The AI Playbook: Tom Gardner of The Motley Fool

AI战略手册：The Motley Fool的Tom Gardner专访 Tom Gardner探讨了The Motley Fool如何围绕快速的AI部署，对其经营了三十年的金融媒体和股票研究业务进行重塑。公司从人工覆盖模式转型为自主、工程师辅助的工作流，使单个分析师能够审视数百只股票。他解释了公司如何转向全球远程招聘，通过自愿买断方案而非传统裁员来确立强制性AI标准，并在缩短的决策周期内对不断发展的前沿模型进行基准测试。Gardner还分析了宏观市场动态，认为超精简型初创公司将颠覆庞大的传统行业巨头，基础设施支出将为二级供应链带来意外红利，且AI模式识别将持续追逐定价失效，而非形成单一的市场共识。 在The Motley Fool，AI编排和算法准则目前负责大约60%的股票研究工作，使单个分析师覆盖的上市公司数量从30家提升至200到400家以上。 (15:14) Tom Gardner预计，AI杠杆很快将催生出以100亿美元被收购的6人初创公司，而24人的团队便能复制拥有1500名员工的传统老牌企业的知识产权。 (36:03) 在大量合成AI简历的冲击下，传统招聘渠道陷入崩溃，促使The Motley Fool采用主动算法寻访以及为期4至8周的付费试用期合同来评估工程师。 (7:19) 为了恪守长期以来不裁员的承诺，同时推行激进的技术适应措施，The Motley Fool将AI采纳定为强制性绩效标准，并向有抵触情绪的员工提供自愿买断离职方案。 (40:18)

In a discussion hosted by NYU Stern, The Motley Fool CEO Tom Gardner argues that artificial intelligence is fundamentally dismantling legacy operating structures across finance and software, shifting the competitive advantage to hyper-lean, AI-orchestrated teams. Reinventing Equity Research and Analyst Capacity (15:14) Integrating frontier AI models has fundamentally expanded the scale of equity analysis while shifting the analytical division of labor. Historically, an equity research analyst manually covered around 30 companies using physical filings and stock sheets. By combining an analyst's domain intellectual property with technical orchestrators, that individual can now track 200 to 400 companies. The modernized analytical pipeline operates on an approximate 60/40 model, where automated quantitative scoring systems handle 60% of the extraction and footnote scrutiny, leaving the remaining 40% to discretionary human synthesis. Disruptive Talent Strategy and Organizational Speed (4:33) Adapting to AI disruption required abandoning traditional office setups and conventional hiring funnels. The Motley Fool moved away from an Alexandria-centric workforce to an asynchronous, remote model, utilizing proprietary algorithmic tools to actively headhunt specialized machine-learning engineers across global hubs like Bucharest, Milan, and the Dominican Republic. Because inbound hiring pipelines are now flooded with synthetic, AI-generated resumes, the company relies on automated candidate scoring and four- to eight-week trial contracts rather than traditional job postings. Operational agility hinges on rapid model evaluation. Internal teams evaluate updates across models such as Claude, GPT, Grok, and Gemini within a strict six-hour Slack window to determine whether computational drift or degraded reasoning requires deprecating an engine from their scoring pipelines. Gardner notes that legacy software firms face a 'melting ice cube' as