1 00:00:03,160 --> 00:00:06,960 Speaker 1: Media. Hello on, Welcome to this week's Better Offline. I'm 2 00:00:06,960 --> 00:00:21,080 Speaker 1: your host ed ze Tron. So yes, you're getting two 3 00:00:21,160 --> 00:00:23,200 Speaker 1: monologues this week because we had a guest pull out 4 00:00:23,280 --> 00:00:25,760 Speaker 1: last minute, leaving me with just my keyboard, my microphone, 5 00:00:25,760 --> 00:00:28,200 Speaker 1: and a diet coke in one of those weird vacuum 6 00:00:28,280 --> 00:00:32,360 Speaker 1: sealed tubes that really only real diet coke freaks on. 7 00:00:33,080 --> 00:00:35,320 Speaker 1: And today I'm going to start with a few questions. 8 00:00:36,000 --> 00:00:39,239 Speaker 1: What's going on in Ananthropic What is Dario Amide up to? 9 00:00:39,680 --> 00:00:42,360 Speaker 1: Why do things keep breaking? Used to be the coders 10 00:00:42,400 --> 00:00:44,720 Speaker 1: could just launch claud and run a bunch of subagents 11 00:00:44,720 --> 00:00:47,720 Speaker 1: to look productive. Think you can rely on claud code. 12 00:00:47,880 --> 00:00:51,400 Speaker 1: Guess again, Dario Amede doesn't care about you, and he 13 00:00:51,520 --> 00:00:55,720 Speaker 1: certainly doesn't care about your family. Anyway, to set the scene, 14 00:00:56,000 --> 00:00:57,800 Speaker 1: I need to give you a few details about claud 15 00:00:57,840 --> 00:01:01,080 Speaker 1: Code in general, which you don't pay for clawed code itself. 16 00:01:01,120 --> 00:01:04,480 Speaker 1: There's not like a CLAWD code subscription. When you subscribe 17 00:01:04,480 --> 00:01:08,200 Speaker 1: to Anthropic service, you pay for a CLAWD subscription and 18 00:01:08,240 --> 00:01:10,480 Speaker 1: you either pay twenty bucks, one hundred bucks, or two 19 00:01:10,560 --> 00:01:12,919 Speaker 1: hundred bucks a month, with each tier having more access 20 00:01:12,959 --> 00:01:15,440 Speaker 1: than the last, and access meaning right limits so you 21 00:01:15,440 --> 00:01:17,800 Speaker 1: can use it more. Though if you're wondering what that 22 00:01:17,880 --> 00:01:21,640 Speaker 1: actually means, that's really by design, as all Anthropic tells 23 00:01:21,680 --> 00:01:23,240 Speaker 1: you is that you've got a five hour limit and 24 00:01:23,280 --> 00:01:26,040 Speaker 1: a weekly limit on them models, with a separate one 25 00:01:26,040 --> 00:01:29,120 Speaker 1: for clawed Opus, which is their more expensive, more complex one. 26 00:01:29,920 --> 00:01:34,039 Speaker 1: An important detail is that these subscriptions are also heavily subsidized. 27 00:01:34,400 --> 00:01:36,920 Speaker 1: When you pay Anthropic two hundred dollars a month, you're 28 00:01:36,920 --> 00:01:39,080 Speaker 1: not paying on a per token rate, which is what 29 00:01:39,120 --> 00:01:41,319 Speaker 1: AI starts have to do. They pay per million input 30 00:01:41,360 --> 00:01:44,720 Speaker 1: and output tokens. No no, no, you just do stuff 31 00:01:44,800 --> 00:01:47,800 Speaker 1: and stuff comes out, and you're able to do the 32 00:01:47,800 --> 00:01:49,320 Speaker 1: same things you would I have used in the API. 33 00:01:49,360 --> 00:01:51,680 Speaker 1: But you're burning tokens, and you can burn over two 34 00:01:51,720 --> 00:01:54,640 Speaker 1: and a half thousand dollars in model tokens on a 35 00:01:54,680 --> 00:01:57,040 Speaker 1: two hundred buck a month subscription, or at least you could, 36 00:01:57,440 --> 00:01:59,640 Speaker 1: and some users have been able to spend upwards of 37 00:01:59,720 --> 00:02:03,560 Speaker 1: five thousand dollars on the same subscription rate. Limits are 38 00:02:03,600 --> 00:02:06,040 Speaker 1: generally used to restrict people from burning as much, but 39 00:02:06,040 --> 00:02:09,440 Speaker 1: the problem has become that for Anthropic to approach anything 40 00:02:09,480 --> 00:02:12,400 Speaker 1: close to profitability would have to rate limit people into 41 00:02:12,440 --> 00:02:15,400 Speaker 1: the well into the depths of Hell at this point. 