WEBVTT - How To Argue With An AI Booster, Part One 0:00:02.800 --> 0:00:07.480 Zone Media. Welcome to Better Offline. I'm your host ed 0:00:07.560 --> 0:00:22.079 zitrun as Eva. Check out the episode notes for merchandise, 0:00:22.200 --> 0:00:25.680 my newsletter, various links, things that I just write for myself. 0:00:25.880 --> 0:00:28.080 I don't actually leave little notes, but maybe I should. 0:00:28.560 --> 0:00:31.000 But as promised in my last monologue, we've got a 0:00:31.000 --> 0:00:34.040 three part episode this week, and we're talking about AI boosters, 0:00:34.080 --> 0:00:36.800 the people who insist despite very little proof and all 0:00:36.840 --> 0:00:39.680 evidence of the contrary. The generative AI is the future, 0:00:39.960 --> 0:00:42.800 open AI and anthropic of perfectly healthy companies. And what 0:00:42.840 --> 0:00:45.760 we're witnessing isn't an inflation of a massive, dangerous bubble 0:00:45.800 --> 0:00:48.800 that might eat our entire economy, but rather the emergence 0:00:48.800 --> 0:00:51.040 of a brand new paradigm in tech that's just as 0:00:51.080 --> 0:00:55.000 meaningful as the smartphone and cloud computing revolutions. I even 0:00:55.120 --> 0:00:57.600 saying that out loud makes me feel a little crazy, 0:00:58.080 --> 0:01:00.560 but yeah, that's what they believe. These people exist. They're 0:01:00.560 --> 0:01:03.680 often wrong. In fact, they're mostly wrong, and they're also 0:01:03.760 --> 0:01:06.800 really really annoying. And so as a public service, I'm 0:01:06.800 --> 0:01:09.080 going to spend the next three episodes explaining how to 0:01:09.120 --> 0:01:11.640 talk with them without losing your sanity, will to live, 0:01:11.720 --> 0:01:14.840 or your ability to operate in society, and now what 0:01:15.000 --> 0:01:17.240 makes me qualified to do this well? In the last 0:01:17.319 --> 0:01:19.880 two years, I've written no less than half a million words, 0:01:19.920 --> 0:01:22.560 with many of them dedicated to breaking both existent and 0:01:22.600 --> 0:01:25.160 previous myths about the state of technology and the tech 0:01:25.200 --> 0:01:29.400 industry itself. Now let's start with something every critic is experienced, 0:01:29.480 --> 0:01:32.680 the massive double standard between those perceived as skeptics and 0:01:32.720 --> 0:01:36.480 those who are optimistic or optimists. To be skeptical of 0:01:36.520 --> 0:01:39.120 ais to commit yourself to a near constant amount of 0:01:39.160 --> 0:01:42.000 demands to prove yourself and endless nags of But what 0:01:42.200 --> 0:01:44.920 about with each one, no matter how small, presented as 0:01:44.959 --> 0:01:47.319 a fact that defeats any points you may have had, 0:01:47.800 --> 0:01:50.920 have or will have in the future. Conversely, being an 0:01:50.920 --> 0:01:53.760 optimist allows you to take things like AI twenty twenty seven, 0:01:53.800 --> 0:01:56.360 which I will fucking get to seriously, to the point 0:01:56.360 --> 0:01:59.080 that you can write an entire feature length fan fiction 0:01:59.200 --> 0:02:01.600 piece in the New York Times and nobody will bat 0:02:01.600 --> 0:02:04.760 an eye lead. In any case, things are beginning to 0:02:04.760 --> 0:02:07.800 fall UPALM two of the actual reporters, the real ones 0:02:07.840 --> 0:02:10.120 at the New York Times rather than column nests, suit 0:02:10.520 --> 0:02:13.760 kind of act like adult babies. Reported in late August 0:02:13.800 --> 0:02:16.560 that Meta is yet again restructuring its AI department for 0:02:16.600 --> 0:02:19.880 the fourth time, and that it's considering downsizing their overall 0:02:19.919 --> 0:02:22.760 AI division, which sure doesn't seem like something you'd do 0:02:22.840 --> 0:02:25.640 if you thought AI was the future. Meanwhile, the markets 0:02:25.639 --> 0:02:28.680 are thoroughly spooked by an MIT study covered by Fortune 0:02:28.720 --> 0:02:31.440 that found that ninety five percent of generative AI pilots 0:02:31.440 --> 0:02:34.480 and companies are failing in providing no ROI. And though 0:02:34.600 --> 0:02:37.000 MIT Nanda has now replaced the link to the study 0:02:37.280 --> 0:02:40.160 with some sort of Google form to request access, you 0:02:40.200 --> 0:02:42.639 can find the full pdf linked in the show notes. 