intellectual property value trends toward zero, while lean startups of 24 individuals (12 engineers and 12 domain experts) can duplicate platforms that previously required up to 1,800 employees. Navigating Cultural Resistance and Buyout Realities (40:18) Transitioning legacy workforces to mandatory AI workflows presents acute organizational friction. After maintaining a strict no-layoffs pledge originating from the 2001 dot-com crash, The Motley Fool addressed internal resistance among its 640 employees by offering voluntary exit packages and structured buyouts. Leadership made AI adoption a mandatory performance benchmark, concluding that allowing staff to casually opt out of modern tools compromises fiduciary obligations to research subscribers. Gardner admits his primary managerial misstep was approaching AI integration too gradually during 2023 out of concern for employee comfort. He argues that sheltering employees from aggressive adoption disadvantages their marketability, as the 'invisible competitor' of autonomous intelligence rapidly reshapes corporate job definitions. Market Volatility, Capital Allocation, and Systemic Risk (25:32) Rather than instantly producing efficient pricing, expanded AI tooling is heightening near-term market volatility. Market participants apply divergent time horizons, ranging from multi-year fundamental investors to high-frequency quantitative shops like Jane Street and Citadel trading over 30-second windows. While AI pattern recognition helps detect accounting discrepancies and governance red flags—such as serial shell-company promoters—Gardner dismisses the idea that unified algorithmic models will ever eliminate market pricing inefficiencies entirely. Macro risks are compounded by capital expenditure cycles and shifting corporate financing. Gardner cautions that vendor incentives, such as hardware bundled with equity warrants, soften revenue quality in AI infrastructure buildouts. Although short-term capital expenditure shocks could trigger market pullbacks between 11% and 29%, Gardner views these dips as operational speed bumps along an inevitable secular monetization trend powered by superior return on invested capital. While AI radically magnifies individual productivity and analytical breadth, it drives capital concentration into private markets where lean startups delay public offerings, locking out retail investors. Sustainable advantage relies on continuous operational reinvention and rigorous human oversight, as over-reliance on unverified models introduces severe governance, cybersecurity, and systemic risks.

