People are already complaining about how the AI training data from recent forums are “contaminated” with outputs from other AIs, if you want something “purely human” to work from then historical pre-2023 data is the best bet.
In the final analysis, nobody cares what Harold Q. Dumpington bought from Amazon in the week of June 4, 2017. That information is technically still stored in Amazon’s databases, but (1) Amazon already has access to it, so encryption is a sort of non-issue, and (2) nobody cares.
The reality is: socially engineering a password or setting up a “man in the middle” attack in a coffee shop WiFi is a hell of a lot easier than attacking encrypted data, but even those attacks are relatively rare, and usually executed against corporations with money. As tempting as it would be for some hacker to get into Jennifer Lawrence’s e-mail or Chris Pratt’s Amazon purchase history, it seems that it’s really not worth the effort to anybody, except in some edge cases.
Putting aside the whole question of what people might want to feed into an AI, why would anybody want that data AT ALL?
You’re making claims about what everyone (and everything) to ever live after this point in time is going to care about. That’s unfounded and kind of presumptuous.
If an AI was being trained to “be” a specific person, why wouldn’t their history of Amazon purchases be useful as part of building up that persona? Or on a broader scale, wouldn’t patterns of purchases be useful for modelling cultural patterns?
Exactly. This is like the biggest problem with today’s society. Everyone thinks they are unique and sigificant… you are not.
The dirty erotic fan fic stories you wrote when you were 14 aren’t going to ruin your life if they are found. What will ruin your life is when you find out how little anyone gives a fuck about you.
Future people do not give a damn about your shopping, your Visa number SSL’d to Cherry-Popping Hot Grampa Action websites that you visit, nor password-protected partitions, no matter how illicit.
And this, it would seem, is your saving grace: the amazing haste of people to forget your name, your face, your litanous list of indefensible indiscretions.
They’ll glance you over, I guess, and then for a bare moment you’ll persist to exist; almost seems like you’re there, don’t it? But you’re not. You’re here. Your name will fade as Front’s will.
That’s a fair question, but I think the answer is obvious. Until the invention of photography, literally the only formal records we had of past events were the things people bothered to write down, paint, or sculpt. And of those, we only have the arts and written records that actually survived. So to find out information about the distant past, we have little choice but to extrapolate from artifacts, dig up old buildings, etc. The artifacts and records that we do find have outsized influence on our understanding of the past, compared to all the information and details that have been lost, which can literally never be recovered.
From the 21st century onward, that relationship is inverted. Any hypothetically useful unit of information about the past will be recorded hundreds or thousands of times, and the useless units of information will outnumber the useful units by many orders of magnitude. Sure, if someone proves to be exceptionally notable, there may be some value in decrypting their past Amazon purchases or cracking the encrypted SSD they left behind. But that’s going to be the exceedingly rare exception, rather than the rule, especially when the world’s data stores are crammed with news articles, photos, videos, interviews, blog posts, reddit posts, journals, and non-encrypted records that appear to tell a complete story of the lives of notable people, and for that matter the day-to-day lives of regular folk.
And that SSD may be every bit as exciting as the Hunter Biden laptop hard disk… that is, barely exciting at all, and full of such routine and irrelevant information as to be an almost pointless exercise in data forensics.
That’s fair. Adding to my point, with the wealth of information future people will have at their disposal, it could be possible to recreate this time era. That is, to simulate entire cities or countries. Who knows what tech they’ll have or what they’ll want to do with it. My point is that the info from this time period, between the advent of the internet and the widespread use of quantum-safe crypto, will be easily accessible to them, and contains such an accurate record of our daily activities. I’ve had the same email address since 2005 and have never deleted messages, so my email alone could probably be used to create a pretty accurate model of a large chunk of my life. Cross-reference that with the information the people I associate with left behind and they definitely could create such a model.
And, adding further, if you were inclined to create such a simulation, you’d likely want to simulate as many people as possible so that the simulation was as realistic as possible.
I can’t believe I forgot about this, but if you really want to explore the question of future people reconstructing the past through AI, watch the movie Marjorie Prime, which is explicitly about this question.
Yeah, almost nothing. SMS is a utility tool for me. I doubt anyone will ever care that my wife wanted more zip-lock bags.
You’d get a better picture of me through old USENET posts (which are unencrypted, of course), or reddit or web forums or Lemmy (all of them unencrypted, I suspect). Good luck, future people.
We are a period-genuine curiosity that will certainly be reconstructed in the future—if the data is available to do so. Our lives, logs, and transactions are a finite resource if simply because they are real.
Imagine a hobby, digging up old logs and piecing together various accounts across deep datsets, working toward a bigger picture… and then realize this is all intuited through whatever present AI. No labor involved, at all, and there is no time limit on this.
We’re all eventually just an intelligent query away from being rediscovered if outside of average in any respect, and even the most average person will become a celebrated oddity.
I think you vastly overestimate the future’s interest in your life.
People are already complaining about how the AI training data from recent forums are “contaminated” with outputs from other AIs, if you want something “purely human” to work from then historical pre-2023 data is the best bet.