0:02:42.840 --> 0:02:44.120 By the way, this is the kind of thing that 0:02:44.280 --> 0:02:46.560 is a PR firm wanting to try and set up interviews, 0:02:46.600 --> 0:02:48.520 not for me, thank you, just like to read the 0:02:48.520 --> 0:02:51.280 thing that you put out very rude. In any case, 0:02:51.320 --> 0:02:53.800 the report is actually grimmer than Fortune made its sound, 0:02:53.880 --> 0:02:56.519 saying that ninety five percent of organizations are getting zero 0:02:56.600 --> 0:02:59.880 return on GENERATIVAI. The report says that adoption is high, 0:03:00.080 --> 0:03:03.160 transformation is low, adding that few industries show the deep 0:03:03.200 --> 0:03:06.880 structural shifts associated with past general purpose technologies, such as 0:03:06.960 --> 0:03:10.440 new market leaders, disrupted business models, or measurable changes in 0:03:10.480 --> 0:03:13.200 customer behavior, and saying that out loud. That is right, 0:03:13.280 --> 0:03:17.040 what is being disrupted by AI? Search? Kind of that 0:03:17.120 --> 0:03:20.280 just means that search sucks anyway. Yet the most damning 0:03:20.360 --> 0:03:22.959 part was there was a part called the five myths 0:03:22.960 --> 0:03:26.160 about AI in the Enterprise, which is probably the most 0:03:26.200 --> 0:03:28.960 wilting takedown of this movement I've ever seen. And I'm 0:03:28.960 --> 0:03:31.840 going to quote an adverbatum okay number one. And they 0:03:32.000 --> 0:03:35.240 present these as the first statement is something that isn't 0:03:35.280 --> 0:03:38.160 going to happen, or they're questioning AI will replace most 0:03:38.240 --> 0:03:40.760 jobs in the next few years. What it actually says 0:03:40.840 --> 0:03:43.720 is research found limited layoffs from JENAI, and only in 0:03:43.720 --> 0:03:47.280 industries that are already affected significantly by AI. There is 0:03:47.320 --> 0:03:51.080 no consensus among executives as the hiring levels over the 0:03:51.160 --> 0:03:54.560 next three to five years. Another statement they challenge generative 0:03:54.600 --> 0:03:58.000 AI is transforming business. No, no, No, adoption is high, but 0:03:58.040 --> 0:04:01.280 transformation is rare. Only five percent of enterprises have AI 0:04:01.360 --> 0:04:04.880 tools integrated into workflows at scale, and seven of nine 0:04:04.920 --> 0:04:07.560 sectors show no real structural change. And I want to 0:04:07.600 --> 0:04:09.760 thank them for putting this there. I made this exact 0:04:09.800 --> 0:04:12.480 point in February in my newsletter. There's no AI revolution. 0:04:12.880 --> 0:04:16.000 Links in the show notes as ever, enterprises are slow 0:04:16.040 --> 0:04:18.720 in adopting new tech is another commonly held thing, and 0:04:18.800 --> 0:04:22.120 they say enterprises are extremely eager to adopt AI, and 0:04:22.200 --> 0:04:26.159 ninety percent of seriously explored buying an AI solution. Another 0:04:26.200 --> 0:04:28.839 statement they challenge, the biggest thing holding back AI is 0:04:28.880 --> 0:04:33.520 model quality, legal data, and risk. And then they say, 0:04:33.560 --> 0:04:37.280 which is my favorite quot, what's really holding it back 0:04:37.360 --> 0:04:39.600 is that most AI tools don't learn and don't integrate 0:04:39.600 --> 0:04:43.800 well into workflows. I really do love that. I love 0:04:43.839 --> 0:04:46.400 that so much because it's like the thing holding AI 0:04:46.480 --> 0:04:49.320 back is that it sucks. And the final one, the 0:04:49.320 --> 0:04:52.120 final challenge statement they make is the best enterprises are 0:04:52.120 --> 0:04:55.479 building their own tools. Internal builds fail twice as often, 0:04:56.320 --> 0:04:59.239 and that's twice as often in a thing where ninety 0:04:59.240 --> 0:05:03.760 five percent anyway, These are brutal, dispassionate points that directly 0:05:03.760 --> 0:05:07.719 deal with the most common boosterisms. Generative AI isn't transforming anything. 