The Motley Fool联合创始人兼首席执行官Tom Gardner指出，人工智能的快速演进正在重塑金融研究与知识经济的底层逻辑；企业若固守传统工作流将面临淘汰，唯有重构组织架构、推进人机协同并主动颠覆既有模式才能维持竞争力。 颠覆传统研究：人机协同重塑股票分析效率 (15:14) 过去二十年间，受限于阅读纸质10-K财报和附注等手动流程，华尔街分析师的合理覆盖上限通常在30家公司左右。Gardner表示，在AI工程师与技术编排团队的配合下，单一分析师能够高质量覆盖200到400家企业。自动化系统可批量抓取数百家公司的SEC申报文件，不仅能秒级核查收入质量，还能扫描过去可能引发类似安然（Enron）或世通（WorldCom）暴雷的隐藏负债与股价触发条款。 当前The Motley Fool的股票分析流程形成了约“六成算法、四成人工”的专业分工：60%的分析依赖自动化定量与定性评分系统，其余40%仍需资深分析师注入尚未提示词化的主观洞察与经验。Gardner建议投资者通过具体案例进行实践，例如将巴菲特（Warren Buffett）一生的投资哲学与框架设定为AI智能体，输入10家小型保险公司的当前财报进行1至100分的打分，从而快速穿透复杂行业的财务机理。 组织范式转移：全球分布式猎头与敏捷微型团队 (7:19) 基础设施由传统纸媒向AWS云迁移之后，The Motley Fool彻底拥抱了无办公室的远程办公模式。Gardner认为，远程办公抹平了层级结构，使公司得以在全球范围内延揽顶尖机器学习人才，包括罗马尼亚、意大利和多米尼加等地区。由于传统求职信充斥着由AI生成的虚假简历，公司转为使用内部自研工具主动搜寻候选人，并设立4至8周的试用合同以检验实战能力。 这支工程团队针对前沿模型更新建立了严密的测试机制。每当Claude、GPT、Grok或Gemini发布新版本，团队会在Slack上启动6小时评估窗口，若发现计算优势下滑或评分偏差便立即在排线中替换。Gardner预计，未来拥有12名AI工程师与12名领域专家构成的24人敏捷初创团队，即可匹敌传统需1500至1800人的中型软件或研究机构，六人估值百亿美元的超级精简企业可能很快出现。 阵痛与重塑：从不裁员承诺到淘汰抵抗文化 (40:18) 面对前沿工具的普及，The Motley Fool的管理层在文化与信义义务之间经历了巨大考验。Gardner自2007年担任CEO后一直坚守不裁员承诺，但生成式AI的冲击使得工具应用成为底线要求。由于部分员工对新兴工具有抵触情绪，公司为640名员工设定了明确的AI绩效指标，并向不愿转型的员工提供自愿买断离职方案，以此化解能力断层。 Gardner坦承，自己最大的管理失误是未能早在2023年就采取强硬立场，过分迁就员工的舒适区反而延误了其市场适应力。他列举了百视达（Blockbuster）、Kmart等巨头在技术变革中瓦解的教训，指出当代企业必须像Shopify要求员工变换开会姿态与鼠标习惯那样，根除心理惯性。如果一线研究人员拒绝借助最新模型排查风险，公司的受托责任便会遭到直接侵蚀。 资本市场与估值重构：短期波动与长期商业化 (25:32) 针对AI能否立刻让市场更有效率的疑问，Gardner指出短期内市场波动性反而会显著放大。尽管前沿技术提供了强大的深挖工具，但各类市场主体的策略与时间框架高度分化：长线价值投资者与毫秒级高频量化机构并存，且围绕底层模型的微小差距展开争夺，使得阿特拉斯（Atlassian）等软件企业在数月内经历剧烈的股价过山车。Gardner驳斥了所有AI终将归拢为单一理性定价实体的论调，坚信追求超额收益的信息竞逐永无止境。 针对数据中心建设是否存在类似次贷危机的泡沫担忧，Gardner指出，虽然部分硬件供应商通过股权权证补贴客户带来了低质量收入，且资本开支受阻可能引发大盘11%至29%的回调，但这属于阶段性减速而非结构性终点。从长期来看，由于软件本身的代码壁垒正在迅速趋向归零，单纯的软件IP价值大幅缩水，而能够借助AI进入“进攻模式”的领域（如药物研发与电力设备）将通过卓越的资本回报率与营业利润率实现全面商业化变现。 该对话表明，AI正在根本性加速知识工作和金融领域的优胜劣汰；然而，随着高价值初创企业长时间保持私有化状态，普通散户正被排除在早期财富积累之外，且网络安全与系统性脆弱性依然构成了企业全面依赖模型分析的重大盲区。

[2:06] Tom Gardner: It doesn't take a lot of deeper thought to realize that financial services will be the first category or one of the first categories disrupted by any new technology as everyone tries to figure out, how can I get an edge? [53:01] Tom Gardner: A dollar of revenue growth without incentives is not a dollar of revenue growth with incentives, right? The latter is weaker. Tom Gardner: And we've already said, and we've already seen, you're gonna see six person companies bought for $10 billion. [4:33] Tom Gardner: And we know that in technology companies, you start to look a little bit more like an NBA team, let's say, you need the top elite players. [20:29] Tom Gardner: Information is coming so quickly, breakthroughs are coming so rapidly that it's not entirely clear to me what investing will look like four years from now relative today.

