I'm a little dense, so at first I took this literally, thought the blue text were actually links, tried to click them so I could figure out wtf is going on and figure out where the author lived...
I'm a little dense, so at first I took this literally, thought the blue text were actually links, tried to click them so I could figure out wtf is going on and figure out where the author lived that blue food was a trend.
Then I realized it was allegory, but still didn't really understand what it was an allegory for, but figured it still had to do with food, so maybe high fructose corn syrup? Salt? Preservatives?
Then I finally realized it was AI, despite maybe 2 minutes ago writing a passionate reply about a different AI story.
It is ASTOUNDING to me how often I am seeing people using ChatGPT as their reference in discussions online, without any resistance at all. A similar thing happened when everybody apparently...
It is ASTOUNDING to me how often I am seeing people using ChatGPT as their reference in discussions online, without any resistance at all. A similar thing happened when everybody apparently started accepting wikipedia as an authoritative source (I don't think I've seen anyone object to a wikipedia link in the past 8 years or so) but the adoption of AI is like 100 times faster.
I try to avoid falling into the trap of "back in my day," but the more I see, the more it really seems like we're boiling the frog into a state of idiocracy that people aren't really too concerned about.
I didn't know what this was about so I asked ChatGPT Oh, now I get it!
I didn't know what this was about so I asked ChatGPT
The blue-food fable is meant as an allegory for artificial-intelligence hype and creep.
The blue coloring behaves exactly the way AI is said to behave today: it’s marketed as harmless and “nearly undetectable,” then it becomes ubiquitous, prices rise once everyone is locked-in, firms re-orient themselves around it (“blue-first company”), and dissenters are told to “just get used to it.” Those beats mirror how AI features are pushed into every product while users’ ability to opt out erodes.
The post carries the tag ai in its topic list, alongside short and story. A follow-up micro-essay on the site (“It’s about lots of things”) clarifies that the story is “even a little bit about AI,” grouping it with other tech annoyances like non-linear feeds, NFTs and “The Algorithm.” 
In short, the blue-food scenario is a tongue-in-cheek stand-in for the way AI is being poured into everything, whether we asked for it or not.
That’s a dangerous game the author is playing. I think the “coming any day now” blue AGI will see right through the author's thoughtcrime and punish it accordingly once it has taken over.
That’s a dangerous game the author is playing. I think the “coming any day now” blue AGI will see right through the author's thoughtcrime and punish it accordingly once it has taken over.
Me: "No, you see, it's fine if it makes sense for it to be blue, like blue raspberry or blueberry" My general attitude is AI isn't inherently bad, but we keep dying everything blue which is...
Me: "No, you see, it's fine if it makes sense for it to be blue, like blue raspberry or blueberry"
My general attitude is AI isn't inherently bad, but we keep dying everything blue which is definitely a huge problem.
Is it a problem? That’s a common occurrence when a new technology arises. There’s a rush to apply to all kinds of problems, because without trying, how would you know if it works or not?...
Is it a problem? That’s a common occurrence when a new technology arises. There’s a rush to apply to all kinds of problems, because without trying, how would you know if it works or not? Eventually, the winners remain and the lost causes fade to history.
Sometimes it is. You see, not all new technologies have the same impact, marketing or capital behind them. In some cases there is so much investment capital behind it that it isn't so much the...
Is it a problem?
Sometimes it is. You see, not all new technologies have the same impact, marketing or capital behind them. In some cases there is so much investment capital behind it that it isn't so much the technology itself that is the main driving factor but the money behind it.
And sometimes when money is the driving factor you end up with situations where the implementations that do "win" are worse than the things they are forcefully replacing. Either because there is so much money invested the powers that be want to see some return on it. Or, in other situations in the past, where marketing budget won the battle for an overal inferior product compared to the competition.
As far as I am concerned, your statement makes sense in a market where using "blue dye" is still a choice amongst other things.
But is that a problem? It's OK for people to try things out. It's weird to make it out like it's not, because that requires prescience that doesn't exist in the world. Over the long term, things...
But is that a problem? It's OK for people to try things out. It's weird to make it out like it's not, because that requires prescience that doesn't exist in the world. Over the long term, things will sort themselves out.