0:05:07.800 --> 0:05:11.520 AI isn't replacing anyone. Enterprises are trying to adopt generative AI, 0:05:11.760 --> 0:05:13.920 but it doesn't fucking work. And the thing holding back 0:05:13.960 --> 0:05:16.560 AI is the fact it doesn't fucking work. This isn't 0:05:16.560 --> 0:05:18.880 a case where the enterprise is suddenly going to save 0:05:18.960 --> 0:05:21.840 these companies because the enterprise has already tried and it 0:05:21.920 --> 0:05:24.520 is not working. And what's hilarious as well is so 0:05:24.560 --> 0:05:27.640 many people say, well, there's just this enterprise opportunity they 0:05:27.640 --> 0:05:31.120 haven't got into yet. Actually they have, They've got into 0:05:31.160 --> 0:05:34.240 the enterprise. They're sticking their fingers and all the bits 0:05:34.279 --> 0:05:36.640 of the sas and it isn't working. And the enterprise 0:05:36.720 --> 0:05:39.000 loves money. The enterprise will do even if it's a 0:05:39.279 --> 0:05:42.799 genuinely evil product. They will push it, Like look at Salesforce, 0:05:42.839 --> 0:05:46.640 an entire company based on them. Anyway, an incorrect read 0:05:46.680 --> 0:05:49.400 of the study that's been going around is that the 0:05:49.480 --> 0:05:53.000 learning gap that makes these things less useful is because 0:05:53.000 --> 0:05:56.239 of human issues like human error, when the study actually 0:05:56.279 --> 0:05:59.279 says that the fundamental gap that defines the GENAI divide 0:05:59.320 --> 0:06:02.680 is that users tools that don't adapt model quality fills 0:06:02.680 --> 0:06:05.800 without context, and UX suffers when systems can't remember this 0:06:05.920 --> 0:06:08.440 isn't something you as a person learn your way out of. 0:06:08.480 --> 0:06:10.440 These products don't do what they're meant to do, and 0:06:10.480 --> 0:06:13.880 people are realizing it when they try and use them. Nevertheless, 0:06:13.880 --> 0:06:15.880 boosters will still find a way to twist this study 0:06:15.880 --> 0:06:18.360 to mean something else. They'll claim that AI is still early, 0:06:18.480 --> 0:06:21.000 that the opportunity is still there, that we didn't confirm 0:06:21.040 --> 0:06:25.400 that the Internet and smartphones with productivity boosting, or that 0:06:25.440 --> 0:06:28.119 we're in the early days of AI somehow three years 0:06:28.160 --> 0:06:31.279 and hundreds of billions of dollars and thousands of articles 0:06:31.279 --> 0:06:33.640 in Yeah, we're in the early days. We're in the 0:06:33.640 --> 0:06:35.600 early days when all the money in all the land 0:06:35.600 --> 0:06:38.560 has gone into this. And I'm tired. I'm tired of 0:06:38.600 --> 0:06:41.160 having the same arguments with these people, and frankly, I'm 0:06:41.200 --> 0:06:44.120 sure you are too. No matter how blindly obvious evidence 0:06:44.920 --> 0:06:47.279 is to the contrary, they will find ways to ignore it. 0:06:47.400 --> 0:06:50.440 They continually make these smug comments about people wishing things 0:06:50.480 --> 0:06:53.400 would be bad or suggesting you are stupid, and yes, 0:06:53.440 --> 0:06:55.560 that is their belief, by the way, for not believing 0:06:55.600 --> 0:06:58.920 that generative AI is disruptive today. And in this three 0:06:59.000 --> 0:07:00.960 part tern I'm going to give you the tools to 0:07:01.000 --> 0:07:03.599 fight back against the AI boosters in your life. I'm 0:07:03.640 --> 0:07:06.280 going to go into the generalities of the booster movement, 0:07:06.600 --> 0:07:08.600 the way they argue, the tropes they cling to, and 0:07:08.640 --> 0:07:10.560 the ways in which they use your own self doubt 0:07:10.560 --> 0:07:13.160 against you. They're your buddy, your boss, a man in 0:07:13.200 --> 0:07:16.200 a Ginghamshire Epic steakhouse who won't leave you the fuck alone. 