[0:07] Melissa Schilling: Hello everyone, I hope you can see the video okay. [0:11] Welcome to our first fireside chat in our brand new AI playbook series. [0:16] In this series, we're going to meet with managers in a wide range of [0:19] industries to hear how they are leveraging AI in their businesses, [0:22] including the challenges and opportunities that entails, [0:26] and the lessons they've learned along the way. [0:28] We're going to have time at the end of each of these chats for Q&A, [0:31] so please drop your questions in [0:33] the Q&A box at the bottom of the screen. [0:36] If you don't see it automatically at the bottom of your screen, [0:38] just choose the option for more and you should see it there. [0:42] Today, we are extremely fortunate to have as our first guest, [0:45] Tom Gardner, CEO and co-founder of The Motley Fool. [0:49] I'm so excited about this. [0:51] The Motley Fool is a private financial and investing advice company [0:55] that's been around since the early 90s that provides stock market news, [1:00] analysis, and subscription-based recommendations for individual investors. [1:04] Tom has authored several books, [1:06] including The Motley Fool Investment Guide, [1:08] You Have More Than You Think, [1:10] and Rule Breakers, Rule Makers. [1:12] He's also been a major advocate for transparency and reform in [1:16] the financial services industry and has even testified [1:18] before US Congress towards that goal. [1:21] He says the mission of The Motley Fool is to make the world smarter, [1:25] happier, and richer. [1:27] So please give a warm welcome to Tom. [1:30] Tom, I'm so excited to have you here. [1:34] Tom Gardner: Mutually enthusiastic and don't you need to, [1:37] in the name of transparency, [1:40] identify that you are a board director of our company? [1:43] Melissa Schilling: Yes. In the interest of transparency, [1:45] I am a board member at The Motley Fool, [1:47] which is part of how I got to know it and got to [1:49] see what exciting things you were doing and probably [1:51] also give me a little edge in getting to have you as a speaker, [1:54] since you're a pretty senior guy who's pretty busy. [1:58] Tom Gardner: Well, I'm very happy to be here. Looking forward to this. [2:00] Melissa Schilling: All right. Cool. [2:02] So for those people who aren't familiar, [2:03] will you tell them a little bit about what The Motley Fool does? [2:06] Tom Gardner: Sure. The Motley Fool started in 1993. [2:11] We were first a print newsletter, [2:14] mostly for friends and family, [2:15] and then we discovered America Online, [2:17] back then Prodigy, CompuServe, [2:19] and the Internet News Groups and found an opportunity to [2:23] share our thinking about long-term investing online. [2:27] We began to very quickly see the benefits of being able to [2:30] talk to other people around the country or around the world, [2:33] share different investing ideas, [2:36] individual companies, or overall strategies. [2:38] Of course, at that point, it felt so new. [2:41] It was very anonymous. [2:43] A lot of the media interviews that we were brought into [2:47] were wondering if it was all a fraud, [2:50] what was happening out online. [2:51] Instead, 33 years later, [2:54] we've been through some substantial ups and downs, [2:56] the ecstasy and the agony of being [2:58] an entrepreneur and being a privately held company. [3:02] Today, yes, you named our mission to help the world invest [3:05] better and to make each of our members and everyone [3:07] associated with our organization smarter, [3:09] happier, richer for their time with us. [3:11] It has certainly been an amazing last 18 months, [3:17] and I'm looking forward to sharing our journey [3:19] because it doesn't take a lot of deeper thought [3:23] to realize that financial services will be [3:25] the first category or one of the first categories [3:27] disrupted by any new technology [3:29] as everyone tries to figure out, [3:31] how can I get an edge? [3:32] How can I get better analysis of a company [3:34] or better information on geopolitical situations? [3:41] Any dynamics around the world that can affect [3:43] how I'm investing would be one of the most important [3:46] breakthrough zones for new technologies, [3:49] and we've seen that with AI. [3:51] Melissa Schilling: Yeah, one of the things I've been constantly amazed about [3:54] by you is how you leap forward into the new things [3:58] that are happening in the world, [3:59] whereas a lot of CEOs, especially founders of companies, [4:03] tend to be a little bit, [4:04] might be a little bit more on their heels [4:06] about something like