For "AI" this is true in particular. It's quite expensive. Compared to, say, a database query, multiplying several gigabyte matrices together is a lot. The reason not to use "AI" is quite evident on your bottom line.
The reason everyone is jumping on the bandwagon is that there is a graveyard of companies who didn't catch on to trends and are now dead because of it. So it's a hedge. You don't want to be Nokia, or Blackberry, who were on top of the world, only be to fall in a matter of years. If AI doesn't work out - you wasted some money, oh well. If AI does work out - you may have saved the entire business from irrelevancy.
If it really, truly is pointless for XYZ purposes, it will be phased out. There's a continual and very present cost to incentivize you not to use "AI" for anything that doesn't need it.
It can be, for consumers, for the reasons I outlined in the comment you are replying to. Things can have a negative impact now. The notion that this potentially will solve itself in the future...
But is that a problem?
It can be, for consumers, for the reasons I outlined in the comment you are replying to.
Things can have a negative impact now. The notion that this potentially will solve itself in the future doesn't change that. Not to mention that sometimes with some disruptors the aftermath is not a return to the status quo for a long time. Simply because infrastructure that was there before has gone the way of the dodo so there isn't necessarily something to return to.
Mind you, there are as many examples of this through out history as there are for your scenario. Personally I rather err on the side of caution here, but hey, that's just me.
What infrastructure are people going to forget about? This is a process that going's to take like 5 years or so, not a millenia, and the "AI" stuff is mostly just stapled on top of existing...
What infrastructure are people going to forget about? This is a process that going's to take like 5 years or so, not a millenia, and the "AI" stuff is mostly just stapled on top of existing products.
It's not that it's a problem that will solve itself, per se, it's that a) we don't know if this is a dead end or a real advancement until time has passed and b) once that time has passed, there's strong incentives for only usable applications to survive.
It's a natural, and good process. Because you only know the quality of something when you do it. It can get wonky when you're in the middle, but it is what it is. Many fads have passed, and many fads have turned into necessary infrastructure.
Personally I rather err on the side of caution here, but hey, that's just me.
I mean, any given company is free to ignore it, and many are. They have that choice, as do the ones that put out a feeler.
You are glossing over quite a few points I made, while focussing on very narrow parts of others. For starters, you are primarily framing this from the business perspective of hedging bets and...
Exemplary
You are glossing over quite a few points I made, while focussing on very narrow parts of others.
For starters, you are primarily framing this from the business perspective of hedging bets and long-term cost savings.
Which completely skips over things like: lost options for users, lower quality and skills or infrastructure that disappear once the "blue dye" goes in. Which is what I am talking about.
When teams glue on LLMs on features today, they rarely leave the old paths in place. Local search that used to run on device now pings an API and costs extra (either in price or privacy). Photo sites/apps bury manual slider behind magic buttons. Phone support lines vanish, replaced by chatbots that loop forever. Once that choice is gone, the engineers who kept it working move on, making any later rollback expensive at the very least.
By talking about infrastructure in this context I am not talking about buildings or servers. I am referring to skills, workflows and alternative products. Something I think I was fairly clear about to be honest. You can't just staple back dismantled functionality five years later. Institutional knowledge in companies degrades quickly, in fact it even can degrade surprisingly quickly in entire industries. Possibly even worsened by the fact that fewer engineers are trained in that period on that knowledge, even fewer engineers in general as many companies put a freeze on junior hiring. And those that do get hired don't gain the same knowledge and if in 5 years time they are asked to dial back AI systems they will have to gain a lot of new knowledge. If that even happens.
The notion that compute costs will police this trend also misses how long cheap capital and FOMO can keep a bad idea alive. WeWork and a stack of crypto outfits burned billions for years before reality caught up. By then plenty of viable alternatives were already dead and users had nowhere else to go. These are just recent examples, as I said, history is littered by examples of this where capital is the driving factor, not the actual value of products.
Experimentation is absolutely fine and personally I also don't have a fundamental issue with LLMs. I am quite happy using various models for various tasks and don't mind products making them optional or non intrusive.
