1 00:00:00,080 --> 00:00:01,600 shooting out. I believe that was 2 00:00:01,600 --> 00:00:05,670 something I've seen. 3 00:00:05,680 --> 00:00:08,160 The trees are quite reminiscent of a 4 00:00:08,160 --> 00:00:10,639 pooh. Deepseek V4 Pro has just been 5 00:00:10,639 --> 00:00:12,960 released in general available version, 6 00:00:12,960 --> 00:00:15,120 which also happened with the V4 Flash 7 00:00:15,120 --> 00:00:17,920 model back about 12 or 13 days ago on 8 00:00:17,920 --> 00:00:21,279 0731. Now, that model when tested proved 9 00:00:21,279 --> 00:00:23,600 to be a significant significant increase 10 00:00:23,600 --> 00:00:25,600 in capability over the preview version 11 00:00:25,600 --> 00:00:27,760 that was previously around. And today 12 00:00:27,760 --> 00:00:29,279 we're going to be finding out if the 13 00:00:29,279 --> 00:00:31,279 same can be said for the newly updated 14 00:00:31,279 --> 00:00:33,920 checkpoint general availability for V4 15 00:00:33,920 --> 00:00:36,000 Pro dubbed0813. 16 00:00:36,000 --> 00:00:37,520 So before we get into it, do feel free 17 00:00:37,520 --> 00:00:39,680 to subscribe so we can hit 70K and then 18 00:00:39,680 --> 00:00:41,680 100K with the plaque. And with that, 19 00:00:41,680 --> 00:00:43,520 basically there's not a lot to say about 20 00:00:43,520 --> 00:00:45,200 this right now because it seems to have 21 00:00:45,200 --> 00:00:47,200 been more or less silently released. 22 00:00:47,200 --> 00:00:48,879 There's no official announcement post 23 00:00:48,879 --> 00:00:50,640 yet on X, which I do believe is 24 00:00:50,640 --> 00:00:52,719 generally where we would see this. But 25 00:00:52,719 --> 00:00:54,719 they do mention here just in the website 26 00:00:54,719 --> 00:00:56,879 a couple references to it where if you 27 00:00:56,879 --> 00:00:58,960 use this model identifier right here for 28 00:00:58,960 --> 00:01:01,440 V4 Pro, you are being routed to the new 29 00:01:01,440 --> 00:01:04,479 V4 Pro0813 version. Additionally to 30 00:01:04,479 --> 00:01:06,880 that, it is also live on open router and 31 00:01:06,880 --> 00:01:08,560 this makes me feel comfortable that it 32 00:01:08,560 --> 00:01:10,640 is actually genuinely released and this 33 00:01:10,640 --> 00:01:12,159 is the model we're going to be using. 34 00:01:12,159 --> 00:01:13,920 So, from the bit of information we have 35 00:01:13,920 --> 00:01:15,840 right here, aside from just the model 36 00:01:15,840 --> 00:01:17,600 identifier now being today's date, 37 00:01:17,600 --> 00:01:19,840 depending on what local one is in, we 38 00:01:19,840 --> 00:01:21,600 can see that the model has a million 39 00:01:21,600 --> 00:01:23,600 context window with a maximum output 40 00:01:23,600 --> 00:01:26,240 length of 384K, which really is a lot. 41 00:01:26,240 --> 00:01:28,080 Now, something that can sometimes happen 42 00:01:28,080 --> 00:01:30,240 with DeepSeek, at least historically, is 43 00:01:30,240 --> 00:01:32,240 there may be a lot of thinking. Now, 44 00:01:32,240 --> 00:01:33,920 this does support selectable thinking 45 00:01:33,920 --> 00:01:35,759 modes, as you can see if we do click on 46 00:01:35,759 --> 00:01:37,759 this. So, it's possible to kind of tone 47 00:01:37,759 --> 00:01:39,520 that down, but in the interest of seeing 48 00:01:39,520 --> 00:01:41,680 the maximum potential performance, all 49 00:01:41,680 --> 00:01:43,600 of our coding tasks for today are going 50 00:01:43,600 --> 00:01:46,560 to be run on this maximum effort. Now, I 51 00:01:46,560 --> 00:01:48,240 want to mention the following with a 52 00:01:48,240 --> 00:01:49,759 grain of salt because I cannot 53 00:01:49,759 --> 00:01:51,600 definitively confirm that this is ground 54 00:01:51,600 --> 00:01:53,600 truth, but I would imagine it probably 55 00:01:53,600 --> 00:01:55,600 is. There were some references and 56 00:01:55,600 --> 00:01:57,439 photos floating around which apparently 57 00:01:57,439 --> 00:01:59,920 came from a group related to deepseek. I 58 00:01:59,920 --> 00:02:01,600 believe they said it was a WeChat group 59 00:02:01,600 --> 00:02:03,759 about Deep Seek or something. And if 60 00:02:03,759 --> 00:02:05,520 this is to be believed, we can see that 61 00:02:05,520 --> 00:02:07,200 these would hypothetically be the 62 00:02:07,200 --> 00:02:09,200 benchmark results for Deepseek V4 63 00:02:09,200 --> 00:02:12,080 Pro0813 or the general availability 64 00:02:12,080 --> 00:02:14,239 model. And basically my main interest 65 00:02:14,239 --> 00:02:15,680 here is just seeing the leap in 66 00:02:15,680 --> 00:02:18,160 capability over the V4 Pro preview to 67 00:02:18,160 --> 00:02:20,959 now because I was initially disappointed 68 00:02:20,959 --> 00:02:23,200 with the V4 Pro Preview as well as the 69 00:02:23,200 --> 00:02:25,440 V4 Flash preview. They felt very 70 00:02:25,440 --> 00:02:27,520 undercooked and probably because they 71 00:02:27,520 --> 00:02:29,280 were named preview. That's actually like 72 00:02:29,280 --> 00:02:31,440 fair. It makes sense. But again, I want 73 00:02:31,440 --> 00:02:33,120 to reiterate the leap in capability 74 00:02:33,120 --> 00:02:35,840 between flash and then the 0731 general 75 00:02:35,840 --> 00:02:37,920 availability version of flash was 76 00:02:37,920 --> 00:02:39,840 absolutely massive. This is like the 77 00:02:39,840 --> 00:02:42,319 best local model right now for like a 78 00:02:42,319 --> 00:02:45,519 dual spark setup. So I mean in a niche, 79 00:02:45,519 --> 00:02:47,760 you know, but still if that leaping 80 00:02:47,760 --> 00:02:49,760 capability remains, then what we can see 81 00:02:49,760 --> 00:02:52,160 right here just in this deep sore 82 00:02:52,160 --> 00:02:55,760 increase from 12.8 to 62.7 would then 83 00:02:55,760 --> 00:02:57,440 hypothetically put this above things 84 00:02:57,440 --> 00:03:00,319 like Opus 4.8 8 and make it very very 85 00:03:00,319 --> 00:03:02,159 competitive being especially that this 86 00:03:02,159 --> 00:03:04,400 is an openweight model or will be. 87 00:03:04,400 --> 00:03:06,400 That's also extremely exciting because 88 00:03:06,400 --> 00:03:09,519 this is not like horrifically gigantic I 89 00:03:09,519 --> 00:03:11,360 guess when considering things like Kimmy 90 00:03:11,360 --> 00:03:14,319 K3 or Quen 3.8 Max. It's still big at 91 00:03:14,319 --> 00:03:16,640 1.6 trillion parameters. It is a mixture 92 00:03:16,640 --> 00:03:19,280 of experts model with how many active? 93 00:03:19,280 --> 00:03:21,120 So if we do estimate that this is going 94 00:03:21,120 --> 00:03:23,120 to be the same size as the preview 95 00:03:23,120 --> 00:03:25,519 version of V4 Pro. It's a 1.6 6 trillion 96 00:03:25,519 --> 00:03:27,760 parameter mixture of experts model with 97 00:03:27,760 --> 00:03:30,159 49 billion active. So it is still very 98 00:03:30,159 --> 00:03:32,799 very large but it's not as gigantic as 99 00:03:32,799 --> 00:03:34,720 like the closing in on 3 trillion 100 00:03:34,720 --> 00:03:37,040 parameter model. So additionally to that 101 00:03:37,040 --> 00:03:38,720 pricing and this is something that we 102 00:03:38,720 --> 00:03:40,720 need to basically take just as it is 103 00:03:40,720 --> 00:03:42,640 right now because they mention and have 104 00:03:42,640 --> 00:03:44,159 been for quite a bit of time that their 105 00:03:44,159 --> 00:03:46,480 API pricing is going to be significantly 106 00:03:46,480 --> 00:03:48,959 increased soon. Now significantly is 107 00:03:48,959 --> 00:03:50,640 relative to what it currently is which 108 00:03:50,640 --> 00:03:53,440 is like very very very cheap. So, right 109 00:03:53,440 --> 00:03:55,519 now, pricing per million input tokens 110 00:03:55,519 --> 00:04:00,480 for DeepSeek V4 Pro is 4.35. 111 00:04:00,480 --> 00:04:03,280 So, okay. No, I know what this is, but 112 00:04:03,280 --> 00:04:05,519 like last time I messed up on this one 113 00:04:05,519 --> 00:04:07,599 and I said it was 2 cents and it wasn't 114 00:04:07,599 --> 00:04:11,200 and then I got roasted. So, this is Z 43 115 00:04:11,200 --> 00:04:13,920 cents per million input tokens and 87 116 00:04:13,920 --> 00:04:16,000 cents per million output tokens is the 117 00:04:16,000 --> 00:04:18,079 current price. Now, if this gets like a 118 00:04:18,079 --> 00:04:20,079 significant increase, say it goes up 119 00:04:20,079 --> 00:04:21,840 like five or 10 times, it would still 120 00:04:21,840 --> 00:04:24,479 hypothetically be competitive. So, it's 121 00:04:24,479 --> 00:04:26,160 all relative to what it currently costs. 