 We had this eG 1020 and let's see what we can improve. So we had those early coordinates. This is what we've got so far. These coordinates are super weird. I don't know why. I think we have these different coordinates from bad. I'll just, I'll just refer to all of them as bots as a full disclaimer using all three. So gpt-button, github, cobalt. So you can see those, these locations are a bit odd. That's not normally how it should look like. I'm just curious where the bot got them from. I don't know if this will generate an image, probably not. It's not there yet. This looks legit. It generates a variable. So you don't understand what the variable of this location is. No, no. Well, we do understand the context of the eG electro locations. So this is what we have in the mod. Obviously the head needs some fixing, not a great head. The ears are not attached to the head, but you get the point. Now, more importantly, I think this, the head should look round and the electro location should fit on top of it. So we have some other locations, so they're obviously from z-minus. No, are they minus one to one or some weird numbers? So we have the names and what else do we need? Yeah, so hopefully we can avoid scaling and we want to plot on a 2D plane. We can have it up to a color function to plot the electro positions. In the 2D plane, we're still thinking we can only have the one option, position, position, position, position and text small. We actually need to use the function plot show, should be it. Are you supposed to solve errors? A much bigger error. It's a dictionary value. Where am I using index? So I need to convert it to their dictionary values. What if I pop it in here? comment this out, just do anything. Yeah, probably not using co-pilot properly. Speedplot, text, position. Okay, we get that, which is not great. What is this? Coordinates are from minus one. Yeah, we have some minus point two in there. There should be one deep line, not two deep line here. Here I can ignore the labels for now. Why? What is this? I'm not worried. Just get a black screen. Why? So the plot is from 0 to 1, both on X and Y. The values are not. Okay, we need to scale first. You can get rid of that for now. Scale is wrong. Proportion, what? Why? A control plot and electro does a circle that looks more like it. They make the OO bit in the plot scaled. This number is wrong. It means max 0 to 1.3. Is it? Okay, so Y, X is going from 0 to 1.3. Y is going from minus 0.4 to 0.2. And so over here, a fit within plot electro location. Make sure the chart should be loaded. What's this number? Right, this number is too large. Right, 0.5. Yes, locations are super weird. Let's see if it updates that. Also, let's get a GPT to generate the same thing or some bad takes forever. Where are they? Look, no, it's not helpful. Obviously, this number should not be a magic number. It should be a circle, a circle size if it's not 0.5. Square root size is proportional to distance from the origin. Why? No, I don't want the size proportion. I want the size to be the same everywhere. Size is proportional to all plot size. Is it based on? No, it's wrong. It actually doesn't look too bad. FP1, F7, F8, or 1 or 2 should in theory be the ease. Anyway, the ones we currently have are totally out of reality. 0.9 should be X, mid and X max. It wasn't outside. The plot area getting error is because the dictionary does not have an index method. Well, that's what the error says. You can connect the dictionary to least using this function and then use the index. The problem is with electrode names get to least. Yeah, it's still having the same error. Why? So we are conveying to this to get the names over the electrodes. Okay, let's get rid of it for a sec just to see if it's just working, but shall we pop it into GPT? What do you say? Used to change GPT so much that now I don't know how to use a co-pilot. Minus 1, values as 0. So it's adding an extra 1. This must be wrong, because the actual values go to... Yeah, this is wrong. So max values of inch electrode locations. This range, there's some problem. X range and these numbers are okay. And let's call this extra 1, make it 2 for a sec. Right now it fits. What's up with those locations? There's only half of them. Something wrong. Okay, those locations are wrong. The interesting thing is that these coordinates are almost correct. I mean, FP1 and FP2 are skewed to the right, but overall, it's pretty close. T5. Yeah, there's a bit of issue with proportions. Cation coordinates on a 2D circle representing a top view of a human head. Pretty sure, don't have to spell anything correctly. The 5th position of key. It's doing some calculations by and it's an interesting one in modern FP2 tried out in one F7, should be F8. I don't know what it is, but it's way off way off. Electrolocation coordinates on to include all electrodes FP1 and FP2 would be on top of the, well, not should be an image, but you know what better way to call it. It's actually nice because it's the coordinates here. I am in the same position. The arrays called EG Electrolocation. It's meant to be the radius, isn't it? This is better. Now it's okay to bring them out, but we really need to plot them. This looks more like it, but not really. Why not just get them the first time around in a correct way? Why did we need to calculate extra coordinates? Why this couldn't be the locations? It's something that the circle that has a problem coordinates to this circle, Cation coordinates on a 2D representation, 2D top view model of a human head with all electrodes. Why didn't it ignore? Yeah, not sure. Why is it not doing it properly? From memory that was wrong as well. It's getting the same thing. Why do I have a feeling this something GPT-4 would be able to solve?