42 00:02:15,880 --> 00:02:18,320 Speaker 1: Moving on. In the middle of March, Anthropics started a 43 00:02:18,360 --> 00:02:21,080 Speaker 1: two week long promotional campaign where they doubled rate limits 44 00:02:21,120 --> 00:02:23,880 Speaker 1: for off peak hours of using Claude, set to end 45 00:02:23,880 --> 00:02:26,600 Speaker 1: on March twenty seven to twenty twenty six, a day 46 00:02:26,600 --> 00:02:28,440 Speaker 1: before it was said to end on March twenty six 47 00:02:28,440 --> 00:02:30,679 Speaker 1: to twenty twenty six, and Tropic would announce it was 48 00:02:30,720 --> 00:02:33,359 Speaker 1: starting peak hours with claud code users maxing out their 49 00:02:33,360 --> 00:02:36,120 Speaker 1: sessions faster than between the hours of five am and 50 00:02:36,160 --> 00:02:40,200 Speaker 1: eleven pm Pacific Manday to Friday, with a spokesperson limply 51 00:02:40,200 --> 00:02:43,640 Speaker 1: adding that efficiency wins aren't named, of course, would offset 52 00:02:43,639 --> 00:02:46,640 Speaker 1: this and only seven percent of users would hit the limits. 53 00:02:47,040 --> 00:02:49,160 Speaker 1: All of this was sold as a result of managing 54 00:02:49,160 --> 00:02:52,600 Speaker 1: the growing demand for Claude. Yes, spoiler alert, more than 55 00:02:52,639 --> 00:02:54,680 Speaker 1: seven percent of users appear to have hit the limits 56 00:02:54,680 --> 00:02:57,679 Speaker 1: and nobody seems to be feeling particularly efficient. Don't know 57 00:02:57,720 --> 00:03:00,840 Speaker 1: where those gains are one use of one hundred dollars 58 00:03:00,840 --> 00:03:03,400 Speaker 1: a month. Max Plan complained about hitting sixty one percent 59 00:03:03,400 --> 00:03:06,200 Speaker 1: of his session limit the five hour one after four prompts, 60 00:03:06,240 --> 00:03:09,120 Speaker 1: which he found out based using a tool called cc 61 00:03:09,280 --> 00:03:12,680 Speaker 1: Usage cost ten dollars and twenty six cents in tokens, 62 00:03:12,720 --> 00:03:15,840 Speaker 1: he still spent ten percent of his subscription and four prompts. 63 00:03:16,400 --> 00:03:18,480 Speaker 1: Another said that they hit sixty three percent of the 64 00:03:18,600 --> 00:03:20,400 Speaker 1: limit on their two hundred a month plan in the 65 00:03:20,400 --> 00:03:22,600 Speaker 1: space of a day, and another hit ninety five percent 66 00:03:22,600 --> 00:03:24,480 Speaker 1: off to twenty minutes of using their max plan, going 67 00:03:24,520 --> 00:03:26,320 Speaker 1: to get one hundred dollars a month on that one. 68 00:03:26,600 --> 00:03:28,640 Speaker 1: Another person hit their max limit off to two or 69 00:03:28,639 --> 00:03:31,520 Speaker 1: three things, don't know what they were, and another vow 70 00:03:31,600 --> 00:03:33,760 Speaker 1: to cancel they're two hundred dollar a month subscription after 71 00:03:33,840 --> 00:03:35,480 Speaker 1: hitting their weekly limit in the space of a day, 72 00:03:35,560 --> 00:03:37,520 Speaker 1: saying that they and I'm going off of a translation 73 00:03:37,680 --> 00:03:41,640 Speaker 1: from fucking Grock, so forgive me expected the premium experience 74 00:03:41,640 --> 00:03:45,000 Speaker 1: between hundred dollars and what they got was constant limit stress. 75 00:03:45,920 --> 00:03:48,600 Speaker 1: I've linked a lot of these in yesterday's free newsletter. 