0:07:16.320 --> 0:07:19.520 A editor, a writer, a founderer, or just a common 0:07:19.640 --> 0:07:22.680 or garden conn artist. Whoever the booster in your life is, 0:07:22.880 --> 0:07:24.960 I want you to have the words to fight them with. 0:07:35.200 --> 0:07:37.720 So an AI booster is not, in many cases, an 0:07:37.760 --> 0:07:42.120 actual fan of artificial intelligence. People like Simon Willison or 0:07:42.160 --> 0:07:44.480 Max Wolf who actually work with LMS on a daily 0:07:44.520 --> 0:07:47.920 basis don't see the need to repeatedly harass everybody or 0:07:47.960 --> 0:07:50.360 talk down to them about their unwillingness to pledge allegiance 0:07:50.400 --> 0:07:53.440 to the graveyard smash of generative AI. In fact, the 0:07:53.480 --> 0:07:56.400 closer I've found someone to actually building things with LLMS, 0:07:56.560 --> 0:07:59.000 the less likely they are to emphatically argue that I'm 0:07:59.040 --> 0:08:03.320 missing out on something by not doing so myself. No, no, no. 0:08:03.400 --> 0:08:07.080 The AI booster is symbolically aligned with generative AI. They're 0:08:07.080 --> 0:08:08.960 fans in the same way that somebody is a fan 0:08:09.000 --> 0:08:11.400 of a sports team. Their houses in blazons with every 0:08:11.440 --> 0:08:14.520 possible piece of tat and crap they can find. They're Sundays, 0:08:14.520 --> 0:08:17.040 living and dying by the successes of the team. Except 0:08:17.080 --> 0:08:19.520 even fans of the Dallas Cowboys of a tighter grasp 0:08:19.560 --> 0:08:24.040 on reality, but not Micah Parsons anyway. Kevin Rus and 0:08:24.120 --> 0:08:26.480 Casey Newton are two of the most notable boosters, and 0:08:26.520 --> 0:08:29.680 as I'll get to later in this series, neither of 0:08:29.720 --> 0:08:32.560 them have a consistent or comprehensive knowledge of AI, despite 0:08:32.679 --> 0:08:35.520 being at The New York Times. Though Casey Newton is 0:08:35.520 --> 0:08:38.559 a contractor, he is a contractor just for a podcast, 0:08:38.600 --> 0:08:43.880 which I can't insult due to my own contractual relationships. Nevertheless, 0:08:43.920 --> 0:08:46.560 they will insist that everybody is using AI for everything, 0:08:46.600 --> 0:08:48.600 which is the title of an article they put out, 0:08:48.800 --> 0:08:51.280 a statement that even a booster should realize is incorrect 0:08:51.280 --> 0:08:54.440 based on the actual abilities of the models. But that's 0:08:54.480 --> 0:08:57.480 because it isn't about what's happening. It's not about what's 0:08:57.520 --> 0:09:02.240 actually happening. It's about allegiance. AI symbolizes something to the 0:09:02.280 --> 0:09:05.080 AI booster, a way that they're better than other people. 0:09:05.120 --> 0:09:08.719 That makes them superior because they're, unlike cynics and skeptics, 0:09:08.760 --> 0:09:12.079 able to see the incredible potential in the future of AI, 0:09:12.200 --> 0:09:14.559 but also how great it is today, though it never 0:09:14.600 --> 0:09:16.760 seemed to be able to explain why it is other 0:09:16.800 --> 0:09:19.040 than it replaced search for me and I use it 0:09:19.040 --> 0:09:21.720 to draw connections between articles I write, which is something 0:09:21.800 --> 0:09:24.400 I do for free without AI with my fucking brain. 0:09:25.240 --> 0:09:28.560 Boosterism is a kind of religion interested in finding symbolic 0:09:28.640 --> 0:09:31.680 proof that things are getting better in some indeterminate way, 0:09:31.880 --> 0:09:34.280 and that anyone that chooses to believe otherwise is ignorant 0:09:34.320 --> 0:09:38.080 or stupid or I actually don't know what it is 0:09:38.120 --> 0:09:40.240 that they're meant to be missing. Let me give you 0:09:40.240 --> 0:09:44.080 an example. Thomas Potastic. He wrote a piece called my 0:09:44.160 --> 0:09:46.840 ais skeptic friends are all nuts, and it was catnip 0:09:46.920 --> 0:09:50.120 for boosters. A software engineer using technical terms like interact 0:09:50.200 --> 0:09:52.880 with GIT and MCP, vague chants, and of course an 0:09:52.880 --> 0:09:56.000 extremely vague statement that says hallucinations aren't a problem, and 0:09:56.040 --> 0:09:58.760 I quote now, I'm sure there are still environments where 0:09:58.800 --> 0:10:02.280 hallucination matters. But hallucination is the first thing developers bring 0:10:02.360 --> 0:10:05.240 up when somebody suggests using llms, despite it being more 0:10:05.360 --> 0:10:11.000 or less a solved problem. Is it anyway? My favorite 0:10:11.000 --> 0:10:13.520 favorite part though, let me quote this. A lot of 0:10:13.640 --> 0:10:17.720 LLM skepticism probably isn't really about llms. It's projection. People 0:10:17.720 --> 0:10:20.080 say llms can't code, when what they