TJ Walker discusses AI video clones that generate personalized audio and video from your content. The technology took five iterations to feel less creepy, and a Delphi CEO demo got 300,000 views but 3,000 negative comments, showing how skeptical people are about the idea.
Questions this video answers
Approximately 300,000 views.
About 3,000 comments; the speaker reports they were overwhelmingly negative.
Chapters
So this, this technology is becoming more and more ubiquitous. You've taken it to the next level through this Delphi company to actually make it more real, like I'm talking to somebody. And by the way, what you have today is going to become infinitely better in a year from now, and it'll become much more realistic than even it is today. And it's already really amazing.
It's radically better than it was on Friday, which.
Exactly, exactly, exactly. And you're using your knowledge base in order to train the model. Essentially, it's. And it's. They're not even. The deli company isn't doing anything that special, I can tell you, because we're not doing anything that special. They've created a language model and essentially it's a private language model that's using OpenAI's AI technology and they're sending that data there and getting responses back based on your content as opposed to the world's content. That's really what's happening. And then it's being translated into either languages, video, audio or what have you. That's all that's happening here. But there's a lot going on and in a year from now, it's going to be infinitely better, as we said.
Yeah. And so, you know, my primary expertise, as you know, is in communication skills, not techno. I'm trying to fuse those two together. But it's, it's not surprising to me that you're, you're incorporating AI into this. But here's my challenge is people think that, oh, text is enough when in. But in real life, if, if an investor wants to meet with you or you want to meet with investor, these days, you either go in person or you do it on Zoom. It's not just you're walking, you're not, you're not just texting back and forth. No. 1. Very few people get significant seven figure, eight figure checks through a text. We're also typically not doing it just through audio. So the visual, I think, is how most important communication takes place. Whether you're in the same room or through Zoom or FaceTime or something like that, visual is most important. And yet so many people have told themselves, oh, I don't need to do that, text is good enough or it's too hard. So that's one area where I see an opening for me, which is why I'm doing a course on how to create your own video clone, but also using my expertise, because just to get the video you saw, I had to go through five different iterations of it.
Right.
Because the problem is, is when engineers are in charge, no offense to the people at Delphi and Haymark at hey Gen and others, they're giving an engineer's perspective. So what they're telling people is be perfectly stiff and nothing's distracting. And then you'll kind of be like this and you won't have any blur or distortion. But the problem is you look like a robot and no one wants to listen to that.
Yeah.
My first reaction is do what I recommend people do on TV and gesture and do that. The problem with that is when someone is a user, is talking to the clone and they see my hands, my hands moving and not I'm not talking, it looks really weird and strange.
Right, right.
Go through all these different versions to get a little bit of movement. Absolutely not show my hands because I wanted it to seem. I wanted it to feel like it's not creepy.
Yep, yep. Absolutely.
Creep. And I want to say hello to Pat, who stopped by to say hello, and maddav and others. And we'll hope, hope to get some of your questions shortly. But the creepiness factor is frankly something I overest or underestimated because the CEO of Delphi went on Fox Business, not the political channel, but Fox Business News with Maria Barnaroma a week and a half ago, did a product demonstration. I thought he did extremely well. Really nothing of controversy was touched on and I looked. So there's a couple of interesting things. His video clip was then posted on the Fox Business News YouTube channel. So not Fox News, not foxbusinessnews.com but just their YouTube channel. The average clip is getting about 30,000 views. His clip, not a celebrity not dealing with controversy, not a well known politician. His clip got 300,000 views in the first few days. It got 3,000 comments. Every one of them was negative. Every one of them was like, this is the worst thing in the world. This is evil. This is criminal. Anyone associated with this should be put in jail. So it does scare a lot of people.
Yep.
Now, I knew that there would be that at some level because every new technology that involves a human face, a voice, does that. First time people saw movies, they jumped out of their seats because they thought the train was coming. People heard a stereo or a phonograph. They thought it was possessed by demons. So new technology that involves anything to do with humans always scares people to death. But I'm wondering if the opposition to this is going to be even more entrenched.
So I think what you find with any new adoption process is that first of all if you're early on. Just like a lot of the other, other very rich, very successful Silicon Valley elites were early on with their ideas, and then they just kind of rode the wave, and they were not overnight successes.
