.
I highly recommend The Century of the Self for a great documentary on the subject. https://en.wikipedia.org/wiki/The_Century_of_the_Self
That phrase has always seemed a bit wishful to me, like when Christians describe our era as the end times or when crypto people say "it's still early days".
I don't think that's particularly controversial. Profitability doesn't imply tastelessness, but profit motive certainly does.
Great logic.
This needs some justification. Profits are what you get when you can do something for less than you charge for it, and be competitive. To not be good enough to make profit you need to be able to force money out of people e.g. with taxes.
Should an entity strive to be profitable? of course. How else will it be self sustaining?
The problem arises when entities maximize for profit with no non-financial values that underpin their decision making.
Why's that? Profit, of course, is just the measure of how much trade is undelivered.
The old: I give you my corn to feed your chickens, and at some point in the future you will give me chickens in return once they are fed and grown. The amount of undelivered chickens are my profit. But eventually you will provide the chickens as promised, theoretically. Fair trade doesn't seem tasteless.
And if I forever hold on to that profit and never expect you to give me the chickens in the future as you originally promised, then I literally gave you the corn for free. How could it be tasteless to help someone out by giving them something for free?
Perhaps you are actually thinking of something like regulatory capture that is oft associated with profit? I can see how that becomes quite tasteless and certainly the tech industry in particular loves to exploit that. I am not sure that underpins our professional efforts, though. The tech industry would still exist even without all the insane laws that surround it.
However I would wager the argument many people have is that they view their professional life as not utilizing their taste and that is reserved for decisions in their private lives (for those who still keep the two relatively distinct). For those who have truly merged professional and personal and gone all in profit -- original point probably stands. Thats pretty tasteless!
It’s very easy for us to tell people to just not do that stuff, but I can tell you from my podcast production days that these annoying trends are often the difference between a 20% rise and a 20% drop in audience. No different than a clever book title turning heads.
To be clear, I find most of these trends incredibly obnoxious and I hated indulging them.
Yes, and if even these people can tell that AI generated stuff is godawful and tasteless, that tells you everything you need to know about AI.
And I think that is not all that surprising, because much of what it was trained on was corporate-speak, which has the same problem.
We can also talk about taste in articles, which seems to have degenerated to "any pro-AI article will be voted up and defended".
edit: As I am thinking about it more, it may be function of age. I am picking up some additional hobbies now and my whole approach has become much more intentional in general.
I don't want to delve into specifics, because it is a public forum. But the difference between learning google syntax and llm handling ( which I suppose would include prompt engineering ) should not be understated.
Can someone point out where this influx is happening?
The author doesn’t provide any references to this trend so I’m a bit confused why this is a big issue, as it’s literally the first time I’ve ever heard of it
In a way, AI does not change at all the problem of having taste. There are more books you'll ever read, movies you'll ever watch, games you'll ever play, software you'll ever use. I remain completely unconvinced that "dead internet/dead youtube" is a problem: you had to filter before, you have to filter now.
What AI does, being highly weird technology, is that it destroys heuristics. Good English used to be one. It used to take effort to write coherent sentences, that's now gone. Code even just compiling used to be evidence that someone at least made the effort to satisfy the type checker. That's gone as well.
I do see an argument that taste, a critical attitude and a good "bullshit detector" are now more important than ever.
Yes. Oddly, for once, English majors may actually benefit, because they may be better prepared than most to prepare prompts for the jobs of tomorrow ( mild sarcasm, coffee didn't kick in yet ).
With all due respect with pure technologists, they just dont understand people, what they need, and how to envision/communicate the benefits.
There still is a cottage industry of people saying one should write this way and that, and by large they have converged to a common consensus of what's Good English. It has been a successful enterprise, and now LLMs excel at generating text inside those parameters :-) .
Now, whenever I review a book, and if it applies, I make a point of saying "the grammar and sentence structure are squeaky clean". Often, that's about the only good thing I can say of the book.
I wonder if Good English is correlated with follow-the-norms attitude in an author+editors team. Because, once you make follow-the-norms your god, it is guaranteed that the writing will be formulaic and uninteresting. And then the only thing that can save your writing (financially) is good marketing.
Tasteless content [manifests] as the following:
— Copying and pasting code without understanding it.
— Designing websites that look exactly like every other company’s website.
— Regurgitating content from the trending influencer of the week.
Where’s the taste here? Where’s the critical judgment, discernment, or appreciation of aesthetic quality that separates mediocrity from excellence?
Not only possible, but exactly what AI does. :)
So I ask them what quality means. So far, I only get the most basic feedback: it should be in X style, pass Y linter, have N% coverage, have documentation...
At the same time, most, if not all manually written repositories do not pass the newfound quality metrics that must apply to AI code to be quality. I'm glad people are thinking about it at least, but let's not pretend like we cared before when it took manual labour. I'm even more glad we are in an age where quality standards can be fully automated.
Those are things that AIs can check by themselves.
What AIs are lacking is common sense.
They have no problem to inline everything they do which makes the codebase unmaintainable for humans
If you tell them to refactor, you get useless abstractions, like functions that get called in random places with no sense of structure.
That's because I trust the code that I write to have quality, because I wrote it, which means I understand it. I may choose not to test something because I am certain that nothing can go wrong with it.
When your repository is thousands of lines of code written by an AI with tendencies to forget critical components, duplicate work, make bonkers editing errors that shuffles everything around to all the wrong places, and invent packages out of thin air, you need a system of accountability.