really mean is 0:10:20.280 --> 0:10:22.400 llms can't write Rust which by the way, is a 0:10:22.440 --> 0:10:26.360 coding language. Fair enough, but people select languages in part 0:10:26.400 --> 0:10:29.000 based on how well lms work with them, So RUSS 0:10:29.040 --> 0:10:33.360 people should get on that. What nobody projects more than 0:10:33.360 --> 0:10:36.000 an AI booster. They thrive on the sense that they're 0:10:36.000 --> 0:10:39.600 oppressed and villainized after years of seemingly every other, every 0:10:39.640 --> 0:10:42.800 goddamn outlet on Earth claiming they're right, regardless of whether 0:10:42.840 --> 0:10:46.199 there's any proof. They sneer and jeer and cry constantly 0:10:46.200 --> 0:10:48.720 the people not showing adequate amounts of awe when an 0:10:48.760 --> 0:10:51.320 AI lab says we did something in private, we can't 0:10:51.320 --> 0:10:54.079 share it with you, but it's so cool, and constantly 0:10:54.120 --> 0:10:57.079 act as if they're victims as they spread outright misinformation, 0:10:57.440 --> 0:10:59.920 either through getting things wrong or never really caring enough 0:10:59.920 --> 0:11:02.760 to check if they're right. Also, none of the booster 0:11:02.880 --> 0:11:06.600 arguments actually survive a thorough response, as Nick Saresh proved 0:11:06.600 --> 0:11:10.439 with his hilarious and brutal takedown of Potastics. Peachs Siesh 0:11:10.520 --> 0:11:12.320 is a great guy. He's been on the show before, 0:11:12.480 --> 0:11:13.960 and I've linked to his piece in the show notes, 0:11:14.000 --> 0:11:15.640 and I'm going to bring him back on. He's written 0:11:15.640 --> 0:11:18.600 for my newsletter as well. Absolute legend now there are 0:11:18.720 --> 0:11:21.600 I believe some people who truly do love using llms, 0:11:21.679 --> 0:11:24.240 yet they are not the ones defending them. But Tastic's 0:11:24.320 --> 0:11:26.920 peace strips with condescension to the point that I feel 0:11:26.920 --> 0:11:29.520 like he's trying to convince himself how good lms are. 0:11:29.559 --> 0:11:33.319 Because boosters are eternal victims. He wrote them a piece 0:11:33.360 --> 0:11:35.920 that they could send around to skeptics saying hh see 0:11:36.040 --> 0:11:38.160 without being able to explain why it was such a 0:11:38.160 --> 0:11:41.520 brutal takedown, mostly because they can't express why other than well, 0:11:42.120 --> 0:11:45.400 this guy gets it. One cannot be a big, smart 0:11:45.400 --> 0:11:47.880 genius that understands the glory and power of AI while 0:11:47.960 --> 0:11:51.080 also acting like a scared little puppy every time somebody 0:11:51.120 --> 0:11:54.920 tells them it sucks. You know what, This is a 0:11:54.960 --> 0:11:57.120 great place to start. This is a great place to 0:11:57.200 --> 0:12:00.959 get into how to deal with AI boosters because AI 0:12:01.080 --> 0:12:04.160 boosters love being victims, and you should not play into it. 0:12:05.559 --> 0:12:07.800 When you speak to an AI booster, you may get 0:12:07.840 --> 0:12:10.120 the instinct to shake them vigorously or respond to their 0:12:10.120 --> 0:12:14.000 posts by saying to do something with your something, or 0:12:14.000 --> 0:12:17.199 that they're stupid. I understand the temptation, but you want 0:12:17.200 --> 0:12:19.400 to keep a level head here. Keep your head on 0:12:19.440 --> 0:12:23.080 the swivel. They thrive on this victimization. I'm sorry. If 0:12:23.080 --> 0:12:25.040 you're an AI booster and this makes you feel bad, 0:12:25.080 --> 0:12:26.920 Please reflect on your work and how many times you've 0:12:26.960 --> 0:12:29.280 referred to somebody who didn't understand AI in a manner 0:12:29.360 --> 0:12:31.760 that suggested that they were ignorant or tried to gaslight 0:12:31.800 --> 0:12:34.800 them by saying AI was powerful while providing no actionable 0:12:34.800 --> 0:12:38.160 ways of proving this or prove or just being able 0:12:38.200 --> 0:12:41.240 to point in it being powerful. You cannot and should 0:12:41.280 --> 0:12:43.160 not allow these people to act as if they're being 0:12:43.240 --> 0:12:46.720 victimized or other Throughout this series, I'm going to address 0:12:46.760 --> 0:12:49.000 a very specific thing. I'm going to use a term 0:12:49.240 --> 0:12:51.680 boost the quip. This refers to the things that they 0:12:51.760 --> 0:12:54.800 say and how often you hear them. These are lines 0:12:54.840 --> 0:12:56.920 that you'll hear them say again and again and again 0:12:57.000 --> 0:13:00.600 and again. They're common arguments, common cliches that demand response. 