122 00:04:26,160 --> 00:04:27,759 And with that, I am going to be using 123 00:04:27,759 --> 00:04:29,919 this from within Codeex. When we tested 124 00:04:29,919 --> 00:04:32,800 V4 Flash, the general availability model 125 00:04:32,800 --> 00:04:35,440 on the 31st of July, I did also use it 126 00:04:35,440 --> 00:04:36,960 through Codeex because I believe they 127 00:04:36,960 --> 00:04:38,800 had suggested to do so and the 128 00:04:38,800 --> 00:04:41,360 performance was very, very good. I was 129 00:04:41,360 --> 00:04:43,040 happy with it. So, I'm going to be using 130 00:04:43,040 --> 00:04:45,199 Codeex as well for this. Now I should 131 00:04:45,199 --> 00:04:48,080 also mention that they did release in 132 00:04:48,080 --> 00:04:50,800 beta a desktop app for codeex or 133 00:04:50,800 --> 00:04:53,280 chatgpt. So I also have it routed and 134 00:04:53,280 --> 00:04:54,960 working through this desktop app as well 135 00:04:54,960 --> 00:04:56,880 and it is doing a skate game for us in 136 00:04:56,880 --> 00:04:58,720 C++. I would imagine that will take 137 00:04:58,720 --> 00:05:00,240 quite a bit of time. So I did just want 138 00:05:00,240 --> 00:05:01,919 to start it and verify this app was 139 00:05:01,919 --> 00:05:03,840 actually working properly on this system 140 00:05:03,840 --> 00:05:06,160 and it does seem to. So keep in mind any 141 00:05:06,160 --> 00:05:09,120 mention of GPT or codecs here is routed 142 00:05:09,120 --> 00:05:11,280 directly to DeepSeek on max reasoning 143 00:05:11,280 --> 00:05:13,759 effort for the 0813 model. And with 144 00:05:13,759 --> 00:05:16,080 that, let's take a look at our Okay, so 145 00:05:16,080 --> 00:05:20,720 apparently the OpenAI beta application 146 00:05:20,720 --> 00:05:22,880 set itself to be the default for all 147 00:05:22,880 --> 00:05:24,720 file types. That's infuriating. All 148 00:05:24,720 --> 00:05:27,759 right. All right. Very Quen 3.8 149 00:05:27,759 --> 00:05:30,080 aesthetic. I'm going to turn sound on 150 00:05:30,080 --> 00:05:31,759 because I did see some mentions of it 151 00:05:31,759 --> 00:05:34,240 just talking about like 152 00:05:34,240 --> 00:05:35,919 sounds. All right. It starts in our 153 00:05:35,919 --> 00:05:39,759 notepad. Can we resize this? All right. 154 00:05:39,759 --> 00:05:43,120 Acceptable. Minimize. Unminimize. Good. 155 00:05:43,120 --> 00:05:45,440 The big one. Is there a right click? 156 00:05:45,440 --> 00:05:48,870 Yeah, there is. 157 00:05:48,880 --> 00:05:50,720 also the correct date and time 158 00:05:50,720 --> 00:05:52,400 in my local in the bottom right. And 159 00:05:52,400 --> 00:05:54,080 there is what I would imagine here to 160 00:05:54,080 --> 00:05:55,759 probably be the special feature. It is 161 00:05:55,759 --> 00:05:57,759 the holographic desktop feature. So, we 162 00:05:57,759 --> 00:05:59,759 will save that for last. Let's start out 163 00:05:59,759 --> 00:06:02,160 by looking at the files app, which I 164 00:06:02,160 --> 00:06:03,840 don't think I've seen a files icon that 165 00:06:03,840 --> 00:06:05,600 looks like this before. All right. We 166 00:06:05,600 --> 00:06:07,199 have some nice transparency to the 167 00:06:07,199 --> 00:06:09,199 application backgrounds and definitely a 168 00:06:09,199 --> 00:06:11,039 nice synth wave theme, which I do 169 00:06:11,039 --> 00:06:12,639 personally like. So, unfortunately, it 170 00:06:12,639 --> 00:06:14,240 seems our fake file system is just 171 00:06:14,240 --> 00:06:15,600 static and we can't browse through it. 172 00:06:15,600 --> 00:06:17,440 Nonetheless, can we create new things in 173 00:06:17,440 --> 00:06:19,680 it? Okay, we can. Can we remove the new 174 00:06:19,680 --> 00:06:23,280 things we created? We also can. Can we 175 00:06:23,280 --> 00:06:24,720 remove things that were there by 176 00:06:24,720 --> 00:06:27,360 default? Yeah, we can. All right, cool. 177 00:06:27,360 --> 00:06:28,960 That makes me a little happier. And then 178 00:06:28,960 --> 00:06:31,360 up would just go there. Oh, okay. Not 179 00:06:31,360 --> 00:06:34,160 bad. Decent bit of immersion, at least 180 00:06:34,160 --> 00:06:35,919 in terms of what directories would exist 181 00:06:35,919 --> 00:06:38,319 were this a real computer. Notepad we 182 00:06:38,319 --> 00:06:40,160 did kind of see because it opened with 183 00:06:40,160 --> 00:06:41,759 this just to give us some information. 184 00:06:41,759 --> 00:06:43,199 The only thing I like to see is will it 185 00:06:43,199 --> 00:06:45,440 save it as a text file on the system. No 186 00:06:45,440 --> 00:06:48,000 such directory. Didn't I delete the 187 00:06:48,000 --> 00:06:50,319 documents directory? Hold on a second. 188 00:06:50,319 --> 00:06:52,720 That's okay. Oh, it remembered that I 189 00:06:52,720 --> 00:06:54,319 did that. I might have messed that up. 190 00:06:54,319 --> 00:06:56,880 So that's going to be that's on me. Uh 191 00:06:56,880 --> 00:07:00,240 okay. This right here is just not right. 192 00:07:00,240 --> 00:07:02,960 It should not have a like visible text 193 00:07:02,960 --> 00:07:05,680 box. That happens sometimes. Help did 194 00:07:05,680 --> 00:07:09,759 give us things. Let's just do Neoetch. 195 00:07:09,759 --> 00:07:11,680 That's actually pretty well done right 196 00:07:11,680 --> 00:07:14,309 there. 197 00:07:14,309 --> 00:07:14,319 there. 198 00:07:14,319 --> 00:07:19,440 Nebula or Nebula. Needula. Still, that's 199 00:07:19,440 --> 00:07:21,039 that's actually all right. I'm just 200 00:07:21,039 --> 00:07:23,759 being kind of mean. Memory. This is 201 00:07:23,759 --> 00:07:26,800 fine. Oh, it actually put a bit of 202 00:07:26,800 --> 00:07:29,840 cleverness in there. 1.44 megs out of 203 00:07:29,840 --> 00:07:31,599 infinity. This is fine. Isn't that how 204 00:07:31,599 --> 00:07:33,280 much a old school floppy disc would 205 00:07:33,280 --> 00:07:35,440 hold? I don't know cuz those are older 206 00:07:35,440 --> 00:07:38,560 than me. Next up, calculator. Okay, so 207 00:07:38,560 --> 00:07:41,520 47 * 208 00:07:41,520 --> 00:07:44,390 8 209 00:07:44,390 --> 00:07:44,400 8 210 00:07:44,400 --> 00:07:45,909 8 376. 211 00:07:45,909 --> 00:07:45,919 376. 212 00:07:45,919 --> 00:07:47,199 I haven't done one of these in a while 213 00:07:47,199 --> 00:07:52,560 divided by six. All right, settings. 214 00:07:52,560 --> 00:07:55,440 Definitely seeing a bit of like Quen 3.8 215 00:07:55,440 --> 00:07:57,280 in here just in terms of some of the 216 00:07:57,280 --> 00:07:59,520 style here. What is that and why do I 217 00:07:59,520 --> 00:08:02,160 like it? 218 00:08:02,160 --> 00:08:05,680 Aurora. Oh, okay. Sunset 219 00:08:05,680 --> 00:08:09,440 City. This No one would understand that 220 00:08:09,440 --> 00:08:11,599 reference. This looks like something one 221 00:08:11,599 --> 00:08:13,199 of my friends in high school was making 222 00:08:13,199 --> 00:08:15,039 as a game project. It actually looks 223 00:08:15,039 --> 00:08:17,440 like a ton like the splash screen for 224 00:08:17,440 --> 00:08:20,319 his project. We have customizable 225 00:08:20,319 --> 00:08:23,280 accents as well. Difficulty easy, 226 00:08:23,280 --> 00:08:26,080 normal, and hard. Oh, for games. Okay, I 227 00:08:26,080 --> 00:08:28,080 probably set it to easy as you may be 228 00:08:28,080 --> 00:08:30,479 familiar with my quick to anger game 229 00:08:30,479 --> 00:08:33,200 playing style. Next up, let's try our 230 00:08:33,200 --> 00:08:36,880 games. The speaker is on. So, City 231 00:08:36,880 --> 00:08:41,110 Heist. 232 00:08:41,120 --> 00:08:46,949 Oh, you absolute. 233 00:08:46,959 --> 00:08:48,640 Okay. So, you have to press enter to 234 00:08:48,640 --> 00:08:52,160 start and not click the button. 235 00:08:52,160 --> 00:08:59,750 Please shift gears. Ah, okay. Okay. 236 00:08:59,760 --> 00:09:03,670 All right. It's 237 00:09:03,680 --> 00:09:11,910 Oh, cool. 238 00:09:11,920 --> 00:09:13,920 Oh, okay. 239 00:09:13,920 --> 00:09:18,470 It's troubled, 240 00:09:18,480 --> 00:09:20,959 but 241 00:09:20,959 --> 00:09:23,040 the sounds for like the smack are kind 242 00:09:23,040 --> 00:09:25,519 of funny. Like, all right. And we 243 00:09:25,519 --> 00:09:27,680 wrecked and we only grabbed a 100red out 244 00:09:27,680 --> 00:09:30,080 of a thousand, but it did work. So, keep 245 00:09:30,080 --> 00:09:31,839 in mind we need to press enter to play 246 00:09:31,839 --> 00:09:33,519 these games and not hit the button that 247 00:09:33,519 --> 00:09:44,389 says launch. Oh. Oh, all right. 