76 00:03:48,640 --> 00:03:51,960 Speaker 1: The SUBPRIMEI crisis is here. Really advise you read it, 77 00:03:52,040 --> 00:03:54,640 Speaker 1: but also you can see how many people are mad, 78 00:03:54,800 --> 00:03:58,040 Speaker 1: or just go on Twitter and search claud limits. It's 79 00:03:58,080 --> 00:04:01,760 Speaker 1: it's not great now. While well, Anthropic technical staff member 80 00:04:01,840 --> 00:04:04,920 Speaker 1: Lydia Halle posted that Anthropic was aware of people hitting 81 00:04:04,960 --> 00:04:07,880 Speaker 1: usage limits in claud code way faster than expected, and 82 00:04:07,880 --> 00:04:10,760 Speaker 1: that some investigation of some sort was taking place. It's 83 00:04:10,800 --> 00:04:12,800 Speaker 1: hard to imagine that Anthropic had no idea that these 84 00:04:12,800 --> 00:04:15,040 Speaker 1: limits were so severe, or that any of this was 85 00:04:15,040 --> 00:04:19,600 Speaker 1: a surprise. Now as I wrote this sentence that I'm 86 00:04:19,640 --> 00:04:21,839 Speaker 1: reading on Tuesday, March thirty first, it doesn't appear that 87 00:04:21,880 --> 00:04:24,560 Speaker 1: any changes have been made. People are still complaining about 88 00:04:24,600 --> 00:04:26,880 Speaker 1: hitting their limits in a few prompts, and Anthropic is 89 00:04:26,960 --> 00:04:29,320 Speaker 1: yet to update anyone. In my opinion, because these are 90 00:04:29,320 --> 00:04:31,599 Speaker 1: the rate limits they decided were necessary to keep the 91 00:04:31,600 --> 00:04:36,159 Speaker 1: business going and roll their nasty ass into IPO. A 92 00:04:36,160 --> 00:04:40,280 Speaker 1: few days previously, though, Anthropic could also accidentally, and I 93 00:04:40,320 --> 00:04:42,720 Speaker 1: put that in quotation, Marks leaked the existence of their 94 00:04:42,760 --> 00:04:46,239 Speaker 1: Cappy Boro and Mythos models to Fortune magazine, by which 95 00:04:46,279 --> 00:04:48,719 Speaker 1: I mean they had a data cash to quote Fortune 96 00:04:48,760 --> 00:04:51,400 Speaker 1: with over three thousand assets that was left open on 97 00:04:51,440 --> 00:04:55,400 Speaker 1: the internet and Fortune somehow found it. You know, Kanna 98 00:04:55,440 --> 00:04:57,760 Speaker 1: reminds me of like a skeezy bloke dropping a Magnum 99 00:04:57,760 --> 00:04:59,320 Speaker 1: condom out of his wallet in front of a woman 100 00:04:59,320 --> 00:05:03,520 Speaker 1: and being like, hey, yeah, you see that, or whoops, whoops. 101 00:05:04,120 --> 00:05:07,160 Speaker 1: In any case, this massive leak also included absolutely fucking 102 00:05:07,200 --> 00:05:10,120 Speaker 1: nothing about the models themselves, other than that they're a 103 00:05:10,240 --> 00:05:14,039 Speaker 1: step change better and that their cybersecurity features were so 104 00:05:14,200 --> 00:05:16,760 Speaker 1: very scary that they would have to roll them out slowly. 105 00:05:17,360 --> 00:05:21,040 Speaker 1: I'm I don't know, man, I no insult to the 106 00:05:21,040 --> 00:05:23,240 Speaker 1: people at Fortune. I just don't think they're on fucking 107 00:05:23,320 --> 00:05:28,080 Speaker 1: Showdan looking up aws buckets. So wonder how this got there? Now? 108 00:05:28,320 --> 00:05:30,680 Speaker 1: I will say, at first I was completely sure that 109 00:05:30,720 --> 00:05:33,680 Speaker 1: this was a deliberate leak, because three thousand assets left 110 00:05:33,720 --> 00:05:35,440 Speaker 1: open on the Internet and none of them have any 111 00:05:35,480 --> 00:05:38,360 Speaker 1: info on the model, but just scary things like ooh 112 00:05:38,360 --> 00:05:41,440 Speaker 1: with too much better. But now I'm not so sure, 113 00:05:41,520 --> 00:05:44,880 Speaker 1: because yesterday