In the final analysis, nobody cares what Harold Q. Dumpington bought from Amazon in the week of June 4, 2017. That information is technically still stored in Amazon’s databases, but (1) Amazon already has access to it, so encryption is a sort of non-issue, and (2) nobody cares.
The reality is: socially engineering a password or setting up a “man in the middle” attack in a coffee shop WiFi is a hell of a lot easier than attacking encrypted data, but even those attacks are relatively rare, and usually executed against corporations with money. As tempting as it would be for some hacker to get into Jennifer Lawrence’s e-mail or Chris Pratt’s Amazon purchase history, it seems that it’s really not worth the effort to anybody, except in some edge cases.
Putting aside the whole question of what people might want to feed into an AI, why would anybody want that data AT ALL?
MC Frontalot has a song about this, Secrets from the Future.
You’re making claims about what everyone (and everything) to ever live after this point in time is going to care about. That’s unfounded and kind of presumptuous.
If an AI was being trained to “be” a specific person, why wouldn’t their history of Amazon purchases be useful as part of building up that persona? Or on a broader scale, wouldn’t patterns of purchases be useful for modelling cultural patterns?
Exactly. This is like the biggest problem with today’s society. Everyone thinks they are unique and sigificant… you are not.
The dirty erotic fan fic stories you wrote when you were 14 aren’t going to ruin your life if they are found. What will ruin your life is when you find out how little anyone gives a fuck about you.
As MC Frontalot opines:
Future people do not give a damn about your shopping, your Visa number SSL’d to Cherry-Popping Hot Grampa Action websites that you visit, nor password-protected partitions, no matter how illicit.
And this, it would seem, is your saving grace: the amazing haste of people to forget your name, your face, your litanous list of indefensible indiscretions.
They’ll glance you over, I guess, and then for a bare moment you’ll persist to exist; almost seems like you’re there, don’t it? But you’re not. You’re here. Your name will fade as Front’s will.
Historians and anthroplogists will likely be the only people who ever regularly access this kind of data if it’s available in the future.
What makes you say that? Who knows what they’ll want to do in the future. Even the most mundane historic records interest today’s archeologists.
That’s a fair question, but I think the answer is obvious. Until the invention of photography, literally the only formal records we had of past events were the things people bothered to write down, paint, or sculpt. And of those, we only have the arts and written records that actually survived. So to find out information about the distant past, we have little choice but to extrapolate from artifacts, dig up old buildings, etc. The artifacts and records that we do find have outsized influence on our understanding of the past, compared to all the information and details that have been lost, which can literally never be recovered.
From the 21st century onward, that relationship is inverted. Any hypothetically useful unit of information about the past will be recorded hundreds or thousands of times, and the useless units of information will outnumber the useful units by many orders of magnitude. Sure, if someone proves to be exceptionally notable, there may be some value in decrypting their past Amazon purchases or cracking the encrypted SSD they left behind. But that’s going to be the exceedingly rare exception, rather than the rule, especially when the world’s data stores are crammed with news articles, photos, videos, interviews, blog posts, reddit posts, journals, and non-encrypted records that appear to tell a complete story of the lives of notable people, and for that matter the day-to-day lives of regular folk.
And that SSD may be every bit as exciting as the Hunter Biden laptop hard disk… that is, barely exciting at all, and full of such routine and irrelevant information as to be an almost pointless exercise in data forensics.
That’s fair. Adding to my point, with the wealth of information future people will have at their disposal, it could be possible to recreate this time era. That is, to simulate entire cities or countries. Who knows what tech they’ll have or what they’ll want to do with it. My point is that the info from this time period, between the advent of the internet and the widespread use of quantum-safe crypto, will be easily accessible to them, and contains such an accurate record of our daily activities. I’ve had the same email address since 2005 and have never deleted messages, so my email alone could probably be used to create a pretty accurate model of a large chunk of my life. Cross-reference that with the information the people I associate with left behind and they definitely could create such a model.
And, adding further, if you were inclined to create such a simulation, you’d likely want to simulate as many people as possible so that the simulation was as realistic as possible.
I can’t believe I forgot about this, but if you really want to explore the question of future people reconstructing the past through AI, watch the movie Marjorie Prime, which is explicitly about this question.
Thanks for the recommendation! I love stuff like that.
We must have VERY different opinions of what our shopping habits or e-mails say about us. My email wouldn’t tell you jack squat.
What about your text messages and phone calls?
Yeah, almost nothing. SMS is a utility tool for me. I doubt anyone will ever care that my wife wanted more zip-lock bags.
You’d get a better picture of me through old USENET posts (which are unencrypted, of course), or reddit or web forums or Lemmy (all of them unencrypted, I suspect). Good luck, future people.
We are a period-genuine curiosity that will certainly be reconstructed in the future—if the data is available to do so. Our lives, logs, and transactions are a finite resource if simply because they are real.
Imagine a hobby, digging up old logs and piecing together various accounts across deep datsets, working toward a bigger picture… and then realize this is all intuited through whatever present AI. No labor involved, at all, and there is no time limit on this.
We’re all eventually just an intelligent query away from being rediscovered if outside of average in any respect, and even the most average person will become a celebrated oddity.