Does Dr. Dray Pass The AI Test?

Okay, it says it's connected, but I'm going to wait here to make sure that it's working as expected. I'm testing a new setup. Okay, I can see myself. So, that's good. It's working. I'm not sure. So, in my last video, I mentioned that I'm doing it well, YouTube didn't monetize every day. And for me to be able to or for us to be able to flag as many of them as possible, we need a faster way to make videos or I need a faster way to make videos.

Streaming should help. um you can just make or put out more information that way and I'll try to make from those streams videos that actually are entertaining because you know people deserve to be entertained. Okay. Uh the point of today's stream or this current stream besides testing my whole setup is I want to test the AI test uh that's evolved so far.

So, the test is always evolving and I want to make sure that it's good enough or it's better as we go into testing more doctors that I really don't know anything about. Uh the one that I'm testing today um I heard was is is a real person and based on a brief look I believe that the the reason that I chose this particular person is because uh Dr. Dre is because they are very productive.

they uh she makes a video every day, which was part of the question that I had for the original um or part of the original test that I had for the Alex Riverly channel. So, what does it actually look like when a presumably real person uh according to one of my commenters, we'll obviously test that, but I'm I'm fairly confident. What does that actually look like?

By having an extreme extremely productive person, a real person as a point of comparison, we can calibrate the test a little bit better. For me, it's not about um it's not about like just a checklist of things. Checklists and processes are like great when you don't want to use your brain so much, right?

Um they just reduce the cognitive effort or cognitive load. I just really have a principle that I'm using to drive the test which is that like as you force AI to introduce variance it makes more errors. Quality control um has to be more rigorous and costs increase. So today we're looking at variance and I have a checklist that I u developed over time.

Okay. Okay. So, we'll start with Dr. Dre. Uh, first thing I I usually do, I think everybody can do this, is just like go to their YouTube page. Uh, okay. I'm just checking the stream here to see. Yeah. So, Angela Lewis. Yeah. So, Angela, thank you for mentioning Dr. Dre. Um, I was really I was really surprised that somebody is so productive and that's why I want I'm testing Dr.

Dre as an extreme example of what it looks like when you have um a real person making videos. So yeah, thank you so much for that. Uh okay, so the first thing that I like to do is like I go and look at the oldest videos. This was 9 years ago. So that's a really good sign already. If if an account is making uh videos for a really long time, like before the AI wave started, which I I'll put around the last four years that it got good enough to fool people.

So 2022, if there are videos before 2022 that are really good, likely that that channel is a person. I'm obviously open to changing that threshold um as I as I you know learn but right now that's what what it's at. I also don't want to discourage new creators, right? You don't you don't want creators to think, oh well the AI wave has started.

I can't compete with AI. Um that's part of why I'm doing this. I want real creators to thrive. Um, yeah. So, don't you get discouraged if you're a real creator. Okay. So, the first vlog uh nine years ago, it's pretty it's a pretty good uh that's a pretty good point of evidence or point for the Dr. Dre channel.

I'll put that as like I don't want to put that as part of the AI test again because of the I don't want to discourage real creators. So, does the video quality make sense? This is why we're looking at the Dr. a great channel. One video a day is like it's really impressive. Uh um so we'll have to to see how she does it.

I I'm very curious. Okay. Um video length long. Is the video length long? Yeah. I mean, a lot of these videos are about 20 minutes, which is like a TV show episode. Okay. You've got one here that's 10 minutes. This is my fake account, by the way. That's why it says A over there. Yeah, I don't know what the average time is.

I don't have the the statistics plugin that I usually have, but it's safe I think it's safe to say that a lot of videos are around 15 to 20 minutes. My cam's too large. Yeah, I could do that. Actually, my editor commented about that. Let me see how what's actually like that might be too small. Uh, okay.

Let's try this. I'm assuming you want to be able to see a good amount of the screen on like what I'm seeing. So, let's just look to see how much you can see. I sort of have some latency here. Let me just refresh my preview page. Okay. So, I'm sort of hiding the bottom right screen. Let me know if that's still too large.

I kind of want be I kind of want to give people the ability to check my facial expressions to if they doubt whether I'm AI, right? Um, maybe I'll try this a little bit smaller. Actually, no. We'll just go with this. We'll just go with this. Let me know if it's too small. I mean, too big, though. Okay.

I do have to make this quick. I'm only doing this 45 more minutes because I have to go to work. Start at 9 central. Okay. Is video length long? Yeah, so video length is about 15 minutes. I'm testing this on Dr. Dre. One video a day is a ton, right? 15 minutes on average. So these like are definitely like I on the higher end of like what is possible I think for a person fatigue wise.

Um, definitely worth documenting historically is it's just really it's admirable. It takes me a lot of time to make a video. Okay, cool. And then here's the here's the core test that I keep I'm going to keep saying and some people don't want to like don't want to absorb it, but I trust that you guys are absorbing it.

We're looking for variance. Uh so in statistics or data science or machine learning variance is like you know changes is just changes and generally it means changes in spec in response to specific variables. We want to see if a if a variable it changes things or if a variable changes another variable.

In this in this case of like analyzing videos, we're just looking overall at how much different type of data does the video have.