248 00:09:44,399 --> 00:09:47,440 Do I go through these? Yeah. I I just 249 00:09:47,440 --> 00:09:49,440 assume based on the sounds. I wonder if 250 00:09:49,440 --> 00:09:50,959 that's how like dogs are since they 251 00:09:50,959 --> 00:09:52,800 don't understand language. But like that 252 00:09:52,800 --> 00:09:54,240 to me seemed like a positive 253 00:09:54,240 --> 00:09:55,839 reinforcement sound. So, it makes me 254 00:09:55,839 --> 00:09:57,760 want to do it more. And I think we did 255 00:09:57,760 --> 00:09:59,440 just discover something about how dog 256 00:09:59,440 --> 00:10:02,240 sprains work. 257 00:10:02,240 --> 00:10:04,959 Okay, the score. I don't know that we 258 00:10:04,959 --> 00:10:07,360 need that many decimal points just to 259 00:10:07,360 --> 00:10:09,040 show a score. I think maybe just 260 00:10:09,040 --> 00:10:12,160 rounding up to 734 or down would have 261 00:10:12,160 --> 00:10:14,800 been acceptable, but I applaud it for 262 00:10:14,800 --> 00:10:18,000 its intricacy. And then, oh, read me. 263 00:10:18,000 --> 00:10:19,680 Okay, so our special feature is 264 00:10:19,680 --> 00:10:21,680 something like holographic. And I would 265 00:10:21,680 --> 00:10:23,440 imagine having a couple apps open just 266 00:10:23,440 --> 00:10:25,279 when we trigger that is probably the 267 00:10:25,279 --> 00:10:28,720 move. So, oh yeah. All right. So, it 268 00:10:28,720 --> 00:10:30,160 just puts like an interesting card 269 00:10:30,160 --> 00:10:31,839 effect. This is what it looks like. This 270 00:10:31,839 --> 00:10:33,839 looks like an Apple Vision Pro like app 271 00:10:33,839 --> 00:10:35,680 demo. This is the way they look from 272 00:10:35,680 --> 00:10:39,120 inside the Vision Pro headset. Um, just 273 00:10:39,120 --> 00:10:42,160 if you have used one, you may resonate 274 00:10:42,160 --> 00:10:45,600 with what I'm saying. Okay. Overall, it 275 00:10:45,600 --> 00:10:47,600 was fine. Everything worked. Nothing 276 00:10:47,600 --> 00:10:49,120 really blew me away. But again, that may 277 00:10:49,120 --> 00:10:51,600 be a limitation of the test and less of 278 00:10:51,600 --> 00:10:54,399 a limitation of the model. So alas, huh, 279 00:10:54,399 --> 00:10:56,560 it's kind of neat. We will continue. So 280 00:10:56,560 --> 00:10:58,800 our skateboard game is still going. And 281 00:10:58,800 --> 00:11:01,120 again, I'm using this with the newly 282 00:11:01,120 --> 00:11:02,959 releasleased beta version of the chat 283 00:11:02,959 --> 00:11:04,959 GPT desktop app. I would imagine this 284 00:11:04,959 --> 00:11:06,720 will take quite a bit of time. So let's 285 00:11:06,720 --> 00:11:08,959 check to see just like Oh no, cuz we 286 00:11:08,959 --> 00:11:11,040 already had looked at the cost. And keep 287 00:11:11,040 --> 00:11:13,920 in mind this is partially for the skate 288 00:11:13,920 --> 00:11:15,839 game as well as the browser OS. So 20 289 00:11:15,839 --> 00:11:18,720 cents total across two separate tasks. 290 00:11:18,720 --> 00:11:20,480 One which is still ongoing. All right. 291 00:11:20,480 --> 00:11:24,320 So, after 68 minutes and 24 seconds, 292 00:11:24,320 --> 00:11:27,839 Skate NYC is complete. So, oh, you 293 00:11:27,839 --> 00:11:29,360 idiot. You spoiled it with these 294 00:11:29,360 --> 00:11:30,959 screenshots and you don't even have 295 00:11:30,959 --> 00:11:33,829 vision. 296 00:11:33,829 --> 00:11:33,839 vision. 297 00:11:33,839 --> 00:11:36,079 I'm infuriated. Nonetheless, now that we 298 00:11:36,079 --> 00:11:38,480 know what the game looks like, I had to 299 00:11:38,480 --> 00:11:40,000 stop myself from saying something a 300 00:11:40,000 --> 00:11:41,760 little more colorful. Let's just take a 301 00:11:41,760 --> 00:11:45,040 peek at at the game. All right, so 302 00:11:45,040 --> 00:11:48,000 here's our NYC skate game. Should be in 303 00:11:48,000 --> 00:11:51,440 outputs, right? Yes. 304 00:11:51,440 --> 00:11:56,160 All right. It's not quite as good as the 305 00:11:56,160 --> 00:11:59,760 Quen 3.8 model, but 306 00:11:59,760 --> 00:12:01,600 that model does have 800 billion more 307 00:12:01,600 --> 00:12:04,870 parameters. 308 00:12:04,880 --> 00:12:07,200 Does that say NYC skate on the ground? 309 00:12:07,200 --> 00:12:09,040 It's inverted, but it does. And it says 310 00:12:09,040 --> 00:12:12,399 established 1999. All right. Uhoh. 311 00:12:12,399 --> 00:12:15,440 Sounded like the beginning of uh the Van 312 00:12:15,440 --> 00:12:18,000 Halen song. I don't know what exactly 313 00:12:18,000 --> 00:12:22,000 is. All right. Uh, let's just maybe not 314 00:12:22,000 --> 00:12:23,760 do that same thing. There are sound 315 00:12:23,760 --> 00:12:25,279 effects, though. I don't normally hear 316 00:12:25,279 --> 00:12:26,639 sound effects. Can we actually move 317 00:12:26,639 --> 00:12:30,959 beyond? All right. Not quite. I will say 318 00:12:30,959 --> 00:12:33,519 it's not wonderful. Oh, that's actually 319 00:12:33,519 --> 00:12:35,760 kind of cool, though. The way it like 320 00:12:35,760 --> 00:12:42,310 stylized the game title. Nice. Nice. 321 00:12:42,320 --> 00:12:43,839 Well, we don't need to do the drum kit 322 00:12:43,839 --> 00:12:45,829 sim. 323 00:12:45,829 --> 00:12:45,839 sim. 324 00:12:45,839 --> 00:12:49,509 The trees are 325 00:12:49,519 --> 00:12:52,959 quite reminiscent of a poo if I were to 326 00:12:52,959 --> 00:12:55,200 be honest, but all right, the the legs 327 00:12:55,200 --> 00:13:02,069 are like kicking. 328 00:13:02,079 --> 00:13:03,839 This may be the funniest skate game 329 00:13:03,839 --> 00:13:05,760 result we've ever seen, which is, you 330 00:13:05,760 --> 00:13:07,120 know, something. All right, we have a 331 00:13:07,120 --> 00:13:10,480 street sign there. The text is a little 332 00:13:10,480 --> 00:13:12,800 sideways, as are the building faces. 333 00:13:12,800 --> 00:13:14,720 This is definitely a troubled result. 334 00:13:14,720 --> 00:13:16,959 However, I do like the water fountain 335 00:13:16,959 --> 00:13:18,480 shooting out. I believe that was 336 00:13:18,480 --> 00:13:22,069 something I've seen. 337 00:13:22,079 --> 00:13:25,120 Okay. All right. I've seen that in I 338 00:13:25,120 --> 00:13:26,959 think the Quen result. Now, here's what 339 00:13:26,959 --> 00:13:29,120 I would like to just quickly do for some 340 00:13:29,120 --> 00:13:30,720 reference. This is the result we 341 00:13:30,720 --> 00:13:32,959 received for the C++ skateboard test 342 00:13:32,959 --> 00:13:35,920 from the preview version of Deep Seek 343 00:13:35,920 --> 00:13:38,160 V4. We can see right here. I'm going to 344 00:13:38,160 --> 00:13:40,560 turn the speaker off. This is pretty 345 00:13:40,560 --> 00:13:43,760 bad. It was like so bad that I actually 346 00:13:43,760 --> 00:13:45,600 remember to go to this specific thing 347 00:13:45,600 --> 00:13:48,160 just to showcase it. So, the leap in 348 00:13:48,160 --> 00:13:50,720 capability based off of what we see 349 00:13:50,720 --> 00:13:52,959 right here to what we just saw in the 350 00:13:52,959 --> 00:13:55,200 skate game that it had created for us, 351 00:13:55,200 --> 00:13:56,639 which we can just look at in these 352 00:13:56,639 --> 00:13:58,800 screenshots here is still pretty 353 00:13:58,800 --> 00:14:00,480 significant. Next up, I'm giving it the 354 00:14:00,480 --> 00:14:02,800 Subway FPS test. I've also added on to 355 00:14:02,800 --> 00:14:04,399 the end of the prompt. Don't worry about 356 00:14:04,399 --> 00:14:06,000 checking over the script. I will 357 00:14:06,000 --> 00:14:07,360 manually check it and if there are 358 00:14:07,360 --> 00:14:09,120 issues I will report them back to you 359 00:14:09,120 --> 00:14:10,800 just because this will really like 360 00:14:10,800 --> 00:14:13,279 extend every single result by like two 361 00:14:13,279 --> 00:14:16,399 times in terms of like timewise length 362 00:14:16,399 --> 00:14:18,079 just in the interest of something I 363 00:14:18,079 --> 00:14:20,320 could do visually in like 5 seconds and 364 00:14:20,320 --> 00:14:21,920 then be like hey this doesn't work. So 365 00:14:21,920 --> 00:14:23,440 we're going to do that just to maybe 366 00:14:23,440 --> 00:14:25,279 like keep things a little tighter. All 367 00:14:25,279 --> 00:14:28,079 right so it did finish the Subway FPS 368 00:14:28,079 --> 00:14:29,680 game. The problem is it just keeps 369 00:14:29,680 --> 00:14:32,079 trying to do like troubleshooting which 370 00:14:32,079 --> 00:14:34,079 it's not good at doing on this system 371 00:14:34,079 --> 00:14:35,440 for some reason. All right, let's check 372 00:14:35,440 --> 00:14:39,199 out our subway FPS. M Rand Avenue. All 373 00:14:39,199 --> 00:14:40,949 right, 374 00:14:40,949 --> 00:14:40,959 right, 375 00:14:40,959 --> 00:14:42,959 going to turn the speaker on. So, it 376 00:14:42,959 --> 00:14:44,800 gave us two options. We can either 377 00:14:44,800 --> 00:14:49,040 explore the station. Okay, it's Hey, you 378 00:14:49,040 --> 00:14:52,560 know what? I think that is the best 379 00:14:52,560 --> 00:14:54,880 subway car that I've ever seen drawn in 380 00:14:54,880 --> 00:14:57,360 this test. And you're going to be like, 381 00:14:57,360 --> 00:14:59,120 well, it doesn't look that great. I 382 00:14:59,120 --> 00:15:00,800 know. It's just that they never really 383 00:15:00,800 --> 00:15:02,639 include these. And when they do, they're 384 00:15:02,639 --> 00:15:06,160 not really very good. So this actually 385 00:15:06,160 --> 00:15:07,680 what was that? If that thing starts 386 00:15:07,680 --> 00:15:10,480 moving I would have probably become more 387 00:15:10,480 --> 00:15:13,440 excited than would be reasonable for you 388 00:15:13,440 --> 00:15:16,160 know responding to such thing. This is 389 00:15:16,160 --> 00:15:18,000 very clean. It's simple. It's nothing 390 00:15:18,000 --> 00:15:19,760 that's blowing me away to be completely 391 00:15:19,760 --> 00:15:21,600 honest with you. Queens Boulevard line. 392 00:15:21,600 --> 00:15:23,279 Okay. We have stairs. We have some 393 00:15:23,279 --> 00:15:24,720 lighting elements and things of the 394 00:15:24,720 --> 00:15:27,120 sort. We have ace if you want to go that 395 00:15:27,120 --> 00:15:31,360 side. And then tunnels 396 00:15:31,360 --> 00:15:33,440 like indicators for the train's trash 397 00:15:33,440 --> 00:15:37,360 bin. Floating particles more simple. And 398 00:15:37,360 --> 00:15:38,639 we can't go up the stairs, which is 399 00:15:38,639 --> 00:15:40,480 fine. We don't want to. Now, of course, 400 00:15:40,480 --> 00:15:43,360 the other feature that was included here 401 00:15:43,360 --> 00:15:46,639 is the FPS. So, let's go back to menu 402 00:15:46,639 --> 00:15:48,639 and then enter the survival FPS. Let's 403 00:15:48,639 --> 00:15:50,000 see if it changes the lighting or 404 00:15:50,000 --> 00:15:52,480 anything. Oh my. That's a good weapon 405 00:15:52,480 --> 00:15:57,600 model. Does it? How? Oh, it does. 406 00:15:57,600 --> 00:15:59,600 I think. 407 00:15:59,600 --> 00:16:02,880 Yeah, it does. It's like smooth now. I 408 00:16:02,880 --> 00:16:05,040 don't see any enemies. Look at when we 409 00:16:05,040 --> 00:16:08,000 run. It's like a waddle. 410 00:16:08,000 --> 00:16:11,279 All right. So, uh the big thing I'm 411 00:16:11,279 --> 00:16:13,040 noticing here is there are no zombie 412 00:16:13,040 --> 00:16:16,160 enemies. So, 413 00:16:16,160 --> 00:16:18,399 good reload. The waddle is just kind of 414 00:16:18,399 --> 00:16:23,030 funny. 415 00:16:23,040 --> 00:16:25,519 Weird. Hey, look. There are shell 416 00:16:25,519 --> 00:16:27,040 casings that actually drop. That's kind 417 00:16:27,040 --> 00:16:29,279 of cool. All right, this is like Yeah, 418 00:16:29,279 --> 00:16:31,440 it's kind of mid. So, I'm going to give 419 00:16:31,440 --> 00:16:34,399 it a feedback piece. Now, I stopped it 420 00:16:34,399 --> 00:16:36,079 because it was just like it had finished 421 00:16:36,079 --> 00:16:38,000 and it was like I can't use this in the 422 00:16:38,000 --> 00:16:39,680 headless environment to check. And at 423 00:16:39,680 --> 00:16:41,040 that point, it's easier to just check it 424 00:16:41,040 --> 00:16:43,440 and tell it what's wrong. So, I'll give 425 00:16:43,440 --> 00:16:45,279 it that honest feedback. So, supposedly 426 00:16:45,279 --> 00:16:47,360 our fixed and polished version of the 427 00:16:47,360 --> 00:16:50,079 Subway FPS is completed. Let's just take 428 00:16:50,079 --> 00:16:52,800 a peek at it. The sound is on. All 429 00:16:52,800 --> 00:16:55,279 right. Interesting. I think it swapped 430 00:16:55,279 --> 00:16:58,560 the positioning of the buttons. So, 431 00:16:58,560 --> 00:17:00,399 all right. 432 00:17:00,399 --> 00:17:02,160 Seven hostiles. It's showing me in the 433 00:17:02,160 --> 00:17:06,480 top right. I unfortunately just can't 434 00:17:06,480 --> 00:17:08,720 locate them. I would have expected it to 435 00:17:08,720 --> 00:17:10,959 have done better. And again, this is one 436 00:17:10,959 --> 00:17:12,559 where we could go back and reference the 437 00:17:12,559 --> 00:17:15,360 previous model of this. So, basically 438 00:17:15,360 --> 00:17:17,760 the preview version, it was 439 00:17:17,760 --> 00:17:19,679 significantly less detailed than this 440 00:17:19,679 --> 00:17:22,710 was. 441 00:17:22,720 --> 00:17:24,720 though unfortunately we don't have 442 00:17:24,720 --> 00:17:26,640 enemies and given the amount of time it 443 00:17:26,640 --> 00:17:28,559 took as well as the kind of multiple 444 00:17:28,559 --> 00:17:30,480 times I told it there were no enemies 445 00:17:30,480 --> 00:17:32,320 I'm regrettably going to call this one 446 00:17:32,320 --> 00:17:34,400 next up I'm giving this a difficult task 447 00:17:34,400 --> 00:17:38,320 to create a 3D model of an RB26 inline 6 448 00:17:38,320 --> 00:17:40,400 engine it must be something that can be 449 00:17:40,400 --> 00:17:42,799 3D printed with little to no support and 450 00:17:42,799 --> 00:17:45,280 fit inside of it a tiny DC motor called 451 00:17:45,280 --> 00:17:47,600 an N20 motor because this is something 452 00:17:47,600 --> 00:17:49,440 that I've told it should like work with 453 00:17:49,440 --> 00:17:51,600 my RC car project actually I don't think 454 00:17:51,600 --> 00:17:53,200 that's specifically in this prompt, but 455 00:17:53,200 --> 00:17:54,960 that's the whole point of this where 456 00:17:54,960 --> 00:17:56,960 this is it should understand and it will 457 00:17:56,960 --> 00:17:59,280 scale it properly to that motor. So far, 458 00:17:59,280 --> 00:18:00,880 I've run this test with a few 459 00:18:00,880 --> 00:18:02,640 intelligent models and I've never really 460 00:18:02,640 --> 00:18:04,000 received anything that was actually 461 00:18:04,000 --> 00:18:06,480 worth printing. So, hopefully sooner 462 00:18:06,480 --> 00:18:08,400 than later that will change. So, our 3D 463 00:18:08,400 --> 00:18:10,400 model task is pretty interesting because 464 00:18:10,400 --> 00:18:12,160 it's just doing a lot of research on 465 00:18:12,160 --> 00:18:14,320 like the actual proper orientation of 466 00:18:14,320 --> 00:18:16,640 specific items on this engine. Like 467 00:18:16,640 --> 00:18:18,160 where is the downpipe, which is an 468 00:18:18,160 --> 00:18:20,080 exhaust component? What side is it? What 469 00:18:20,080 --> 00:18:21,840 is the intake side? What is the exhaust 470 00:18:21,840 --> 00:18:24,880 side? Intake plenum left or right? So, I 471 00:18:24,880 --> 00:18:26,960 do like to see the amount of like 472 00:18:26,960 --> 00:18:28,880 intricate research that's going into the 473 00:18:28,880 --> 00:18:31,200 culmination of this model creation. And 474 00:18:31,200 --> 00:18:33,760 it would be really awesome for like us 475 00:18:33,760 --> 00:18:35,840 to get a model for this task that's 476 00:18:35,840 --> 00:18:37,120 actually printable. It would be the 477 00:18:37,120 --> 00:18:39,840 first one. So, in 1 hour and 10 minutes, 478 00:18:39,840 --> 00:18:42,640 we received our 3D model response. This 479 00:18:42,640 --> 00:18:45,120 is a very, very intricate just bit of 480 00:18:45,120 --> 00:18:47,520 text about the thing it made. Now, of 481 00:18:47,520 --> 00:18:48,799 course, the other part of that is 482 00:18:48,799 --> 00:18:50,480 actually just taking a look at what it 483 00:18:50,480 --> 00:18:52,880 looks like. Though, once I remember what 484 00:18:52,880 --> 00:18:54,559 I named this folder, we will be able to 485 00:18:54,559 --> 00:18:56,720 take a peek. Assembled display STL. I 486 00:18:56,720 --> 00:18:58,559 would imagine that to be the case. 487 00:18:58,559 --> 00:19:00,799 Please be good. I've really like I need 488 00:19:00,799 --> 00:19:04,080 a good one for these. 489 00:19:04,080 --> 00:19:07,039 Actually, it's sideways. Son of a gun. 490 00:19:07,039 --> 00:19:10,480 This did a really good job. 491 00:19:10,480 --> 00:19:13,840 This is actually like 492 00:19:13,840 --> 00:19:17,360 so far the best one. It's not perfect. 493 00:19:17,360 --> 00:19:19,520 We have some sideways turbocharger 494 00:19:19,520 --> 00:19:22,160 issues, though. I mean, this is All 495 00:19:22,160 --> 00:19:23,760 right, let's let's peek at this in 496 00:19:23,760 --> 00:19:25,360 OpenCAD. And there are a bunch of 497 00:19:25,360 --> 00:19:27,200 individual pieces for this. So, we also 498 00:19:27,200 --> 00:19:29,039 want to see if the engine that we told 499 00:19:29,039 --> 00:19:30,640 it needs to fit inside the little remote 500 00:19:30,640 --> 00:19:32,799 control car engine. Let's see if it has 501 00:19:32,799 --> 00:19:36,400 any hope of properly fitting research. 