But, what I (and the article for that matter) am pushing back on is rolling half baked AI into every product by default. Leaving consumers to eat the risk while companies chase a hedge. That isn't caution, it is just basic due diligence. If we want "only usuble applications to survive" we also need real opt-outs, real comparisons and time to prove the tech. Not a oneway door that closes the moment the marketing budget arrives.
As a closing note, clearly you love technology. Which is cool, I like technology too. But it is also fine to love technology, be excited about the future of it while being critical of where and how it is being used. I'd argue it is even essential, otherwise your entire premise of the market solving itself out is a non starter to begin with.
I agree with this. From a purely consumer perspective, the fact that all the new features come without any off switch - I can't get rid of AI chat in Snapchat or FB Messenger, disable AI summaries...
I agree with this. From a purely consumer perspective, the fact that all the new features come without any off switch - I can't get rid of AI chat in Snapchat or FB Messenger, disable AI summaries with a switch on Google, remove Android 16's AI search icon right where my right thumb hits the default unremovable search bar for web/phone search - makes it feel like computing costs won't be a controlling factor, at least not for years, and is actively making my day to day experience worse. It's not like the blue food is on the shelves, the regular food is gone and only the blue food is left.
It's so easy to accidentally hit the "blue food please" buttons by accident, I'm sure it's easy to say that everyone wants their food blue now.
ETA someone just compared accidentally hitting the AI search button to hitting the "internet" button on your flip phone and trying to back out before you (or your parents) got charged for data and that feels very accurate. Perhaps mixed with accidentally hitting Clippy because no I don't need help.
I would argue there’s a much larger graveyard of companies who followed the hype train off the edge of a cliff. There are far more failed companies who never really got started than there are...
The reason everyone is jumping on the bandwagon is that there is a graveyard of companies who didn't catch on to trends and are now dead because of it. So it's a hedge. You don't want to be Nokia, or Blackberry, who were on top of the world, only be to fall in a matter of years. If AI doesn't work out - you wasted some money, oh well. If AI does work out - you may have saved the entire business from irrelevancy.
I would argue there’s a much larger graveyard of companies who followed the hype train off the edge of a cliff. There are far more failed companies who never really got started than there are companies who did quite well for a while before then declining later, but I feel like your examples are more worried about the latter examples than the former.
I meant my use of "problem" to be more colloquial than dramatic, but the color analogy is pretty on point in that many additions will be weird like blue-tinged chicken breast, not strictly...
I meant my use of "problem" to be more colloquial than dramatic, but the color analogy is pretty on point in that many additions will be weird like blue-tinged chicken breast, not strictly beneficial in many contexts, and just kinda give things that weird tinge: The de-facto standard web browsers, OSes and phone OSes currently all have some degree of AI integration. Every major, non-subscription/purpose-built search engine has some sort of generalized chat/search AI. Microsoft has it in every facet of their software. I think it's understandable to not want harmlessly blue cheeseburgers in much the same way I think it's okay to not want an AI sidepane on my web browser.
Your comment reminded me of two early bicycle adaptations when bicycles were still new-ish: sail bikes land-and-water velocipedes Might seem silly and impractical in hindsight, but man would I...
Your comment reminded me of two early bicycle adaptations when bicycles were still new-ish:
Squeezing out berries, you usually get either red or purple, so maybe the point really is that almost nothing should be “naturally” blue? :P Edit: Few things edible, great point.
Squeezing out berries, you usually get either red or purple, so maybe the point really is that almost nothing should be “naturally” blue? :P
Love it. It reminded me of an older Steve Hughes bit about television from his "While it's still legal" special. Australian levels of profanity, including 1 drop of the r word, but the core...
Love it.
It reminded me of an older Steve Hughes bit about television from his "While it's still legal" special. Australian levels of profanity, including 1 drop of the r word, but the core message is sound.
Punchline that applies to AI too
Stop being so cynical, so negative, and eat your shit.
The important question is: does this person who lives in such a blue world drive a blue Corvette and have a girlfriend who is also blue? Korg M1 piano starts
The important question is: does this person who lives in such a blue world drive a blue Corvette and have a girlfriend who is also blue?
Yeah I think I actually find that a questionable follow up, sure I guess it's about other things but, only one of those was a tag on the original post? Idk I get the broader point being made but I...
Yeah I think I actually find that a questionable follow up, sure I guess it's about other things but, only one of those was a tag on the original post?
