Mark Cuban

Meaning what exactly went into the final determination that the AI set out, right? So then that's where you have your board who says, okay, this is what the AI recommends. We have these additional comments to make on additional concerns. And then the board makes the final decision. But you consolidate that review process from what in some cases could be years, a decade to weeks.

You still want some human intervention. You need some common sense. And AIs can do a lot of things, but they're best with repetitive things and best with taking vast volumes of data and information and applying those to processes. And so let the AI do what the AI can do, but leave the final decision to people who understand the context.

You still want some human intervention. You need some common sense. And AIs can do a lot of things, but they're best with repetitive things and best with taking vast volumes of data and information and applying those to processes. And so let the AI do what the AI can do, but leave the final decision to people who understand the context.

So if your job is answering the question yes or no all the time, AI is going to have an impact. If your job requires you to think, AI won't have much of an impact. Really? Yeah. So in this particular case, let's go back to the NIPA type example. Somebody's got to make sure the model is getting all the information and that it's structured correctly. It's all data entry.

So if your job is answering the question yes or no all the time, AI is going to have an impact. If your job requires you to think, AI won't have much of an impact. Really? Yeah. So in this particular case, let's go back to the NIPA type example. Somebody's got to make sure the model is getting all the information and that it's structured correctly. It's all data entry.

No, it's not even just typing, right? It takes intellectual capacity. So somebody who understands what the goal is, somebody who's been doing this for years, has got to be able to input feedback on everything that the models collect and are trained on. You don't just assume the model knows everything. You want somebody to check, to grade their responses and make corrections.

No, it's not even just typing, right? It takes intellectual capacity. So somebody who understands what the goal is, somebody who's been doing this for years, has got to be able to input feedback on everything that the models collect and are trained on. You don't just assume the model knows everything. You want somebody to check, to grade their responses and make corrections.

And so while that model is being trained, is being developed and trained, there's a huge need for people to contribute value. The better example is if you want an artificial intelligence that talks about Shakespeare and spits out things about Shakespeare, somebody's got to teach it, contribute the nuanced things that you're not going to find just from reading their books.

And so while that model is being trained, is being developed and trained, there's a huge need for people to contribute value. The better example is if you want an artificial intelligence that talks about Shakespeare and spits out things about Shakespeare, somebody's got to teach it, contribute the nuanced things that you're not going to find just from reading their books.

What do you got? No, it's the exact opposite right now. Here's the real problem. I'm glad you brought this up. Another great question, Jon Stewart. Boom! Come on! End of year, baby! Let's do this! Damn! Okay, so IP, you've seen the New York Times and others sue different...

What do you got? No, it's the exact opposite right now. Here's the real problem. I'm glad you brought this up. Another great question, Jon Stewart. Boom! Come on! End of year, baby! Let's do this! Damn! Okay, so IP, you've seen the New York Times and others sue different...

creators of large language models like OpenAI, Meta, et cetera, because they just automatically went out and spidered and input their text and their intellectual property into their models. That... intellectual property is still owned by the creator.

creators of large language models like OpenAI, Meta, et cetera, because they just automatically went out and spidered and input their text and their intellectual property into their models. That... intellectual property is still owned by the creator.

And so you're not going to find the models being able to just go out and get everything, everything that was on the internet up to this minute, whatever it is that was openly available and stuff that's added that's openly available. Yes.

And so you're not going to find the models being able to just go out and get everything, everything that was on the internet up to this minute, whatever it is that was openly available and stuff that's added that's openly available. Yes.

Yeah, and I've been through my credit card. Oh, I'm not a subscriber. The ASO, where's my credit card? No, but think it in terms of medicine. And so your MD Anderson or your Stanford Medical or pick all the different, UPMC, and you have a doctor or a researcher that's really good in this specialty.

Yeah, and I've been through my credit card. Oh, I'm not a subscriber. The ASO, where's my credit card? No, but think it in terms of medicine. And so your MD Anderson or your Stanford Medical or pick all the different, UPMC, and you have a doctor or a researcher that's really good in this specialty.

You're not going to just say to Meta or OpenAI or Google, hey, you just take it for your medical models. You just go ahead and take it. You're going to say, no, I'll do my own. I don't need you, and I'll take my specialty. Let's just say my specialty is oncology and cancer. I have all these experts and I'm gonna create my own model.

You're not going to just say to Meta or OpenAI or Google, hey, you just take it for your medical models. You just go ahead and take it. You're going to say, no, I'll do my own. I don't need you, and I'll take my specialty. Let's just say my specialty is oncology and cancer. I have all these experts and I'm gonna create my own model.

I'll go out and use an open source model that's spidered everything that's open and free on the internet, but I'm gonna add all the expertise that I have and all the expertise that I create in the future so that it is specific to me and I can charge for it or use it however I deem fit. And what's gonna end up happening is within medicine, some of the models will pay.