Editing Openai/69557275-0da8-8004-a03a-8eb02e10a61f (section)

=== User: Okay, we are getting ever closer. ===
Okay, we are getting ever closer. What's missing? I'm going to give you very detailed item-by-item edits yet. Where I think we're still struggling is to thingify each of these to what I want is the title. I mean, we're going for deeper, interesting things, but each one needs to be a specific thing that pops. And we need to catch that thing in the title. And so I don't want vague descriptions of concepts that kind of float. I want concentric circles where we have the inner circle of a specific thing we're pointing the camera at it. And then we're maybe talking a little bit about the broader implication. So I'm going to go item-by-item through your list and try to call that out. And these are my editorial comments. I invite you to think critically about them. But here's my new list of 14. So number one, what I want that to be, and that's going to be an element if you're number one and you're number four, the idea that everyone has an answering service. And we should mention here about how CEOs or heads of states don't read and answer all of their emails or texts, that they have sort of gatekeepers, and that it's going to be like that. Everyone is going to have an answering service that reads all of their emails, reads all of their texts if they wanted to. They'll have different access levels. They can respond to them instantaneously and can flag for the person the ones that they respond to. And I want to be playful here and think about some of what would emerge, which would be certain people would get specific codes that they would use that would allow you to bypass the AI assistant. AI would just automatically pass it on if that person has that code, or I don't know exactly how it will play out. But you can think a little bit about how this will evolve when everyone has an answering service. Number two, AI becomes your financial bloodhound. I think that is great, as you've written it now in your number two. But I think we should just mention that this is going to be sequential. People can do this right now, right, by using prompts and upload their bank statement and look for areas where they're being ripped off. They can create more elaborate prompts that actually do sort of like benchmarking and detailed things. Eventually, it will become fully, fully agentic. And eventually, it will be having your own procurement requirement that will review all of your correspondence. And eventually, people will get to the point that they demand from their banks or from their health insurance company sort of communications or explanations of benefits or charges that can be integrated into their AI, which will be really interesting. See, these second and third order things, I want to zoom in on the concentric circle to give the thing and what it is. And then I want to talk about some of the interesting implications in that further out concentric circles. Number three, your number three, compute. I want this to be bigger. Not just compute becomes a household budget item, but compute time becomes one of the most coveted goods in our society. But you need to show me a little bit more like what the levels of cost would be. There's the free AI. There's the $20 a month AI. There's the $200 a month. Is there a higher one? And what does that allow me to do at each level? And I want you to apply some of the like Nick Bostrom deep utopia here. You know, think about why computation is going to become even more brilliant. And the idea that in like three or four years, everyone can be making their own customized 4K movies every day, two hour movies or something, bespoke movies, like bespoke simulated worlds, things like that. But you need to show me some of the things that are going to be in the AI price range that look more like not $2,000 a year, more like $20,000 a year. Or is there even a $200,000 a year tier in real terms? And what would that look like? So that's number three. Number four, your number four, but let's make it better. I don't like the human guaranteed becomes a luxury tier of reality. What I think it's going to be more interesting is customer service cyborgs. So what I'm imagining is like I call a customer service. I get greeted by an AI. Then it's just AI and human become interchangeable. Like at some point, a human might join the line while the AI is still on the line. The human might provide some direction and then say, OK, you know, the named AI will take it from here. So the AI is almost like the sort of intern trainee that can be on the line with the human. The human kind of interfaces as necessary when flagged by the AI. And then you go back off and do the resolution with the AI. And so it becomes like the customer service experience is an integrated human AI cyborg. That's number four. I'm going to skip your number five. We'll come back to that. So skip your number five. Let's go to your number six, which is this idea of privacy becomes a social behavior, not a setting. I want to recast that. And I want to think a little bit about, you know, how people use technology now during social relationship formation. They share photos, share YouTube videos, share Instagram photos on a first date. You know, people bring their phone out on a when they're together. Friends do this, first dates do this, and they show each other stuff. I want to examine that's going to become more and more AI enabled. And they're going to interact with AI. They're going to have prompts. They're going to have activities that are AI oriented. You're going to share your favorite prompts, share your favorite AI friend, things like that. So AI becomes a part of how we humans socially communicate with each other. So that's number five. Number six I think is good. You can temporarily borrow a specialist. Your number seven is good. The title is maybe a little bit clunky, but we can work on that. Number seven on my list is, okay, it's number eight on yours. The dead become a new kind of address. No, we need to broader that. So death becomes more complicated in that everyone has a complicated AI ecosystem around them. And it's going to be ultimately the choice of the dead person, but involving their ecosystem around them to determine what level of their AI ecosystem, which includes a lot of things, lives on after them. But that's going to become a decision that everyone makes. So that's number seven. Number eight, which is your number nine, social life gets clunky. Let's make this more visible is we increasingly use AI referees for complex human-human interactions, whether they be relationship debates, business strategy discussions. Whenever humans reach an impasse where they're talking past each other, they're not getting each other, it's going to become increasingly common to say, hey, let's get an AI read on this and help us. And that's going to be extraordinarily weird. One of the things we're trying to do is some of these things are more positive, some are negative. We're having fun with just how weird they're going to be. Right. So let's try to make that manifest throughout. OK, my number nine, your number 10, you can smell optimized communication. No, I don't like that. What I want to play with is this idea that, you know, when BlackBerrys first came out and iPhones in business, people would have this little tag at the end of their I'm trying to get much more visceral. Do you see this? Like they would have this little tag at the end of their email saying, please forgive typos. This was written. People are going to have AI assist tags on their communication, and it's going to be incredibly interesting and nuanced because different people are going to have different ways, divide and give AI different rights and come up with their own different way to convey like what of that email is authentically them or authentically them piped into AI versus a pure AI response. And they're going to have their customized language and it's going to be really interesting how these like people stamp their communications with regard to the nature of human interaction. I don't know what to call that. It's going to be you can think about it. Prompt speak becomes a thing. That's my number 10, your number 11. That's good. Let's just maybe make some of the more conceptual types of edits I'm talking about with all these other ones, but I think that that's that's very good. Your number 12. I think this is good. Give me a little. This is my number 11 your number 12. Truth starts coming with stakes. I want to think about some of the more second and third order questions like there should be like there should be referenced bonded, and then somehow, like there should be. I don't know process behind it where I'm thinking if the news publishes an article, and then I have my AI check it and my AI finds an error. What's the process I send that back and it's, it's like anyone can use their AI to go back to the AI behind the, the agentic AI that's linked to the article who may have written it or is now responsible for edits and those get I can submit a truth claim. And so, everything that's published kind of becomes like Wikipedia, in a way, but it's agentic Wikipedia, but obviously there have to be a huge amount of safeguards and permissions will be different and CNN isn't going to probably let an agent change an article just because someone's agent said something but I want to show the weirdness of this.