Anthropic accidentally leaked the entire source code of 114 00:05:44,880 --> 00:05:47,960 Speaker 1: its claud code coding interface. The leak was a result 115 00:05:47,960 --> 00:05:50,800 Speaker 1: of a reference in claud Code's MPM package, the thing 116 00:05:50,839 --> 00:05:53,440 Speaker 1: that you queried download clord code onto your computer via 117 00:05:53,440 --> 00:05:55,800 Speaker 1: the terminal that led right back to a ZIP archive 118 00:05:55,839 --> 00:05:59,239 Speaker 1: on Anthropic servers per the register that contained the source code. 119 00:06:00,040 --> 00:06:01,960 Speaker 1: I will add that there's an ongoing discussion about what 120 00:06:02,000 --> 00:06:05,520 Speaker 1: actually caused this problem with someone hacker news, hacker news, 121 00:06:05,760 --> 00:06:08,159 Speaker 1: how can use I'm just going to keep it saying 122 00:06:08,160 --> 00:06:10,920 Speaker 1: it might be a problem with bun, the packaging tool 123 00:06:11,040 --> 00:06:14,279 Speaker 1: used to allow people to download clawed code that Anthropic 124 00:06:14,320 --> 00:06:17,760 Speaker 1: acquired in December of last year. To be clear, this 125 00:06:17,800 --> 00:06:20,640 Speaker 1: isn't a leak of Anthropics models, but it's still an 126 00:06:20,720 --> 00:06:23,919 Speaker 1: unbelievably large leak, one that exposed claud code's innards to 127 00:06:23,960 --> 00:06:26,840 Speaker 1: the entire Internet and all of their competitors. And while 128 00:06:26,839 --> 00:06:29,640 Speaker 1: I imagine using the source code is illegal on some level, 129 00:06:29,920 --> 00:06:32,480 Speaker 1: I can't imagine there's any reason their competitors can't take 130 00:06:32,480 --> 00:06:34,320 Speaker 1: a look or that they're you think that they're all 131 00:06:34,360 --> 00:06:37,240 Speaker 1: sitting around being like, oh, I absolutely can't, I mustn't. 132 00:06:37,520 --> 00:06:40,240 Speaker 1: It's not okay, especially when this is a company that 133 00:06:40,400 --> 00:06:43,760 Speaker 1: fucked over just about anyone building anything that you build 134 00:06:43,800 --> 00:06:46,280 Speaker 1: on top of a claud code subscription, thinking about how 135 00:06:46,279 --> 00:06:49,560 Speaker 1: they treated open code. By the way, anyway, now is 136 00:06:49,600 --> 00:06:51,919 Speaker 1: a great time to remind you that claud code create 137 00:06:51,960 --> 00:06:54,440 Speaker 1: a Boris Churney said at the end of December that 138 00:06:54,480 --> 00:06:57,040 Speaker 1: one hundred percent of his contributions to clawd code were 139 00:06:57,040 --> 00:07:00,520 Speaker 1: written in claud code, and told Lenny Richitski in February 140 00:07:00,520 --> 00:07:03,400 Speaker 1: twenty twenty six the coding is now solved for most 141 00:07:03,520 --> 00:07:06,919 Speaker 1: use cases. I assume that the use cases that haven't 142 00:07:06,920 --> 00:07:10,080 Speaker 1: been solved include making sure that there isn't a direct 143 00:07:10,080 --> 00:07:12,640 Speaker 1: link from the claud code installer to a cloud Flare 144 00:07:12,840 --> 00:07:15,840 Speaker 1: storage bucket with its source code ready to download. That 145 00:07:15,880 --> 00:07:18,680 Speaker 1: one just doesn't seem like they got that one pinned down. 146 00:07:19,080 --> 00:07:21,240 Speaker 1: I've also had that. The same interview added that Cherney 147 00:07:21,240 --> 00:07:23,720 Speaker 1: hadn't written a single line of code since November twenty 148 00:07:23,760 --> 00:07:27,640 Speaker 1: twenty five, which is I mean, those of you playing 149 00:07:27,680 --> 00:07:31,640 Speaker 1: at home, I know many of you really enjoy my 150 00:07:31,760 --> 00:07:34,360 Speaker 1: game show is that good? And I just have to 151 00:07:34,400 --> 00:07:39,040 Speaker 1: ask you is that good? Anyway? I feel like I 152 00:07:39,200 --> 00:07:42,040 Speaker 1: very recently warned everybody about the very obvious dangers of 153 00:07:42,080 --> 00:07:44,960 Speaker 1: allowing lms to write all of your code. These models 154 00:07:44,960 --> 00:07:47,200 Speaker 1: do not have thoughts or knowledge, or really anything other 155 00:07:47,240 --> 00:07:50,600 Speaker 1: than probabilistic generations of outputs based on training data and 156 00:07:50,640 --> 00:07:53,840 Speaker 1: a lot of it, and they're quite complex. Nevertheless, this 157 00:07:53,960 --> 00:07:56,440 Speaker 1: means that any time you choose to just accept the 158 00:07:56,480 --> 00:07:59,640 Speaker 1: code they generate without reading it thoroughly, you're choosing to 159 00:07:59,640 --> 00:08:03,480 Speaker 1: trust thing inherently untrusted, whether it doesn't think or have knowledge. 160 00:08:03,640 --> 00:08:05,680 Speaker 1: This is just the beginning of us finding out the 161 00:08:05,800 --> 00:08:09,480 Speaker 1: ugly cost of software engineers trusting large language models to 162 00:08:09,520 --> 00:08:13,200 Speaker 1: write their code at scale. Llms are good at writing 163 00:08:13,280 --> 00:08:15,520 Speaker 1: lots of code, which in turn means that the code 164 00:08:15,600 --> 00:08:19,360 Speaker 1: requires far more time to review, which assumes you do 165 00:08:19,480 --> 00:08:21,880 Speaker 1: review it, which becomes far more difficult when you're constantly 166 00:08:21,880 --> 00:08:24,120 Speaker 1: pressured to ship more and more and more software every 167 00:08:24,160 --> 00:08:26,520 Speaker 1: fucking week. Or if you're Boris Journey and you believe 168 00:08:26,560 --> 00:08:28,800 Speaker 1: that shipping software fast is the same thing is shipping 169 00:08:28,800 --> 00:08:32,719 Speaker 1: good software. I'm now thinking about the launch of anthropics 170 00:08:32,800 --> 00:08:35,840 Speaker 1: claud cowork, a product that can allegedly manage files on 171 00:08:35,880 --> 00:08:38,600 Speaker 1: your computer or draft documents or some such Billshere nobody 172 00:08:38,640 --> 00:08:40,920 Speaker 1: seems to be able to explain it. At the time, 173 00:08:41,080 --> 00:08:44,000 Speaker 1: Boris Journey proudly boasted that clawed cowork was built in 174 00:08:44,040 --> 00:08:47,440 Speaker 1: around a week and a half, which makes perfect sense, says. 175 00:08:47,640 --> 00:08:50,040 Speaker 1: Almost immediately, a story came out of a guy who 176 00:08:50,120 --> 00:08:53,280 Speaker 1: tried to reorganize his wife's desktop files using cowork, which 177 00:08:53,280 --> 00:08:56,280 Speaker 1: is one of the literal things on the cowork website. 178 00:08:56,280 --> 00:08:58,440 Speaker 1: It tells you to do, only to watch it delete. 179 00:08:58,520 --> 00:09:01,280 Speaker 1: Every single one of his folk goes permanently. He was 180 00:09:01,320 --> 00:09:03,640 Speaker 1: only able to recover them thanks to a nicloud backup. 181 00:09:04,679 --> 00:09:08,120 Speaker 1: This is the future that these so called large language 182 00:09:08,160 --> 00:09:11,800 Speaker 1: model companies want for you. Bad software chipped quickly hyped 183 00:09:11,800 --> 00:09:13,800 Speaker 1: by a captured media, doesn't give a damn about whether 184 00:09:13,800 --> 00:09:15,679 Speaker 1: the services are the things you pay for or functional 185 00:09:15,760 --> 00:09:18,120 Speaker 1: or useful, and doesn't even bother using the tools or 186 00:09:18,200 --> 00:09:22,120 Speaker 1: understanding them. And nobody else appears to be discussing how 187 