this, [4:07] which has potential to be quite disruptive [4:09] and create a lot of change, [4:10] and change can be costly. [4:11] That's not your personality at all. [4:14] Your personality is to say, [4:15] okay, there's this new thing. [4:17] We're going to charge forward. [4:18] So why don't you tell people a little bit about [4:21] how you are currently leveraging AI at The Motley Fool [4:25] or maybe how you sort of started [4:27] and whether there were any stumbling blocks [4:30] or just tell us about that journey a little bit. [4:33] Tom Gardner: Well, I'm a big believer in Dr. Schilling's work, [4:37] and you really helped us when we began our cloud migration, [4:41] which was late. [4:42] So I do think we're much more technically [4:45] at the vanguard today, [4:46] but it started for us with the cloud migration, [4:48] realizing we were behind. [4:50] We're a private company. [4:51] We're a publishing subscription business for the most part. [4:55] There weren't necessarily any early breakthroughs for us [4:59] in some of the technologies [5:00] after we were online, [5:01] publishing successfully in a subscription format. [5:04] Once you move to the cloud, [5:05] you start to realize, [5:06] when you're, in our case, on AWS, [5:08] there are a variety of new skillsets [5:10] that aren't going to be developed quickly [5:12] for your internal group [5:14] if they haven't been upscaling throughout. [5:16] So that was a big cultural change for us moving on to AWS. [5:19] But once we landed there, [5:20] we had brought in a lot of advisors, [5:22] Alyssa among them, [5:23] and I was able to get very active advice every day [5:26] from experts around the world [5:29] on what will the impact be [5:31] for the earliest forms of chat GPT in 2022, 2023. [5:37] How can we use these tools? [5:39] And I think a big thing for us is that post COVID, [5:41] we chose not to mandate back to the office. [5:45] That has some disadvantages, [5:47] but some of the advantages for us [5:48] have been contracting talent around the world, [5:51] everyone being in the same size digital square. [5:53] So if you're in an office [5:55] and there are eight of you in the meeting room [5:56] and one person on the monitor, [5:58] I mean, eight people have a connection advantage [6:02] by being in that room. [6:03] So we really have been bringing expertise [6:05] in from around the world. [6:07] And that's probably the most important thing for us [6:10] is that we're heavily on Zoom and Slack 24 seven. [6:13] We have great AI engineers in Bucharest, [6:19] in Milan, in Dominican Republic. [6:23] So we've become much more of a global talent force. [6:27] And that also has meant that tenure isn't very important. [6:31] And we know that in technology companies, [6:33] you start to look a little bit more like an NBA team, [6:35] let's say, you need the top elite players. [6:38] You're gonna take them anywhere you can get them. [6:40] You're gonna put them in the game [6:41] to win competitively in the sport. [6:44] And that's been a transformation to our culture [6:46] because I'd say for the most part, [6:48] we were much more single location, Alexandria, Virginia, [6:52] again, publishing model. [6:55] And we hadn't had as much change inside of our company [6:58] as we've seen in the last two years in the last 20. [7:01] Melissa Schilling: Wow. [7:02] Okay, wait. [7:02] So I have several questions from that. [7:04] But the first thing I wanna know is, [7:07] how did you locate, [7:08] you talk about this enabling you to tap [7:10] a global talent force, [7:12] finding the best people around the world. [7:13] But how would you do that? [7:15] I mean, how'd you even know what you were looking for? [7:16] And how would you reach those people? [7:19] Tom Gardner: Well, a shift for us was to move from listing jobs [7:22] and sort of traditional recruiting [7:24] to more of a headhunting approach [7:26] and building AI powered tools to find and score applicants. [7:30] As we know, anyone who's in recruiting today [7:33] is getting a lot of fake AI submissions. [7:36] So I do think that a lot of recruiting [7:38] has become more finding and selecting [7:40] versus taking incoming applications. [7:43] We were relatively early on with that. [7:45] One of our contractors, Jonathan Wilder, [7:47] built a tool that helps to score candidates [7:49] based on synthetic job descriptions in any domain. [7:52] But our primary focus was, [7:54] how are we gonna find AI engineers [7:56] that have multiple years of experience? [7:58] Because the gap, for example, [7:59] between somebody with, let's say, [8:02] PhD level interest