0:13:00.920 --> 0:13:03.640 And let's start with our first booster quip. You are 0:13:03.760 --> 0:13:06.760 just being a hater for attention. Contrarians just do it 0:13:06.800 --> 0:13:10.360 for clicks and headlines. First and foremost, there are boosters 0:13:10.480 --> 0:13:13.199 are pretty much every major think tank, government agency, and 0:13:13.240 --> 0:13:17.040 media out there. It's extremely lucrative being an AI booster. 0:13:17.320 --> 0:13:20.360 You're showered with panel invites access to executives, and you're 0:13:20.360 --> 0:13:22.360 able to get headlines by saying how scared you are 0:13:22.360 --> 0:13:24.560 of the computer. And it's really easy to do. So 0:13:24.920 --> 0:13:28.080 being a booster is easy. And I must be clear 0:13:28.120 --> 0:13:31.760 when I say booster, it doesn't always have to mean 0:13:31.920 --> 0:13:33.920 over It could just mean the things you choose not 0:13:34.000 --> 0:13:36.360 to do. It could mean the things you choose not 0:13:36.360 --> 0:13:38.360 to criticize them for, or the things you just write 0:13:38.400 --> 0:13:40.880 down that they say. But really we're talking about the 0:13:40.880 --> 0:13:43.560 worst of them. But if you hear that sentence and 0:13:43.640 --> 0:13:46.520 you don't think you're a booster, you can be a booster, 0:13:46.840 --> 0:13:49.400 by the way, if you just choose not to criticize them. 0:13:49.400 --> 0:13:52.440 But we're talking about the real assholes. Being a critic 0:13:52.559 --> 0:13:54.920 requires you to constantly have to explain yourself in a 0:13:54.920 --> 0:13:57.480 way that boosters never have to. Now, if a booster 0:13:57.600 --> 0:13:59.559 says this to you, if they say you're just being 0:13:59.600 --> 0:14:02.200 a hater for attention, you're just doing this for clicks, 0:14:02.559 --> 0:14:04.840 ask them to explain first of all what they mean 0:14:04.880 --> 0:14:07.120 by clicks or attention, and how they think you are 0:14:07.120 --> 0:14:10.600 monetizing it, how this differs in its success from say, 0:14:10.760 --> 0:14:14.000 anybody who interviews and quotes Sam Altman or Dario Wario, 0:14:14.080 --> 0:14:17.960 Ama Day or whomever from Anthropic on Hard Fork. Ask 0:14:18.040 --> 0:14:20.600 them what the difference is, and ask them why do 0:14:20.680 --> 0:14:23.360 they believe your intentions as a critic are somehow malevolent 0:14:23.400 --> 0:14:25.640 as opposed to those literally reporting what the rich and 0:14:25.680 --> 0:14:28.680 powerful want them to. There's no answer here, because this 0:14:28.840 --> 0:14:32.480 is not a coherent point of view. Boosters are more successful, 0:14:32.680 --> 0:14:35.360 get more perks, and are in general treated better than 0:14:35.360 --> 0:14:39.120 any critic and pretty much every major out there. Fundamentally, 0:14:39.160 --> 0:14:41.400 these people exist in the land of the vague, and 0:14:41.440 --> 0:14:43.600 they don't like it when you force them to get specific. 0:14:44.000 --> 0:14:46.440 They will drag you toward what's just on the horizon, 0:14:46.480 --> 0:14:48.600 but never quite define what the thing that dazzles you 0:14:48.720 --> 0:14:51.840 so much will be or when it will arrive. Really, 0:14:51.880 --> 0:14:54.400 their argument comes down to one thought. You must get 0:14:54.480 --> 0:14:56.400 on board now because at some point it will be 0:14:56.480 --> 0:14:59.600 so good you'll feel so stupid for not believing something 0:14:59.640 --> 0:15:02.520 that it sucks wouldn't be really good. And if this 0:15:02.640 --> 0:15:04.640 line sounds familiar, it's because you've heard it a million 0:15:04.640 --> 0:15:10.560 times before, most notably with cryptocurrency, NFTs, Mataverse, Clubhouse, tons 0:15:10.600 --> 0:15:13.680 of movements. Actually, they will make you define what would 0:15:13.720 --> 0:15:16.160 impress you, which is not your job. In the same way, 0:15:16.200 --> 0:15:18.960 finding the use case for them isn't your job. In fact, 0:15:19.160 --> 0:15:22.400 you're the customer, you're the consumer. You are the person 0:15:22.520 --> 0:15:25.160 AI needs to prove itself too, not the other way around. 