502 00:19:36,400 --> 00:19:38,240 It has a research folder right here just 503 00:19:38,240 --> 00:19:41,679 showing like a bunch of images and stats 504 00:19:41,679 --> 00:19:44,559 for this motor. Very cool though. And 505 00:19:44,559 --> 00:19:46,160 here we have some better preview 506 00:19:46,160 --> 00:19:48,230 renders. 507 00:19:48,230 --> 00:19:48,240 renders. 508 00:19:48,240 --> 00:19:51,440 This is definitely well done. 509 00:19:51,440 --> 00:19:55,280 Turbo side. Okay. Underside. All right. 510 00:19:55,280 --> 00:19:56,880 It's not something that's going to 511 00:19:56,880 --> 00:19:59,440 really easily be 3D printable. Although, 512 00:19:59,440 --> 00:20:01,200 I guess I can't make that determination 513 00:20:01,200 --> 00:20:03,360 because this comes in multiple pieces. 514 00:20:03,360 --> 00:20:05,440 So, it's not designed to be printed as 515 00:20:05,440 --> 00:20:07,520 we first saw it just in one lump like 516 00:20:07,520 --> 00:20:09,120 this. It's designed to be printed 517 00:20:09,120 --> 00:20:11,600 separately. I would like to see. Let's 518 00:20:11,600 --> 00:20:14,720 look now at just the single first piece, 519 00:20:14,720 --> 00:20:16,480 which is ideally where it would be 520 00:20:16,480 --> 00:20:19,440 placing the 3D printed motor that would 521 00:20:19,440 --> 00:20:21,760 go in it. 522 00:20:21,760 --> 00:20:23,840 Unfortunately, printing this without 523 00:20:23,840 --> 00:20:26,559 supports is just not an option in any 524 00:20:26,559 --> 00:20:28,400 potential orientation that we see right 525 00:20:28,400 --> 00:20:30,320 here, which is a bit frustrating. It 526 00:20:30,320 --> 00:20:32,000 almost seems like it has these posts and 527 00:20:32,000 --> 00:20:33,760 holes in it to be able to snap fin on 528 00:20:33,760 --> 00:20:36,799 other parts, which is cool. Okay. Yep. 529 00:20:36,799 --> 00:20:38,559 And I think those are definitely like 530 00:20:38,559 --> 00:20:40,480 attachment posts, which is pretty cool. 531 00:20:40,480 --> 00:20:42,400 This one would be able to be printed 532 00:20:42,400 --> 00:20:44,240 more or less without any supports if we 533 00:20:44,240 --> 00:20:47,280 were to put it flat. Then this should 534 00:20:47,280 --> 00:20:49,760 have the Nissan logo, or at least the 535 00:20:49,760 --> 00:20:53,440 text Nissan written on it. 536 00:20:53,440 --> 00:20:56,080 Yep. And this one's quite well done. No 537 00:20:56,080 --> 00:20:58,799 issues here. The oil pan. Oh, no. Okay. 538 00:20:58,799 --> 00:21:01,280 This is where we This is where we decide 539 00:21:01,280 --> 00:21:02,400 we're not going to be trying to print 540 00:21:02,400 --> 00:21:05,840 this. I'm not disappointed, but I'm not 541 00:21:05,840 --> 00:21:08,799 super impressed either. So, I suppose I 542 00:21:08,799 --> 00:21:10,799 am lukewarm when it comes to judging 543 00:21:10,799 --> 00:21:12,880 this. Definitely, if we just look at 544 00:21:12,880 --> 00:21:15,440 this as a render, even forget about 545 00:21:15,440 --> 00:21:16,960 looking at the individual parts there, 546 00:21:16,960 --> 00:21:19,120 cuz it's not really going to be feasible 547 00:21:19,120 --> 00:21:21,600 to print this, it did a really decent 548 00:21:21,600 --> 00:21:23,600 job just assembling it. The research 549 00:21:23,600 --> 00:21:25,280 that went into this, I think, is almost 550 00:21:25,280 --> 00:21:27,679 probably like I hate to sound like an AI 551 00:21:27,679 --> 00:21:29,600 here, but like the story isn't the 552 00:21:29,600 --> 00:21:31,120 model, it's the research that went into 553 00:21:31,120 --> 00:21:33,520 it. That's like generic like AI slob. It 554 00:21:33,520 --> 00:21:36,240 really did a lot of decent research 555 00:21:36,240 --> 00:21:38,720 here. I mean, did it find this and 556 00:21:38,720 --> 00:21:42,000 download it just to see? No. Okay. So, 557 00:21:42,000 --> 00:21:44,240 this is something it maybe made or maybe 558 00:21:44,240 --> 00:21:46,320 probably found online just to be able to 559 00:21:46,320 --> 00:21:48,080 better assess what the size of this 560 00:21:48,080 --> 00:21:50,159 asset is, but it took a lot of care to 561 00:21:50,159 --> 00:21:51,840 ensure that what it tried to generate 562 00:21:51,840 --> 00:21:53,840 would have hypothetically fit the 563 00:21:53,840 --> 00:21:55,760 specific specifications that were 564 00:21:55,760 --> 00:21:58,960 denoted in the original prompt. So, not 565 00:21:58,960 --> 00:22:01,200 half bad. I'm also going to give this a 566 00:22:01,200 --> 00:22:02,880 simple front-end design test where it 567 00:22:02,880 --> 00:22:04,720 needs to create the watch website. So, 568 00:22:04,720 --> 00:22:06,240 it has to have a pretty cinematic 569 00:22:06,240 --> 00:22:08,320 panning effect over a 3D watch asset 570 00:22:08,320 --> 00:22:10,799 that it creates. I didn't specifically 571 00:22:10,799 --> 00:22:12,720 Okay, no, I did specifically denote that 572 00:22:12,720 --> 00:22:14,559 the watch needs to be a 3D model. So, 573 00:22:14,559 --> 00:22:16,240 we'll see how competent it is at putting 574 00:22:16,240 --> 00:22:18,720 the entirety of this together. So, in 1 575 00:22:18,720 --> 00:22:21,280 hour and 30 minutes, we got our watch 576 00:22:21,280 --> 00:22:23,520 front end result. It again seems to have 577 00:22:23,520 --> 00:22:25,760 a pretty substantial like just list of 578 00:22:25,760 --> 00:22:27,679 what was built here. So, let's take a 579 00:22:27,679 --> 00:22:29,600 look and see how it is. beautiful watch 580 00:22:29,600 --> 00:22:32,880 front end website. 581 00:22:32,880 --> 00:22:35,280 That's actually it's so close to 582 00:22:35,280 --> 00:22:39,039 excellent. The big problem is this, I 583 00:22:39,039 --> 00:22:40,559 would assume, is the watch strap 584 00:22:40,559 --> 00:22:43,039 sticking out. This is 3D, so don't be 585 00:22:43,039 --> 00:22:45,760 fooled by its 2Dishness look. It is 586 00:22:45,760 --> 00:22:48,320 actually 3D. It's slowly panning, too. I 587 00:22:48,320 --> 00:22:50,480 really Okay, so a couple things I do 588 00:22:50,480 --> 00:22:52,880 like the wooden background and then this 589 00:22:52,880 --> 00:22:55,360 like leather card or pad that this watch 590 00:22:55,360 --> 00:22:57,520 is placed on. Additionally to that, the 591 00:22:57,520 --> 00:22:59,280 face looks really good. The materials 592 00:22:59,280 --> 00:23:01,520 look very good. It's just like an 593 00:23:01,520 --> 00:23:04,320 overall it's a high-end aesthetic, which 594 00:23:04,320 --> 00:23:06,080 is exactly what we go for when 595 00:23:06,080 --> 00:23:09,039 performing this test. Now, that's 596 00:23:09,039 --> 00:23:11,200 actually that right there looks good. 597 00:23:11,200 --> 00:23:14,149 The 598 00:23:14,149 --> 00:23:14,159 The 599 00:23:14,159 --> 00:23:16,320 Let's just sit and appreciate for a 600 00:23:16,320 --> 00:23:18,640 minute. Okay. Yeah, that maybe I could 601 00:23:18,640 --> 00:23:21,200 have done without that, but it's 602 00:23:21,200 --> 00:23:23,600 definitely quite something. And then 603 00:23:23,600 --> 00:23:25,039 look at the way that goes away as soon 604 00:23:25,039 --> 00:23:27,760 as we start to scroll down. Quiet luxury 605 00:23:27,760 --> 00:23:30,480 measured in seconds. 