in machine learning, [8:06] in data analytics, who was on GPT, [8:10] maybe even before ChatGPT arrived, [8:12] certainly exploring and asking questions, [8:14] mostly the gains that they've had [8:16] over the average user of these technologies today [8:19] are around judgment, [8:21] the ability to understand what is this tool capable of? [8:24] And now with the pace of upgrades [8:26] across the frontier models, [8:27] having a team of AI engineers, [8:29] I think it's essential for any company [8:31] in a category that's being disrupted, [8:33] not commercial landscaping, [8:35] not coin-operated laundromat, let's say, [8:38] but the center of finance [8:41] where everyone's racing to get the most information [8:45] analyzed and assessed most effectively, [8:48] as quickly as possible. [8:50] All of a sudden, you start to realize [8:52] that there is a massive gap in capabilities [8:54] between somebody who's been deploying these models [8:57] in as many ways as possible to analyze information, [9:01] but also who is now sitting on the pace of upgrades [9:05] across Claude and GPT and Grok and Gemini, [9:10] and realizing that each of those upgrades [9:11] in the first six hours with a team, [9:13] in our case, on Slack, real-time saying, [9:16] we're going to have to remove this one [9:18] because the computational benefits [9:21] have actually diminished a little bit, [9:23] and it's not going to be as strong for us [9:24] in scoring companies. [9:25] So for us, finding that group [9:29] through that recruiting tool that we built, [9:31] deploying it, bringing people in on contract initially [9:34] and saying, let's try it out for four weeks, [9:36] six weeks, eight weeks. [9:37] In a number of cases, we had academics [9:39] that weren't really in a business-working [9:41] corporate environment before and have come in. [9:43] I think it's been particularly exciting for them [9:45] because they've been theorizing about what's possible, [9:49] and now they're able to apply it [9:50] in a category that needs it. [9:52] Toward a company, I think, [9:53] that has a mission that's inspiring in general [9:56] to the people that come to work at The Motley Fool, [9:58] the idea that we're not trying to pack this knowledge up [10:00] into a hedge fund and gain a particular advantage [10:02] for a narrow expression. [10:03] We're actually trying to bring it out [10:05] to the general public as quickly as possible. [10:07] So we've gotten some pretty remarkable people [10:10] from those tools, and it's a fascinating, [10:13] almost what we know, Melissa, we've all read or heard, [10:16] that those who are utilizing these tools [10:18] are actually working a lot more. [10:20] In fact, there's a lot of risk of burnout from enthusiasm, [10:25] not from pressure, not like, [10:26] hey, you need this by Tuesday at four, [10:28] but like, I couldn't sleep last night [10:30] because the breakthroughs that are available right now [10:34] are so brand new and so exciting to me. [10:36] So it does have a late-1990s Motley Fool feel to me, [10:39] what's happening with all the tools [10:40] and what they're making available. [10:42] Melissa Schilling: Wow, wow. [10:43] Okay, so I just wanna repeat back [10:44] something I think I heard you say, [10:45] and that was that you used to have a hiring model [10:48] where you would post a job listing [10:49] and then all these applications would come in [10:51] and you would sift through them. [10:52] Now you're using a proprietary tool [10:54] to go out and seek out candidates [10:56] that fit your criteria. [10:59] So you've moved from a push model to a pull model. [11:03] You're going out and searching out the candidates you want. [11:07] And how would someone who, [11:09] and I presume other companies are doing the same thing, [11:11] and how would somebody who wants to get spotted [11:14] by a tool like that make themselves [11:16] most likely to get spotted by a tool like that? [11:19] I feel like it's kind of funny [11:20] we're skipping straight to the placement implications, [11:22] but I think people in the audience might wanna know. [11:26] Tom Gardner: Well, I think the game, let's say, [11:31] of expressing who you are professionally on LinkedIn, [11:37] these were search tools that existed [11:40] years ago within LinkedIn. [11:42] I remember talking to one of their prior CEOs [11:44] who told me, as a quick distraction aside, [11:47] following the talent flows [11:49] inside of LinkedIn's platform and database [11:52] is apparently to him quite predictive [11:56] of value creation in the public markets. [11:58] When you see a number of