Idk I get the broader point being made but I think it actually weakens the very strong, very focused (IMO) original post.
I have to disagree. AI is far from the first "new normal" that's been imposed on us in our lifetimes. This story of blue has played out so many times in so many ways, that we've all been the blue...
I have to disagree. AI is far from the first "new normal" that's been imposed on us in our lifetimes. This story of blue has played out so many times in so many ways, that we've all been the blue hater or part of the accepting public at various times. The most powerful stories are the ones that echo through time, and limiting the parable to just AI really just limits its potential.
As I said I get the broader applicability. But I find it relevant that the original blogpost didn't tag any of those other things. It feels disingenuous to throw AI in the middle of a long list of...
As I said I get the broader applicability. But I find it relevant that the original blogpost didn't tag any of those other things.
It feels disingenuous to throw AI in the middle of a long list of things it's "about." Had the follow up post been like, yeah it's about AI, because .... But also it's about ...... (Long list), I'd respect it more.
The topics only say "short" "story" and "AI" for a reason.
And they posted that follow up for a reason. What can I say, I'm a firm believer of the Death of the Author, and I can respect someone who will pull it on themselves.
And they posted that follow up for a reason.
What can I say, I'm a firm believer of the Death of the Author, and I can respect someone who will pull it on themselves.
Well I think arguably as they're the author, their POV is not Death of the Author. As I said I think it's a questionable follow up. But like they obviously can write as they like. I just don't like it
Well I think arguably as they're the author, their POV is not Death of the Author.
As I said I think it's a questionable follow up. But like they obviously can write as they like. I just don't like it
Yeah sharing this to my coworkers equally frustrated at being told we're not leveraging AI sufficiently
I'm a little dense, so at first I took this literally, thought the blue text were actually links, tried to click them so I could figure out wtf is going on and figure out where the author lived that blue food was a trend.
Then I realized it was allegory, but still didn't really understand what it was an allegory for, but figured it still had to do with food, so maybe high fructose corn syrup? Salt? Preservatives?
Then I finally realized it was AI, despite maybe 2 minutes ago writing a passionate reply about a different AI story.
I think I might be stupid.
You know what I hear is good for that? That we all need more of? AI, apparently.
I saw a really funny tweet yesterday which was:
It is ASTOUNDING to me how often I am seeing people using ChatGPT as their reference in discussions online, without any resistance at all. A similar thing happened when everybody apparently started accepting wikipedia as an authoritative source (I don't think I've seen anyone object to a wikipedia link in the past 8 years or so) but the adoption of AI is like 100 times faster.
I try to avoid falling into the trap of "back in my day," but the more I see, the more it really seems like we're boiling the frog into a state of idiocracy that people aren't really too concerned about.
You're not alone. Going purely off of the title, I thought it was about the FDA licencing a new 'natural' blue food dye.
I didn't know what this was about so I asked ChatGPT
Oh, now I get it!
"...whether we asked for it or not."
Chat? Is that really the right pronoun there?
With how much ChatGPT has impacted people’s decisionmaking, I feel like Chatty can use “we” lmao.
That’s a dangerous game the author is playing. I think the “coming any day now” blue AGI will see right through the author's thoughtcrime and punish it accordingly once it has taken over.
Me: "No, you see, it's fine if it makes sense for it to be blue, like blue raspberry or blueberry"
My general attitude is AI isn't inherently bad, but we keep dying everything blue which is definitely a huge problem.
Is it a problem? That’s a common occurrence when a new technology arises. There’s a rush to apply to all kinds of problems, because without trying, how would you know if it works or not? Eventually, the winners remain and the lost causes fade to history.
Sometimes it is. You see, not all new technologies have the same impact, marketing or capital behind them. In some cases there is so much investment capital behind it that it isn't so much the technology itself that is the main driving factor but the money behind it.
And sometimes when money is the driving factor you end up with situations where the implementations that do "win" are worse than the things they are forcefully replacing. Either because there is so much money invested the powers that be want to see some return on it. Or, in other situations in the past, where marketing budget won the battle for an overal inferior product compared to the competition.
As far as I am concerned, your statement makes sense in a market where using "blue dye" is still a choice amongst other things.