0:15:25.880 --> 0:15:28.560 But let's go to another booster quip. When they go, 0:15:29.760 --> 0:15:33.920 you just don't get it, here's a great place to start. Say, 0:15:34.080 --> 0:15:36.920 that's a really weird thing to say. It is peculiar 0:15:36.960 --> 0:15:40.240 to suggest that somebody that doesn't get how to use 0:15:40.280 --> 0:15:44.360 a product is weird, and that we as the consumer, 0:15:44.480 --> 0:15:48.920 as the customer, must justify ourselves to our own purchases. No, no, 0:15:48.920 --> 0:15:51.560 no no. If I don't get it, it's the booster's 0:15:51.600 --> 0:15:56.280 job to tell me. Why make them justify their attitude. 0:15:56.560 --> 0:15:59.680 Just like any product, we buy software to serve a need. 0:16:00.080 --> 0:16:02.800 This is meant to be artificial intelligence. Why is it 0:16:02.880 --> 0:16:05.040 so fucking stupid that I have to work out why 0:16:05.080 --> 0:16:07.600 it's useful? The answer is, of course, that it has 0:16:07.640 --> 0:16:10.600 no intellect. It is not intelligent, and large language models 0:16:10.600 --> 0:16:13.000 are being pushed up mounted by a cadre of people 0:16:13.000 --> 0:16:16.080 who are either easily impressed or invested, either emotionally or 0:16:16.160 --> 0:16:19.200 financially in its success. Due to the company. They keep 0:16:19.200 --> 0:16:21.760 all their intentions for the world. And if a booster 0:16:21.840 --> 0:16:24.120 suggests you just don't get it, ask them to explain 0:16:24.160 --> 0:16:27.880 the following What am I missing, What specifically is it 0:16:27.960 --> 0:16:30.240 that is so life changing about this product based on 0:16:30.280 --> 0:16:33.040 your own experience, not on any anecdotes from other people, 0:16:33.240 --> 0:16:35.040 because they will say, well, I heard of a guy 0:16:35.080 --> 0:16:37.040 who are ten billion lines of coat, and then the 0:16:37.080 --> 0:16:40.880 baby looked at me and I cried. They don't have 0:16:40.960 --> 0:16:44.520 real things themselves, So cut off the exits brought up 0:16:44.560 --> 0:16:47.560 the doors. And then also ask them what use cases 0:16:47.560 --> 0:16:51.680 are truly transformative about AI. Don't let them say well, 0:16:51.680 --> 0:16:54.680 I heard in an industry. Actually make them prove themselves. 0:16:55.120 --> 0:16:57.440 The use cases will likely be the AI is replaced. 0:16:57.480 --> 0:17:00.480 Search for them that they use it for brainstorming or journaling, 0:17:00.520 --> 0:17:02.680 proof reading an article, or looking through a big pile 0:17:02.720 --> 0:17:05.159 of their notes or some other corpus of information and 0:17:05.200 --> 0:17:09.119 summarizing it or pulling out insights. Who gives a shit? 0:17:10.160 --> 0:17:12.920 Sorry not to be too a cerbic, but really, who 0:17:12.960 --> 0:17:16.240 fucking cares? That shit's so boring? Hundreds of billions of 0:17:16.280 --> 0:17:18.959 dollars of wasted investment on this and this is what 0:17:19.000 --> 0:17:22.880 We've got three years in fucking Humpty dumpty. Could never 0:17:22.920 --> 0:17:26.199 have it this good anyway. Our next booster quip is 0:17:26.280 --> 0:17:29.800 one of my faves. Ahem, AI is powerful and getting 0:17:29.840 --> 0:17:33.280 exponentially more powerful. Now, if a booster ever refers to 0:17:33.320 --> 0:17:36.680 AI being powerful and getting more powerful, ask them the following, 0:17:37.040 --> 0:17:40.400 what does powerful mean? In the event that they mentioned benchmarks, 0:17:40.440 --> 0:17:43.360 ask them how those benchmarks apply to real world scenarios. 