606 00:23:30,480 --> 00:23:32,799 Sunburst dial. Okay, it has like a 607 00:23:32,799 --> 00:23:35,280 material card here. The sapphire crystal 608 00:23:35,280 --> 00:23:37,200 is perhaps something I would have maybe 609 00:23:37,200 --> 00:23:39,120 omitted from this section, but the dial 610 00:23:39,120 --> 00:23:41,200 looks good. Two watches, one point of 611 00:23:41,200 --> 00:23:43,200 view. Please render them. Oh, look at 612 00:23:43,200 --> 00:23:45,360 that. That is sick. That really looks 613 00:23:45,360 --> 00:23:46,799 pretty good. I love the way they're 614 00:23:46,799 --> 00:23:48,480 placed on the table. Really the big 615 00:23:48,480 --> 00:23:49,840 thing letting it down is just this 616 00:23:49,840 --> 00:23:52,320 leather slab is inhibiting our view of 617 00:23:52,320 --> 00:23:54,559 these beautiful straps. This one is the 618 00:23:54,559 --> 00:23:56,880 stainless steel option. This one is the 619 00:23:56,880 --> 00:23:59,440 gold option and priced accordingly I 620 00:23:59,440 --> 00:24:01,679 would imagine. Yeah. Although I would 621 00:24:01,679 --> 00:24:05,520 say the jump in price of like it's a 622 00:24:05,520 --> 00:24:07,039 negligible amount to get the gold one. 623 00:24:07,039 --> 00:24:08,320 I'd probably go with this one 624 00:24:08,320 --> 00:24:10,159 personally. Oh, cool. And these ones we 625 00:24:10,159 --> 00:24:12,720 can freely rotate so that helps a bit. 626 00:24:12,720 --> 00:24:15,440 There is no movement to the watch hands 627 00:24:15,440 --> 00:24:16,799 but that's okay. I'll forgive it for 628 00:24:16,799 --> 00:24:19,760 that. But these are pretty darn cool. 629 00:24:19,760 --> 00:24:22,880 These are well drawn, I would say. 630 00:24:22,880 --> 00:24:26,400 Especially the numeral indicators. Look 631 00:24:26,400 --> 00:24:28,320 at the way that they actually have multi 632 00:24:28,320 --> 00:24:30,320 material on them. So, they're not just 633 00:24:30,320 --> 00:24:32,960 like one static material. It's gold and 634 00:24:32,960 --> 00:24:35,360 then black infill or silver and then 635 00:24:35,360 --> 00:24:36,880 black infill as well. So, that's some 636 00:24:36,880 --> 00:24:39,039 attention to detail I do like seeing. As 637 00:24:39,039 --> 00:24:41,520 well as the kind of brushed material 638 00:24:41,520 --> 00:24:43,120 look on the stainless steel and the 639 00:24:43,120 --> 00:24:44,480 gold. It's probably weird we're like 640 00:24:44,480 --> 00:24:47,440 assessing like a 3D watch, but it just 641 00:24:47,440 --> 00:24:49,840 shows like some capability. Okay, that 642 00:24:49,840 --> 00:24:51,120 looks pretty good right there. And then 643 00:24:51,120 --> 00:24:52,960 we have a clean footer as well. I like 644 00:24:52,960 --> 00:24:54,640 this. I'm happy with this. Even though 645 00:24:54,640 --> 00:24:58,720 it's like not perfect, it's darn close. 646 00:24:58,720 --> 00:25:00,480 So, next up, just here from within 647 00:25:00,480 --> 00:25:02,000 codeex. And I can't show this full 648 00:25:02,000 --> 00:25:04,080 prompt because it has an API key. I'm 649 00:25:04,080 --> 00:25:06,159 giving this the full-on cinematic Steve 650 00:25:06,159 --> 00:25:08,080 the PC repair game test where it needs 651 00:25:08,080 --> 00:25:10,320 to create a cinematic game. It needs to 652 00:25:10,320 --> 00:25:12,320 create voice assets for these characters 653 00:25:12,320 --> 00:25:14,559 just by using an open AI API key that 654 00:25:14,559 --> 00:25:16,720 I've given it and it needs to make a fun 655 00:25:16,720 --> 00:25:19,200 game that is enjoyable and playable. I 656 00:25:19,200 --> 00:25:21,120 will assess this is probably going to 657 00:25:21,120 --> 00:25:22,720 take north of an hour. So, we're just 658 00:25:22,720 --> 00:25:24,480 going to basically leave Codeex here in 659 00:25:24,480 --> 00:25:26,640 the background, the beta desktop version 660 00:25:26,640 --> 00:25:28,960 of it to complete this task and we'll 661 00:25:28,960 --> 00:25:30,400 maybe work on some other things in the 662 00:25:30,400 --> 00:25:32,799 meantime. So, our Steve the PC repairman 663 00:25:32,799 --> 00:25:35,679 cinematic game completed in basically 65 664 00:25:35,679 --> 00:25:37,840 minutes. So, a little over an hour. All 665 00:25:37,840 --> 00:25:40,159 right, here's how to play. I'll probably 666 00:25:40,159 --> 00:25:42,400 just do what it says here, not assume 667 00:25:42,400 --> 00:25:44,159 that it's already running this in some 668 00:25:44,159 --> 00:25:47,360 server for us. So, okay, episode one, 669 00:25:47,360 --> 00:25:50,720 the Azure palm job. Okay, click to 670 00:25:50,720 --> 00:25:52,880 begin. Headphones recommended. We have a 671 00:25:52,880 --> 00:25:55,110 speaker. 672 00:25:55,110 --> 00:25:55,120 speaker. 673 00:25:55,120 --> 00:25:58,159 >> All right. 674 00:25:58,159 --> 00:25:59,600 >> My name is Steve. 675 00:25:59,600 --> 00:26:00,240 >> This is good. 676 00:26:00,240 --> 00:26:02,159 >> I fix computers. 677 00:26:02,159 --> 00:26:04,240 >> Most of my clients are old enough to 678 00:26:04,240 --> 00:26:06,480 remember when a hard drive was a luxury 679 00:26:06,480 --> 00:26:09,440 item. They pay what they can afford and 680 00:26:09,440 --> 00:26:14,400 I make it work. Mostly honest business. 681 00:26:14,400 --> 00:26:16,880 >> Oh, thank you, dear. The screen kept 682 00:26:16,880 --> 00:26:19,520 telling me the date was 1980 and I 683 00:26:19,520 --> 00:26:20,960 thought all fixed. 684 00:26:20,960 --> 00:26:22,640 >> I have been traveling long enough. 685 00:26:22,640 --> 00:26:24,799 >> Fresh battery. Your computer now 686 00:26:24,799 --> 00:26:25,110 believes 687 00:26:25,110 --> 00:26:25,120 believes 688 00:26:25,120 --> 00:26:26,640 >> I will bring you oatmeal cookies on 689 00:26:26,640 --> 00:26:27,110 Tuesday. 690 00:26:27,110 --> 00:26:27,120 Tuesday. 691 00:26:27,120 --> 00:26:27,919 >> Same as the rest. 692 00:26:27,919 --> 00:26:29,440 >> Give Mr. Buttons a scratch behind me 693 00:26:29,440 --> 00:26:30,240 ears for me. 694 00:26:30,240 --> 00:26:34,080 >> Let me walk you out, ma'am. 695 00:26:34,080 --> 00:26:36,720 >> I'll be right with you, Mr. Thomas, 696 00:26:36,720 --> 00:26:38,720 >> the speech overlays itself. 697 00:26:38,720 --> 00:26:40,640 >> A good boy, Steven. 698 00:26:40,640 --> 00:26:43,360 >> If that quiet man needs fixed, you be 699 00:26:43,360 --> 00:26:43,909 patient. 700 00:26:43,909 --> 00:26:43,919 patient. 701 00:26:43,919 --> 00:26:45,039 patient. >> Punctual. 702 00:26:45,039 --> 00:26:48,159 >> I like punctual. Door is unlocked. 703 00:26:48,159 --> 00:26:49,039 Coffee is warm. 704 00:26:49,039 --> 00:26:50,880 >> A man. What is the job? 705 00:26:50,880 --> 00:26:55,120 >> Kane for 2 years. My life. 706 00:26:55,120 --> 00:26:57,120 >> See the the head tilt movement back and 707 00:26:57,120 --> 00:26:59,200 forth is actually there's three people 708 00:26:59,200 --> 00:27:00,960 talking now. And he forgets the entire 709 00:27:00,960 --> 00:27:04,320 name Charles buy him penthouse private 710 00:27:04,320 --> 00:27:06,310 security 711 00:27:06,310 --> 00:27:06,320 security 712 00:27:06,320 --> 00:27:07,919 wants his legend some people would like 713 00:27:07,919 --> 00:27:10,000 to ask him question him slowly the 714 00:27:10,000 --> 00:27:13,120 client pays 100% upront I would like is 715 00:27:13,120 --> 00:27:15,840 also up in the case all expenses first 716 00:27:15,840 --> 00:27:17,840 class two tickets this is why we work 717 00:27:17,840 --> 00:27:20,070 together 718 00:27:20,070 --> 00:27:20,080 together 719 00:27:20,080 --> 00:27:22,480 the top tray is a normal file on a 720 00:27:22,480 --> 00:27:28,149 laptop 721 00:27:28,159 --> 00:27:31,120 it Quiet is my second favorite thing. 722 00:27:31,120 --> 00:27:32,880 Professional. I will not ask what the 723 00:27:32,880 --> 00:27:35,200 first is. The phone rings when the 724 00:27:35,200 --> 00:27:37,120 helicopter is ready. 725 00:27:37,120 --> 00:27:40,080 >> 40. 726 00:27:40,080 --> 00:27:42,880 >> Cut scene over. Thank goodness. Click to 727 00:27:42,880 --> 00:27:46,400 take control. All right. Aside from the 728 00:27:46,400 --> 00:27:48,559 sounds being just totally messed up, 729 00:27:48,559 --> 00:27:51,520 this is actually 730 00:27:51,520 --> 00:27:56,880 pretty decent. Okay. The cat is 731 00:27:56,880 --> 00:27:59,840 There's definitely some 732 00:27:59,840 --> 00:28:01,919 geometric issues. What are we supposed 733 00:28:01,919 --> 00:28:04,080 to do? Take the briefcase when you're 734 00:28:04,080 --> 00:28:07,039 ready. Okay. Where is it? 735 00:28:07,039 --> 00:28:10,000 Well, pet buttons are like, 736 00:28:10,000 --> 00:28:12,240 >> "Don't look at me like that. Buttons." 737 00:28:12,240 --> 00:28:12,799 >> Easter egg. 738 00:28:12,799 --> 00:28:14,240 >> A man's got to eat. 739 00:28:14,240 --> 00:28:17,919 >> Okay. Where's the briefcase? 740 00:28:17,919 --> 00:28:19,520 All right. 741 00:28:19,520 --> 00:28:22,960 We have inventory. Zip ties. Okay. E to 742 00:28:22,960 --> 00:28:24,399 continue. All right. Is it going to put 743 00:28:24,399 --> 00:28:24,630 us 744 00:28:24,630 --> 00:28:24,640 us 745 00:28:24,640 --> 00:28:25,279 us >> Miami? 746 00:28:25,279 --> 00:28:25,679 >> Nice. 747 00:28:25,679 --> 00:28:28,960 >> Palm trees, humidity, and one very bad 748 00:28:28,960 --> 00:28:31,840 man, Mr. Colton. 749 00:28:31,840 --> 00:28:32,559 >> Champagne perhaps. 750 00:28:32,559 --> 00:28:33,919 >> I'm just going to click through cuz it's 751 00:28:33,919 --> 00:28:34,789 like 752 00:28:34,789 --> 00:28:34,799 like 753 00:28:34,799 --> 00:28:37,919 >> black like my well coffee 754 00:28:37,919 --> 00:28:40,559 >> asset at the resort. 755 00:28:40,559 --> 00:28:41,919 >> Poolside at noon. 756 00:28:41,919 --> 00:28:43,120 >> Oh, poolside. 