technically expert people [12:02] leaving a large software business [12:03] and going to a series C emerging data software business, [12:13] it's kind of telling if that group is very talented. [12:16] So anyway, that's just an interesting aside. [12:17] But I would say that certainly the most important thing [12:21] would be to evaluate your areas of greatest interest [12:28] because we now have a new competitor, [12:32] an invisible competitor with a rising level [12:36] of sophisticated intelligence [12:38] that will impact domain after domain after domain. [12:42] Obviously not all at the same time, [12:44] but to look at what the skills are that are required [12:46] for the role that I'm interested in now, [12:49] let's say I wanna be a doctor, [12:51] what is it gonna be like to be a doctor five years from now? [12:54] And I think it's hard in these situations [12:56] when we get into a groove of the passions that we have [13:01] and we kind of expect that they're gonna persist. [13:03] We do have a benefit for us at The Motley Fool [13:05] in terms of transforming our culture [13:08] because we sat there in the front row [13:09] and watched MTV and People Magazine, [13:11] two of the dominant media brands of the 1990s. [13:15] I mean, I don't know in 2002 or 2003 [13:17] what would be considered more almost guaranteed [13:21] of their future than the companies like MTV, [13:24] brands like MTV and People Magazine now. [13:25] They have basically been rendered insignificant entirely. [13:30] And MTV could have, it's easy to play 2020 hindsight, [13:35] but could have become Spotify. [13:37] And People Magazine could have become Instagram. [13:39] I mean, every celebrity, [13:39] everybody, all the agencies [13:41] and everyone was trying to get you in People Magazine [13:43] and like trying to get you on MTV, [13:45] but the shift you would have had to make [13:47] from those organizations, [13:48] you would have to essentially tear down [13:50] all of your workflows. [13:52] And a wonderful playbook for that, [13:53] because I think companies are gonna have to do this, [13:57] is Andy Grove's book, Only the Paranoid Survive, [13:59] because the move from memory chips to CPUs inside of Intel, [14:03] I remember one of the lines [14:04] that friends were turning against each other [14:06] inside the business that had been partnered [14:08] and working together for a decade [14:10] and had many families that were vacationing together [14:15] and all of a sudden now they're on opposite sides [14:17] of the transformation and the inflection point is here. [14:20] So I think when we look at what's happening [14:23] in today's work environment, [14:25] it's really important to use the tools [14:28] to generate what your job is gonna look like [14:31] three years from five years from now, [14:32] and to begin building those skills as quickly as possible. [14:35] Classic cliche, escape where the puck is going, [14:38] not where it is. [14:39] And thereby by developing those skills, [14:41] have those skills articulated on your LinkedIn page. [14:45] And so that you're searchable by organizations [14:47] that are trying to push the vanguard of their category [14:51] with advanced AI. [14:56] Melissa Schilling: Wow. [14:57] Wow. [14:57] Okay. [14:58] That's a lot. [15:00] I wanna shift a little bit to talk about [15:02] how you're actually using AI at The Motley Fool. [15:06] Like what are some of the things that have changed, [15:08] the processes that have changed [15:10] or the products that have changed? [15:14] Tom Gardner: Some of the things that have not changed, [15:16] we continue to be a subscription model. [15:19] So I believe that will persist. [15:24] We continue to be focused on providing [15:27] the best investment guidance [15:29] to find the companies of tomorrow today [15:32] and to help our members build diversified portfolio [15:37] of stocks and ETFs. [15:39] And to overall just be a little bit more active [15:41] in their financial life [15:42] and to take a more of an ownership role. [15:44] I certainly think it would be good in society [15:46] if everyone's 18th birthday, [15:48] part of the celebration was, [15:50] you are the CEO of your financial life. [15:53] The sooner you realize, [15:54] obviously everyone's in different situations [15:56] moving through their teens and 20s. [15:57] At some point you're gonna wake up and realize [15:59] you are the CEO of your financial life. [16:00] You may outsource it to a money manager. [16:02] You may have a spouse or partner [16:05] that's gonna take care of most of it. [16:06] But overall, it would be good for us to