0:17:43.680 --> 0:17:47.000 If they bring up swe bench the standard benchmark for coding, 0:17:47.160 --> 0:17:48.960 ask them if they can code, and if they cannot, 0:17:49.600 --> 0:17:52.960 ask them for another example. I mean, they will tell 0:17:53.000 --> 0:17:55.119 you that they've spoken with coders. I've talked with a 0:17:55.119 --> 0:17:57.120 lot of coders. I have a great one with Colt 0:17:57.200 --> 0:18:01.880 Vogie coming up, another software engineer talking about MS. It's 0:18:01.920 --> 0:18:04.959 so funny. It's so funny when you actually lay this 0:18:05.000 --> 0:18:08.040 stuff out, how weak their arguments are. But in the 0:18:08.160 --> 0:18:11.120 event they mentioned reasoning, ask them to define them. Once 0:18:11.119 --> 0:18:14.000 they've defined reasoning, ask them to explain in plain English, 0:18:14.000 --> 0:18:16.879 what reasoning allows you to do on a use case level, 0:18:17.160 --> 0:18:19.679 not just how it works. They will likely bring up 0:18:19.680 --> 0:18:22.000 the gold medal performance the open AI's model got on 0:18:22.000 --> 0:18:25.399 the meth OLYMPIAD. Ask them why open ai hasn't released 0:18:25.400 --> 0:18:28.160 that model. Then ask them what the actual practical use 0:18:28.200 --> 0:18:30.840 case is that this success has opened up. They will 0:18:30.840 --> 0:18:33.320 say it's an innovation. You've got to be patient, and 0:18:33.359 --> 0:18:37.240 then pairper spray them. No, don't, don't pepper spray anyone anyway. 0:18:38.119 --> 0:18:40.040 You should also then ask them what use cases have 0:18:40.160 --> 0:18:42.920 arrived as a result of models becoming more powerful. If 0:18:42.960 --> 0:18:45.360 they say vague things like oh in coding and oh 0:18:45.480 --> 0:18:48.880 in medicine, ask them to get specific, and then ask 0:18:48.880 --> 0:18:51.240 them what new products have arrived as a result. If 0:18:51.240 --> 0:18:53.399 they say coding LMS, they will likely add that this 0:18:53.480 --> 0:18:57.320 is replacing coders. Ask them where that has happened and 0:18:57.400 --> 0:19:01.040 ask them to show you proof. Blinks and it will 0:19:01.040 --> 0:19:03.879 not be sufficient to say that a CEO mentioned that 0:19:03.960 --> 0:19:09.160 they did something with AI and efficiency. Get numbers. They 0:19:09.240 --> 0:19:11.800 just won't. They won't do this. They will turn into 0:19:11.840 --> 0:19:13.679 a pillar of salt. And we got two more fucking 0:19:13.720 --> 0:19:16.000 parts of this. I mean you're gonna have a ball 0:19:16.040 --> 0:19:18.800 with this. Look. The core of the AI booster's argument 0:19:18.920 --> 0:19:20.840 is that they need to make you feel bad. They 0:19:20.920 --> 0:19:23.440 like the gaslight, and you should. You need to refuse 0:19:23.520 --> 0:19:25.439 to let them. You need to push back heavily. If 0:19:25.520 --> 0:19:27.879 there's ever a point where you feel like they are 0:19:27.880 --> 0:19:29.840 trying to make you feel stupid, ask them why they're 0:19:29.840 --> 0:19:33.159 doing so. And to be clear, anyone with a compelling 0:19:33.240 --> 0:19:35.399 argument doesn't have to make you feel bad to convince you. 0:19:35.720 --> 0:19:39.880 The iPhone didn't need a spurious hype cycle of proof 0:19:40.080 --> 0:19:42.679 of people saying you must look at this, it is 0:19:42.760 --> 0:19:46.679 so important. When it didn't really work. It worked immediately. 0:19:47.160 --> 0:19:49.000 Now I said in my news that it didn't need 0:19:49.040 --> 0:19:52.040 a fucking marketing campaign. Yes there was marketing dollars behind 0:19:52.040 --> 0:19:55.680 the iPhone. Eric newcomer, come at me with better WORKMATEE 0:19:55.720 --> 0:19:58.119 respond to the rest of this, But we all kind 0:19:58.119 --> 0:19:59.479 of got it with the iPhone. In fact, the moment 0:19:59.520 --> 0:20:02.600 Steve job announced their piece of shit that he was, 0:20:02.400 --> 0:20:05.280 he was. He said, here's the phone, here's an IPO, 0:20:05.680 --> 0:20:08.240 here's here's ema. You could do all these on one device, 0:20:08.320 --> 0:20:12.120 and everyone, oh, yeah, that is good. That is really 0:20:12.160 --> 0:20:15.359 obvious that that was good. It was impressive because it 0:20:15.440 --> 0:20:19.880 was impressive. And boosters will suggest you are intentional or 0:20:19.960 --> 0:20:21.879 in not liking AI because you're a hater or a 0:20:21.880 --> 0:20:24.439 cynic or a luddie. They'll suggest that you're ignorant for 0:20:24.480 --> 0:20:28.199