But is that a problem? It's OK for people to try things out. It's weird to make it out like it's not, because that requires prescience that doesn't exist in the world. Over the long term, things will sort themselves out.
For "AI" this is true in particular. It's quite expensive. Compared to, say, a database query, multiplying several gigabyte matrices together is a lot. The reason not to use "AI" is quite evident on your bottom line.
The reason everyone is jumping on the bandwagon is that there is a graveyard of companies who didn't catch on to trends and are now dead because of it. So it's a hedge. You don't want to be Nokia, or Blackberry, who were on top of the world, only be to fall in a matter of years. If AI doesn't work out - you wasted some money, oh well. If AI does work out - you may have saved the entire business from irrelevancy.
If it really, truly is pointless for XYZ purposes, it will be phased out. There's a continual and very present cost to incentivize you not to use "AI" for anything that doesn't need it.
It can be, for consumers, for the reasons I outlined in the comment you are replying to.
Things can have a negative impact now. The notion that this potentially will solve itself in the future doesn't change that. Not to mention that sometimes with some disruptors the aftermath is not a return to the status quo for a long time. Simply because infrastructure that was there before has gone the way of the dodo so there isn't necessarily something to return to.
Mind you, there are as many examples of this through out history as there are for your scenario. Personally I rather err on the side of caution here, but hey, that's just me.
What infrastructure are people going to forget about? This is a process that going's to take like 5 years or so, not a millenia, and the "AI" stuff is mostly just stapled on top of existing products.
It's not that it's a problem that will solve itself, per se, it's that a) we don't know if this is a dead end or a real advancement until time has passed and b) once that time has passed, there's strong incentives for only usable applications to survive.
It's a natural, and good process. Because you only know the quality of something when you do it. It can get wonky when you're in the middle, but it is what it is. Many fads have passed, and many fads have turned into necessary infrastructure.
I mean, any given company is free to ignore it, and many are. They have that choice, as do the ones that put out a feeler.
You are glossing over quite a few points I made, while focussing on very narrow parts of others.
For starters, you are primarily framing this from the business perspective of hedging bets and long-term cost savings.
Which completely skips over things like: lost options for users, lower quality and skills or infrastructure that disappear once the "blue dye" goes in. Which is what I am talking about.
When teams glue on LLMs on features today, they rarely leave the old paths in place. Local search that used to run on device now pings an API and costs extra (either in price or privacy). Photo sites/apps bury manual slider behind magic buttons. Phone support lines vanish, replaced by chatbots that loop forever. Once that choice is gone, the engineers who kept it working move on, making any later rollback expensive at the very least.
By talking about infrastructure in this context I am not talking about buildings or servers. I am referring to skills, workflows and alternative products. Something I think I was fairly clear about to be honest. You can't just staple back dismantled functionality five years later. Institutional knowledge in companies degrades quickly, in fact it even can degrade surprisingly quickly in entire industries. Possibly even worsened by the fact that fewer engineers are trained in that period on that knowledge, even fewer engineers in general as many companies put a freeze on junior hiring. And those that do get hired don't gain the same knowledge and if in 5 years time they are asked to dial back AI systems they will have to gain a lot of new knowledge. If that even happens.
The notion that compute costs will police this trend also misses how long cheap capital and FOMO can keep a bad idea alive. WeWork and a stack of crypto outfits burned billions for years before reality caught up. By then plenty of viable alternatives were already dead and users had nowhere else to go. These are just recent examples, as I said, history is littered by examples of this where capital is the driving factor, not the actual value of products.
Experimentation is absolutely fine and personally I also don't have a fundamental issue with LLMs. I am quite happy using various models for various tasks and don't mind products making them optional or non intrusive.
But, what I (and the article for that matter) am pushing back on is rolling half baked AI into every product by default. Leaving consumers to eat the risk while companies chase a hedge. That isn't caution, it is just basic due diligence. If we want "only usuble applications to survive" we also need real opt-outs, real comparisons and time to prove the tech. Not a oneway door that closes the moment the marketing budget arrives.
As a closing note, clearly you love technology. Which is cool, I like technology too. But it is also fine to love technology, be excited about the future of it while being critical of where and how it is being used. I'd argue it is even essential, otherwise your entire premise of the market solving itself out is a non starter to begin with.