757 00:28:43,120 --> 00:28:46,720 >> 170 enters. Camera coverage. 758 00:28:46,720 --> 00:28:50,080 Elevator. There is no dog. Blind is the 759 00:28:50,080 --> 00:28:53,440 west service stairwell. 760 00:28:53,440 --> 00:28:54,720 Well, I think we just saw a glimpse of 761 00:28:54,720 --> 00:28:56,559 the pool. 762 00:28:56,559 --> 00:29:01,279 All right. Oh, great. Azure Palm Resort. 763 00:29:01,279 --> 00:29:03,600 Click to take control. 764 00:29:03,600 --> 00:29:06,480 It's even got like the like low rent 765 00:29:06,480 --> 00:29:08,240 wallpaper that you'd expect at one of 766 00:29:08,240 --> 00:29:11,200 these joints. Oh, wow. This is actually 767 00:29:11,200 --> 00:29:12,880 like kind of intricate. Oh, good. Okay. 768 00:29:12,880 --> 00:29:14,159 I was worried we wouldn't be able to 769 00:29:14,159 --> 00:29:16,720 just walk through the walls, but okay. 770 00:29:16,720 --> 00:29:18,799 We can toss a coin in the fountain. Your 771 00:29:18,799 --> 00:29:22,080 wish. No more CEOs errors. 50/50. It 772 00:29:22,080 --> 00:29:24,559 seems like Steve's shop and the airplane 773 00:29:24,559 --> 00:29:27,360 scene is like oddly placed inside of the 774 00:29:27,360 --> 00:29:29,760 location we're supposed to be at. Let's 775 00:29:29,760 --> 00:29:31,760 just ignore that. 776 00:29:31,760 --> 00:29:34,320 So, check in at the front desk and then 777 00:29:34,320 --> 00:29:37,840 reach the 17th floor. 778 00:29:37,840 --> 00:29:40,960 >> Welcome to the Azure Palm. Checking in, 779 00:29:40,960 --> 00:29:42,880 Mr. Me. 780 00:29:42,880 --> 00:29:45,120 >> Coulter. Steven Coulter. 781 00:29:45,120 --> 00:29:45,679 >> Mr. 782 00:29:45,679 --> 00:29:48,720 >> I believe I have a reservation. Room 783 00:29:48,720 --> 00:29:52,080 1408, partial ocean view. I do 784 00:29:52,080 --> 00:29:54,080 apologize. The ice machine on your floor 785 00:29:54,080 --> 00:29:55,600 is out of order. 786 00:29:55,600 --> 00:30:00,149 >> Story of my life. 787 00:30:00,159 --> 00:30:01,760 >> All right. We now have a room key in our 788 00:30:01,760 --> 00:30:09,590 inventory. So, is this the elevator? Oh, 789 00:30:09,600 --> 00:30:12,000 is that 790 00:30:12,000 --> 00:30:15,440 reach? Okay. West service stairwell. 791 00:30:15,440 --> 00:30:20,789 Let's just assume this is north. 792 00:30:20,799 --> 00:30:42,549 Good 793 00:30:42,559 --> 00:30:44,080 like 794 00:30:44,080 --> 00:30:48,399 what? Well, that is 795 00:30:48,399 --> 00:30:50,640 frustrating because it gave us this 796 00:30:50,640 --> 00:30:53,279 really like anticipatory music. I'm sure 797 00:30:53,279 --> 00:30:54,799 there's a far better terminology for 798 00:30:54,799 --> 00:30:56,000 what I was about to say. 799 00:30:56,000 --> 00:30:57,200 >> My name is Steve. 800 00:30:57,200 --> 00:30:59,840 >> Oh, no. Don't. 801 00:30:59,840 --> 00:31:02,000 >> Most of my clients are old enough to 802 00:31:02,000 --> 00:31:04,720 remember when a hard drive was a luxur. 803 00:31:04,720 --> 00:31:07,600 >> All right. We we got most of the way. I 804 00:31:07,600 --> 00:31:10,720 think this is a test that more often 805 00:31:10,720 --> 00:31:13,360 than not, we don't complete it. I will 806 00:31:13,360 --> 00:31:15,120 say the elements of this judged 807 00:31:15,120 --> 00:31:17,120 individually. Definitely the pieces were 808 00:31:17,120 --> 00:31:18,559 all there. I don't know that the puzzle 809 00:31:18,559 --> 00:31:20,399 made the correct picture that was shown 810 00:31:20,399 --> 00:31:22,000 on like the puzzle box, but it was 811 00:31:22,000 --> 00:31:24,159 pretty darn close and I saw some 812 00:31:24,159 --> 00:31:25,919 truthfully this did a better job than I 813 00:31:25,919 --> 00:31:27,760 expected, especially considering like 814 00:31:27,760 --> 00:31:29,679 the subway scene game and things like 815 00:31:29,679 --> 00:31:32,799 that. So, I'm actually 816 00:31:32,799 --> 00:31:35,679 I'm okay with this. So, the final thing 817 00:31:35,679 --> 00:31:38,000 was of course the street yeet game, 818 00:31:38,000 --> 00:31:40,320 which is weird because this piece of 819 00:31:40,320 --> 00:31:42,080 text moving up and down like that while 820 00:31:42,080 --> 00:31:43,840 this one remains completely static makes 821 00:31:43,840 --> 00:31:46,480 me uncomfortable in some level. So, this 822 00:31:46,480 --> 00:31:49,760 is a Bejian issue. All right, good. So, 823 00:31:49,760 --> 00:31:52,480 I neglected to um just run it in a 824 00:31:52,480 --> 00:31:57,279 server. One fist, zero mercy. 825 00:31:57,279 --> 00:32:04,310 I already like this. Oh, it seems like 826 00:32:04,320 --> 00:32:06,399 Good music. All right. All right. All 827 00:32:06,399 --> 00:32:10,470 right. 828 00:32:10,480 --> 00:32:16,789 Let's just explore the city real quick. 829 00:32:16,799 --> 00:32:20,000 I don't even know. Okay, so I saw a car 830 00:32:20,000 --> 00:32:25,509 get like 831 00:32:25,519 --> 00:32:34,789 This one's actually kind of fun. 832 00:32:34,799 --> 00:32:38,640 Excuse me. Excuse me. Pardon me. It does 833 00:32:38,640 --> 00:32:40,960 like a like a 834 00:32:40,960 --> 00:32:42,880 Okay. I don't know that we hit. Is that 835 00:32:42,880 --> 00:32:47,669 a Can we hit the bike? 836 00:32:47,679 --> 00:32:54,549 Bonk. 837 00:32:54,559 --> 00:33:00,950 Uh-oh. Jay-walking pedestrian. 838 00:33:00,960 --> 00:33:02,960 Can we punch this parking meter? Yeah, 839 00:33:02,960 --> 00:33:04,720 we can just everything except the 840 00:33:04,720 --> 00:33:06,159 buildings. All right. There's 841 00:33:06,159 --> 00:33:08,240 unfortunately the orientation of some of 842 00:33:08,240 --> 00:33:10,240 these things is not quite right. We can 843 00:33:10,240 --> 00:33:12,000 see that we're moving sideways when we 844 00:33:12,000 --> 00:33:15,750 move forward. 845 00:33:15,760 --> 00:33:18,880 And the cars are also 846 00:33:18,880 --> 00:33:21,440 Sorry, pal. 847 00:33:21,440 --> 00:33:23,110 Oh, 848 00:33:23,110 --> 00:33:23,120 Oh, 849 00:33:23,120 --> 00:33:25,200 and it's interesting how they uh 850 00:33:25,200 --> 00:33:26,960 interact with the scenery as well. Can 851 00:33:26,960 --> 00:33:28,720 we eat this statue? I don't think we 852 00:33:28,720 --> 00:33:34,149 can, but Oh, yeah, we can. All right. 853 00:33:34,159 --> 00:33:36,640 I will say the cartooness of this game 854 00:33:36,640 --> 00:33:38,960 is actually kind of well done. I don't 855 00:33:38,960 --> 00:33:40,799 know how easy that is to assess just 856 00:33:40,799 --> 00:33:42,720 watching someone play it, but like the 857 00:33:42,720 --> 00:33:44,880 pacing is decent. I think maybe that's 858 00:33:44,880 --> 00:33:47,600 attributed to the music as well, but the 859 00:33:47,600 --> 00:33:49,039 way that when you actually All right, 860 00:33:49,039 --> 00:33:50,399 the the loop is getting a little 861 00:33:50,399 --> 00:33:52,720 repetitive when you actually yeet the 862 00:33:52,720 --> 00:33:54,799 things and they also go and interact 863 00:33:54,799 --> 00:33:56,799 with other items in the environment. So 864 00:33:56,799 --> 00:33:59,360 like how that will bounce off the 865 00:33:59,360 --> 00:34:00,880 building and then maybe hit other things 866 00:34:00,880 --> 00:34:03,039 in the path. That's pretty cool. And 867 00:34:03,039 --> 00:34:05,039 this was 868 00:34:05,039 --> 00:34:07,760 actually quite arcade style and 869 00:34:07,760 --> 00:34:09,430 playable. 870 00:34:09,430 --> 00:34:09,440 playable. 871 00:34:09,440 --> 00:34:11,280 Our score is going up. We have combo 872 00:34:11,280 --> 00:34:14,480 multipliers overall. 38 here is a 873 00:34:14,480 --> 00:34:16,560 success. All right, so that is probably 874 00:34:16,560 --> 00:34:18,079 going to conclude our first look and 875 00:34:18,079 --> 00:34:20,560 test of the general availability DeepS 876 00:34:20,560 --> 00:34:23,760 V4 Pro, the 0813 model designation. Now, 877 00:34:23,760 --> 00:34:25,520 this is definitely an improvement over 878 00:34:25,520 --> 00:34:27,359 the preview version that felt pretty 879 00:34:27,359 --> 00:34:28,960 undercooked and I was not at all 880 00:34:28,960 --> 00:34:30,879 impressed with it. But if we look back 881 00:34:30,879 --> 00:34:32,560 at some of the tests that were run in 882 00:34:32,560 --> 00:34:34,159 that video, which is on the channel, we 883 00:34:34,159 --> 00:34:36,079 ran some of the same ones today. This is 884 00:34:36,079 --> 00:34:38,879 a significant improvement. I would say 885 00:34:38,879 --> 00:34:40,399 though, I think the flash model 886 00:34:40,399 --> 00:34:42,159 impressed me more than this, especially 887 00:34:42,159 --> 00:34:44,320 considering the size of it. This has a 888 00:34:44,320 --> 00:34:46,720 bit more competition, but again, it is a 889 00:34:46,720 --> 00:34:48,720 big leap on its predecessor, which I 890 00:34:48,720 --> 00:34:50,320 guess is what we like to see. I was 891 00:34:50,320 --> 00:34:52,159 impressed a few times with some things 892 00:34:52,159 --> 00:34:54,240 like basically results within results. 