00:09:22,160 --> 00:09:26,360 Speaker 1: inherently deceptive these companies have. Become a subscriber to Claude 188 00:09:26,360 --> 00:09:28,920 Speaker 1: who paid for an annual subscription in December twenty twenty 189 00:09:28,920 --> 00:09:31,600 Speaker 1: five now as a product that has significantly less value 190 00:09:31,720 --> 00:09:34,800 Speaker 1: thanks to egregious rate limits. And that's also like if 191 00:09:34,840 --> 00:09:37,199 Speaker 1: you can even do your job anymore, I imagine people 192 00:09:37,200 --> 00:09:39,880 Speaker 1: in the twenty buck a month subscription are really fucking suffering. 193 00:09:41,200 --> 00:09:44,000 Speaker 1: And in general, anthropics models seem to go and look 194 00:09:44,000 --> 00:09:46,360 Speaker 1: at the Reddit if you want to understand more. They 195 00:09:46,360 --> 00:09:48,880 Speaker 1: seem to oscillate in utility and effixi based on the 196 00:09:48,920 --> 00:09:51,440 Speaker 1: time of day you use them and their proximity to 197 00:09:51,520 --> 00:09:54,840 Speaker 1: a new model launch. This is a really weird story 198 00:09:54,880 --> 00:09:56,560 Speaker 1: that I've never been able to get to the bottom of, 199 00:09:56,800 --> 00:09:59,280 Speaker 1: but there have been reports for like over a year 200 00:09:59,440 --> 00:10:02,840 Speaker 1: of different mode from Anthropic from OpenAI from whomever getting 201 00:10:02,960 --> 00:10:06,160 Speaker 1: dumber at random times. No one's able to get to 202 00:10:06,200 --> 00:10:08,400 Speaker 1: the bottom of it. I've heard people saying they quantize 203 00:10:08,400 --> 00:10:10,960 Speaker 1: the models make them smaller during the day. I don't 204 00:10:10,960 --> 00:10:14,000 Speaker 1: know if you have any information about this, shoot me 205 00:10:14,040 --> 00:10:16,600 Speaker 1: at shoot me a piss on clerk, or email me 206 00:10:16,679 --> 00:10:19,840 Speaker 1: a easy at better offline dot com. That's Echo Zeta 207 00:10:20,040 --> 00:10:23,680 Speaker 1: at better offline dot com. What we do know is 208 00:10:23,720 --> 00:10:26,840 Speaker 1: that claud Code, one of the few popular AI products, 209 00:10:27,040 --> 00:10:29,960 Speaker 1: is built with the same disregard for safety and customer 210 00:10:30,000 --> 00:10:33,680 Speaker 1: happiness as the rest of Anthropics, astonishingly shitty business that 211 00:10:33,760 --> 00:10:37,199 Speaker 1: burns billions of dollars with no end in sight. Chief 212 00:10:37,200 --> 00:10:40,360 Speaker 1: claud Code slopagandist Boris Cherney doesn't give enough of a 213 00:10:40,400 --> 00:10:42,760 Speaker 1: fuck about his customers to actually read his code, and 214 00:10:42,800 --> 00:10:44,480 Speaker 1: I imagine the same goes for a lot of other 215 00:10:44,520 --> 00:10:48,240 Speaker 1: engineers within the company a big tech at large. The 216 00:10:48,320 --> 00:10:51,319 Speaker 1: consequences are already obvious. In the last few months, A 217 00:10:51,600 --> 00:10:54,920 Speaker 1: coding tools brought down Aws twice and lost Amazon hundreds 218 00:10:54,960 --> 00:10:57,480 Speaker 1: of thousands of orders and led to a security breach 219 00:10:57,480 --> 00:11:00,840 Speaker 1: inside Meta mere weeks ago. And now how these tools 220 00:11:00,840 --> 00:11:04,760 Speaker 1: of leaked claud code source code. How much is enough 221 00:11:04,800 --> 00:11:06,880 Speaker 1: to make people wake up to the inherent dangers of 222 00:11:06,960 --> 00:11:10,240 Speaker 1: using these models the right software at scale. I guess 223 00:11:10,280 --> 00:11:12,960 Speaker 1: we're gonna find out. I'll see you on Friday for 224 00:11:13,000 --> 00:11:14,600 Speaker 1: another monologue.