I agree with this. From a purely consumer perspective, the fact that all the new features come without any off switch - I can't get rid of AI chat in Snapchat or FB Messenger, disable AI summaries with a switch on Google, remove Android 16's AI search icon right where my right thumb hits the default unremovable search bar for web/phone search - makes it feel like computing costs won't be a controlling factor, at least not for years, and is actively making my day to day experience worse. It's not like the blue food is on the shelves, the regular food is gone and only the blue food is left.
It's so easy to accidentally hit the "blue food please" buttons by accident, I'm sure it's easy to say that everyone wants their food blue now.
ETA someone just compared accidentally hitting the AI search button to hitting the "internet" button on your flip phone and trying to back out before you (or your parents) got charged for data and that feels very accurate. Perhaps mixed with accidentally hitting Clippy because no I don't need help.
I would argue there’s a much larger graveyard of companies who followed the hype train off the edge of a cliff. There are far more failed companies who never really got started than there are companies who did quite well for a while before then declining later, but I feel like your examples are more worried about the latter examples than the former.
I meant my use of "problem" to be more colloquial than dramatic, but the color analogy is pretty on point in that many additions will be weird like blue-tinged chicken breast, not strictly beneficial in many contexts, and just kinda give things that weird tinge: The de-facto standard web browsers, OSes and phone OSes currently all have some degree of AI integration. Every major, non-subscription/purpose-built search engine has some sort of generalized chat/search AI. Microsoft has it in every facet of their software. I think it's understandable to not want harmlessly blue cheeseburgers in much the same way I think it's okay to not want an AI sidepane on my web browser.
Your comment reminded me of two early bicycle adaptations when bicycles were still new-ish:
sail bikes
land-and-water velocipedes
Might seem silly and impractical in hindsight, but man would I love to try them out.
off topic, but the "blue raspberry" flavor is largely made up of cherry and pineapple-based flavors.
Squeezing out berries, you usually get either red or purple, so maybe the point really is that almost nothing should be “naturally” blue? :P
Edit: Few things edible, great point.
There are a few animals I can think of with natural blue colouring, and many are extremely poisonous.
I think you're omitting birds from your blue, non poisonous/venomous model. And fish. And butterflies.
(The sea slugs and octopi are right out though)
Love it.
It reminded me of an older Steve Hughes bit about television from his "While it's still legal" special. Australian levels of profanity, including 1 drop of the r word, but the core message is sound.
Punchline that applies to AI too
Stop being so cynical, so negative, and eat your shit.
The important question is: does this person who lives in such a blue world drive a blue Corvette and have a girlfriend who is also blue?
Korg M1 piano starts
I don't think I'd ever actually listened to the song properly before today to know there was anything to it besides the chorus lol.
The author has posted a little follow up: https://blog.foxtrotluna.social/its-about-lots-of-things/
Yeah I think I actually find that a questionable follow up, sure I guess it's about other things but, only one of those was a tag on the original post?
Idk I get the broader point being made but I think it actually weakens the very strong, very focused (IMO) original post.
I have to disagree. AI is far from the first "new normal" that's been imposed on us in our lifetimes. This story of blue has played out so many times in so many ways, that we've all been the blue hater or part of the accepting public at various times. The most powerful stories are the ones that echo through time, and limiting the parable to just AI really just limits its potential.
As I said I get the broader applicability. But I find it relevant that the original blogpost didn't tag any of those other things.
It feels disingenuous to throw AI in the middle of a long list of things it's "about." Had the follow up post been like, yeah it's about AI, because .... But also it's about ...... (Long list), I'd respect it more.
The topics only say "short" "story" and "AI" for a reason.
And they posted that follow up for a reason.
What can I say, I'm a firm believer of the Death of the Author, and I can respect someone who will pull it on themselves.
Well I think arguably as they're the author, their POV is not Death of the Author.
As I said I think it's a questionable follow up. But like they obviously can write as they like. I just don't like it
I wish I hadn't seen the tag before I read the post, because knowing it involved AI ahead of time made it obvious.
FWIW it felt very obvious to me and I didn't see the tags. But not everything hits the same for everyone.