Having poor taste (or more charitably, having no taste) can be covered up or ignored by the ability to choose from a pre-curated tasteful menu of options. This is what happens when people who "hate shopping" pick a mainstream clothing brand and stick with it. Or pick a car (most of them). Or a frying pan. I've never seen an offensively ugly frying pan. You could pick one out blindfolded and end up OK 100% of the time.
But when you put a tool like generative AI into this person's hands, they are exposed. The palette of possibilities is open. The curation is on you. And if someone with taste isn't in the mix, it will ultimately become apparent when you share your creation with the world.
Maybe let someone who can see colours pick something usable? You don't have to drag everyone to your level in the name of accessibility.
My TV has its backlight set to '0' (OLED, and a non-real property to set as no 'backlight', but still a metric). If I set it to 100 my eyes bleed. My current monitor with a real backlight has it set to 5. Yes, 5 out of 100.
I think grey on grey works, if the very walls behind you are being bleached by the intensity of the light coming off the monitor.
If you look back through the past, you can see some horrid design choices. Thus, some designs we think as awesome right now, will be seen as horrid to our descendants.
So if that's true, what if taste is social? And if it's social, then... well, all people have is peer pressure taste.
And your words show the truth in this, to a degree. Pre-curated options, to ensure "good taste" in choice. And how style conveys social status in some capacity, I don't mean "this style means success" but "this style means you have good taste".
Hair styles can be described as "taste", just as a taste in clothing. Yet hair styles suddenly become "ugly" where a decade before they were "tasteful".
Even beauty changes. One century it's skin and bones, the next more corpulent. Sometimes it's muscular, other times slim.
It's all peer pressure, all social status.
On the other hand deliberately tasteless art is a thing, it’s a bit in the eye of the beholder.
It’s true that many musicians cater to people who don’t really like music, they want to hear a good story with a beat. And that’s fine.
To have taste is to have developed a point of view, it’s not a mystical gift, it’s something you can develop over time. And not everyone needs that.
if you had bad taste, your taste is just badder faster
i think this is why so far there hasnt been any real moment of innovation from AI
because its not doing anything new. same crap as before just faster
I mean another way to think about it is - eating artificially produced food.
What happens when you consume it? Your taste for good stuff is eroded away - somehow, the artificial crap is acceptable. Most notably because of its lower price.
The same phenonomon with food, will prevail with information.
Why not just enjoy the things you enjoy? And if the things you enjoy drift over time as you experience more and notice more patterns then fine, but this does not mean the new enjoyable thing is in any objective sense better then the old.
Finally, I'm completely fine with a website that looks "exactly like every other website".
Its not as subjective.
Think Bootstrap or Material UI landing sites. Using a variant of those designs - big hero images, colourful buttons - makes your website look like it fits into the world. But use the default themes and it just looks lazy.
So there is always a tension in being attractive, you have to look unique, but still part of the species.
- search engine algorithms used be be the main place of information discovery. Before 200x it would involve not using javascript for any text you wanted to be readable by a bot
- "best viewed in x browser" which happened in the late 90s and early 00s. If a website looked crap, use the other browser.
- social graph metadata. Have a better image, title, description for people who see a snippet of your page on a social network
Nowadays everything is best viewed in Chrome/Safari, Firefox does have some issues.
Google owns the majority of the search market.
Facebook/Twitter/Linkedin at least in the Western world drive most social traffic.
I would guess the 'taste' of AI has been predetermined by these strong factors on the web.
An alternative could be a DMOZ like directory with cohorts voting on the value of things, maybe with the help of AI. It does seem like the web has been 'shaped' for the past 15 years or so.
People have trouble thinking 2 years out, let alone 5, 10, 15, 20 years...
To me it's undeniable that the web has become more centralised, more homogenised, and certain agents find that very convenient.
even wiki(pedia|data) is very convenient for large scale training, and most of their sources are from the 'open' web.
But with vibe coding, there is no knowledge transfer.
This is fundamentally how humans work. Our civilization is far too large and complex for you to be a bona fide expert in everything. And even worse, the more time you spend dedicated to expertise in one thing, the less well rounded you are in the ~infinity other things.
So what do you do if you cannot have authentic "taste" in everything?
You trust people. You find "influencers" that seem to succeed and have knowledge and you trust their opinion.
In practice, the vast majority of everything we think and do are not because of our personal expertise, but because of our un-expert belief in other people who do these things. Taste isn't just a measure of your personal expertise, but a measure of your ability to find and listen to experts.
So to call this out as a negative is harmful. You need to assemble a group of influencers not just professionally, but in every part of your life, who themselves are experts and can lend you their taste so you can make decisions without the paralysis of needing personal expertise in all things. That, or you need infinite time to become an infinite expert.
Perhaps I'm in a bubble (or outside of a bubble), but I don't know what the opening premise of this article is talking about. Namely:
> There’s been an influx of people telling others to develop taste to use AI. Designers. Marketers. Developers. All of them touting the same message.
I have not seen this influx. I have never heard anyone telling others to "develop taste to use AI". I work with & talk to a lot of developers, designers, marketers across a span of areas so this seems surprising to me. I've talked to outright skeptics, blind fanatics & some in between - the discussions with those "in betweeners" tend to centre more around functional contributions of AI as an aid rather than it's "creative output" & our judgements of same. I have yet to encounter anyone who has strong thoughts on taste w.r.t. AI beyond the polar extremes of "AI can do my creativity for me" & "only use AI for explicitly functional non-taste-related tasks".
Am I the only one missing this influx?
You could say that propaganda authors had more taste back in my day.
fluxusars•1h ago
jmeaden•50m ago