893 00:34:54,240 --> 00:34:56,560 The cinematic Steve the PC repair game, 894 00:34:56,560 --> 00:34:58,640 it actually had some elements to it that 895 00:34:58,640 --> 00:34:59,920 were like, hey, you know, it kind of 896 00:34:59,920 --> 00:35:01,599 thought this through. Just the way when 897 00:35:01,599 --> 00:35:03,040 they were speaking, their heads were 898 00:35:03,040 --> 00:35:04,640 like moving back and forth. Not the 899 00:35:04,640 --> 00:35:06,160 weird chin thing that was happening. I'm 900 00:35:06,160 --> 00:35:08,240 not sure what happened there. The music 901 00:35:08,240 --> 00:35:10,079 it included in both that game and then 902 00:35:10,079 --> 00:35:12,240 the Street Eat game were actually like 903 00:35:12,240 --> 00:35:14,480 they definitely fit the scene that it 904 00:35:14,480 --> 00:35:16,160 was drawing here and it did a good job 905 00:35:16,160 --> 00:35:17,599 making the Street Eek game very 906 00:35:17,599 --> 00:35:20,000 cartoonish and the playability of it 907 00:35:20,000 --> 00:35:21,680 felt good, I guess, for lack of a better 908 00:35:21,680 --> 00:35:24,000 term. The model that we got of the 909 00:35:24,000 --> 00:35:27,040 engine was actually it wasn't printable. 910 00:35:27,040 --> 00:35:29,119 So, unfortunately, maybe it failed to 911 00:35:29,119 --> 00:35:31,040 like check that box in the specific 912 00:35:31,040 --> 00:35:32,880 task. However, if we just look at one of 913 00:35:32,880 --> 00:35:34,960 the previews right here, it did a pretty 914 00:35:34,960 --> 00:35:37,119 decent job. Definitely better than the 915 00:35:37,119 --> 00:35:38,880 other models I've tested this with. 916 00:35:38,880 --> 00:35:40,640 Though, I have not tested this specific 917 00:35:40,640 --> 00:35:42,560 prompt with any big beefy Frontier 918 00:35:42,560 --> 00:35:44,640 models yet. So, keep that in mind as 919 00:35:44,640 --> 00:35:46,079 well. But, it definitely had some chops 920 00:35:46,079 --> 00:35:47,839 here. And the amount of research it did 921 00:35:47,839 --> 00:35:50,560 to actually get to that point was quite 922 00:35:50,560 --> 00:35:52,480 a lot. We have all the things in here 923 00:35:52,480 --> 00:35:54,000 downloading these images when it doesn't 924 00:35:54,000 --> 00:35:55,760 have vision. So, I'm not quite sure what 925 00:35:55,760 --> 00:35:57,760 to make of that. But nonetheless, I 926 00:35:57,760 --> 00:35:59,359 think probably the biggest point of 927 00:35:59,359 --> 00:36:02,560 frustration for me is it just goes all 928 00:36:02,560 --> 00:36:04,880 out when trying to actually ensure that 929 00:36:04,880 --> 00:36:07,040 every aspect of every line of code it 930 00:36:07,040 --> 00:36:09,520 made is correct. And with a model like 931 00:36:09,520 --> 00:36:11,920 this where it does not have vision, 932 00:36:11,920 --> 00:36:14,320 sometimes it takes longer and there is a 933 00:36:14,320 --> 00:36:16,000 cost thing to it as well. Assuming 934 00:36:16,000 --> 00:36:17,440 you're going to be paying to use these 935 00:36:17,440 --> 00:36:20,640 models, it just sometimes is simpler to 936 00:36:20,640 --> 00:36:22,480 have it make the result, not try to do 937 00:36:22,480 --> 00:36:24,000 like overviews and checks and 938 00:36:24,000 --> 00:36:25,839 follow-ups. you look at it yourself and 939 00:36:25,839 --> 00:36:27,359 then tell it what's wrong and tell it to 940 00:36:27,359 --> 00:36:30,079 fix it. Having to just go all out and it 941 00:36:30,079 --> 00:36:32,079 trying to fix it adds a lot of time and 942 00:36:32,079 --> 00:36:34,960 expense and tokens to the task, which 943 00:36:34,960 --> 00:36:36,960 maybe is more of a consideration when 944 00:36:36,960 --> 00:36:39,599 doing zeroot tests like this, but I find 945 00:36:39,599 --> 00:36:41,280 that's something I've consistently had 946 00:36:41,280 --> 00:36:43,119 frustrations with as well as with the 947 00:36:43,119 --> 00:36:45,119 flash model where if I'm having them 948 00:36:45,119 --> 00:36:47,040 make like a say street heat, it's easier 949 00:36:47,040 --> 00:36:48,720 if they just spit the script out, I test 950 00:36:48,720 --> 00:36:50,160 it and say, "Hey, the buildings are 951 00:36:50,160 --> 00:36:51,920 disappeared." and then it fixes it 952 00:36:51,920 --> 00:36:53,599 rather than it trying to go through and 953 00:36:53,599 --> 00:36:54,960 like well let me try to spin up a 954 00:36:54,960 --> 00:36:56,640 headless browser and then I'll take a 955 00:36:56,640 --> 00:36:58,640 screenshot and like do a heat map 956 00:36:58,640 --> 00:37:00,320 because it can't actually see it or some 957 00:37:00,320 --> 00:37:02,720 other function. So it's a little 958 00:37:02,720 --> 00:37:04,230 tokenhappy. 959 00:37:04,230 --> 00:37:04,240 tokenhappy. 960 00:37:04,240 --> 00:37:06,320 I did run it on the highest potential 961 00:37:06,320 --> 00:37:07,760 reasoning mode. So that could be 962 00:37:07,760 --> 00:37:10,560 attributed to it as well, but I still 963 00:37:10,560 --> 00:37:12,800 would like to be able to do that without 964 00:37:12,800 --> 00:37:14,640 it like going all out to try to make 965 00:37:14,640 --> 00:37:16,480 sure it's polishing every aspect of the 966 00:37:16,480 --> 00:37:18,480 script because I want max performance. I 967 00:37:18,480 --> 00:37:20,400 just don't need that level of depth. So, 968 00:37:20,400 --> 00:37:22,720 that may be a personal complaint, but 969 00:37:22,720 --> 00:37:24,640 you know. So, for the final thing, let's 970 00:37:24,640 --> 00:37:26,320 take a look at everything we spent for 971 00:37:26,320 --> 00:37:28,960 the entirety of this test. We can see 972 00:37:28,960 --> 00:37:31,920 right here that all cost $1.91. 973 00:37:31,920 --> 00:37:33,760 So, that was pretty darn cheap. And you 974 00:37:33,760 --> 00:37:35,839 do have to remember this is going to be 975 00:37:35,839 --> 00:37:38,320 increased their API prices. They say 976 00:37:38,320 --> 00:37:40,640 significant increase expected, but I 977 00:37:40,640 --> 00:37:42,400 mean a significant increase might be 978 00:37:42,400 --> 00:37:44,320 like, okay, the entirety of this test 979 00:37:44,320 --> 00:37:46,640 was $8 or something, which is still in 980 00:37:46,640 --> 00:37:49,359 perspective not horrible. The Quen 3.8 8 981 00:37:49,359 --> 00:37:51,760 max test video that we did a few weeks 982 00:37:51,760 --> 00:37:54,000 ago or like a week or two ago that was 983 00:37:54,000 --> 00:37:56,640 about $32. So keep that in mind as well. 984 00:37:56,640 --> 00:37:58,560 That may be a bit more performant than 985 00:37:58,560 --> 00:38:01,040 this, but there is going to be a pricing 986 00:38:01,040 --> 00:38:03,359 consideration as well. And this was very 987 00:38:03,359 --> 00:38:05,760 very very cheap. It won't remain that 988 00:38:05,760 --> 00:38:08,160 cheap, but even if it gets five times as 989 00:38:08,160 --> 00:38:09,920 expensive, it's still not horrible. I 990 00:38:09,920 --> 00:38:11,359 suppose that's going to conclude our 991 00:38:11,359 --> 00:38:14,000 first look and test of the 0813 Deepseek 992 00:38:14,000 --> 00:38:16,560 V4 Pro model, the general availability 993 00:38:16,560 --> 00:38:18,560 version. And this is now out of preview. 994 00:38:18,560 --> 00:38:20,560 And that's just pretty exciting. 995 00:38:20,560 --> 00:38:22,720 Basically, now I think I'm going to test 996 00:38:22,720 --> 00:38:24,800 Grock 4.6 because that also came out 997 00:38:24,800 --> 00:38:26,960 today. And then I believe Friday Quen 998 00:38:26,960 --> 00:38:29,520 3.827B is coming out. So that'll be 999 00:38:29,520 --> 00:38:31,200 exciting as well. Seems like there's 1000 00:38:31,200 --> 00:38:32,960 been this has been like a pretty busy 1001 00:38:32,960 --> 00:38:34,720 week in terms of model releases. So 1002 00:38:34,720 --> 00:38:37,599 that's cool also. And when Musepark 1.2 1003 00:38:37,599 --> 00:38:39,359 comes out, that's going to be awesome. I 1004 00:38:39,359 --> 00:38:40,720 don't know how big it is, but I'm very 1005 00:38:40,720 --> 00:38:42,720 excited to try to run it locally. So 1006 00:38:42,720 --> 00:38:44,400 with that, if you have any questions, 1007 00:38:44,400 --> 00:38:45,760 please feel free to leave them in the 1008 00:38:45,760 --> 00:38:49,200 comments. And thanks for watching.