realize, [16:08] hey, we're activated financially now. [16:09] We need to be thinking through our strategy. [16:12] So we want to be that partner for our members. [16:15] Some of the things that have changed. [16:17] Well, let's just go maybe first to stock analysis. [16:21] So there was fairly good research. [16:25] It was really published maybe 20 years ago [16:27] to say that an individual stock analyst [16:28] can really only effectively cover 30 companies. [16:32] So in a lot of cases on Wall Street, [16:33] you get assigned to an industry. [16:35] You're following utilities. [16:37] You're given 30 companies. [16:38] You're there to assess everything about them. [16:40] You're contacting management. [16:41] You're doing everything you can [16:42] to know those 30 businesses as well as possible. [16:45] I can talk about some of the problems [16:47] with that model on its own, [16:48] but I want to just talk about the technical shift [16:50] that you're asking about, [16:50] which is now it is much more possible [16:53] to follow 200 companies more effectively [16:58] than you could have covered 30. [17:00] And I might even go farther to like 400 companies [17:03] than you could have effectively. [17:05] So if I had an investor that said, [17:09] just let me follow these 30 businesses. [17:12] I'm going to do it in my old school way. [17:15] You know, that happened there 1993, 94, right? [17:17] Investors were using the S&P stock guide. [17:19] They were using the value line printed information. [17:23] They were getting 10 Ks sent to them in the mail [17:25] and those workflows were disrupted [17:28] by the ability to get all that information real time, [17:30] to talk to other investors and all the other benefits [17:33] and difficulties that emerged [17:36] by having connected networked devices [17:40] and being able to learn together online. [17:43] So now it's the ability [17:45] to process information differently. [17:47] So I'm going to take the person who says, [17:50] I will deliver for you excellent analysis [17:54] of 400 companies with these tools [17:57] and with orchestrators that are capable [18:00] of the technical work that I'm not capable of to help me. [18:03] I will cover 400 companies [18:05] rather than I will manually cover 30. [18:06] Because once you do that, [18:07] let's just say across the 400 companies, [18:10] you can run evaluations of the footnotes, [18:15] scoring the footnotes, [18:16] or you can score revenue quality. [18:18] You can score any aspect of any company [18:20] and find information that it would have been [18:25] as hard for you to find across 30 companies. [18:28] So you think of, there were a lot of institutions [18:30] that owned shares of Enron, right? [18:32] There was, right? [18:33] So of WorldCom, these were widely held businesses [18:36] that were flourishing, [18:37] but there were little bits of information [18:38] that journalists identified, [18:42] business journalists or analysts [18:43] or hedge fund managers identified. [18:46] And in some cases were shouting [18:47] and trying to say, look at how this, [18:49] like Enron's debt is tied to the price of their stock. [18:53] If their stock declines, that debt can be called. [18:55] They're gonna, this thing could unravel pretty quickly. [18:58] And so it was harder to even using the internet [19:02] and connecting with everyone else [19:03] to be able to find that little bit of information. [19:05] But now you can find that information very, very quickly. [19:09] So I would say, for example, the identification of fraud [19:12] and to go outside the public markets [19:15] and just say the identification of corruption, [19:19] it should be and is easier and easier with these tools. [19:23] But the other thing I would say that has changed [19:26] is it's essential to have your AI engineering partner [19:30] or team, because what they're capable of doing, [19:33] both with the discipline of sourcing the data [19:37] that's most relevant that you're guiding them perhaps [19:39] as the subject matter expertise, [19:40] but then cleaning that data, red teaming, [19:42] pushing back on the scores. [19:45] Now, to me, it is much more, [19:48] it is more sort of 60% [19:50] about the scoring system you're building. [19:53] And then the other 40% does trail down into manual work [19:56] where you're still using that expertise you have [19:59] that you haven't been able to put into a prompt system yet, [20:02] let's say, but that's a long-winded way of saying [20:05] just in the core of stock analysis now, [20:07] I would say that the gap between the productivity [20:11] of one person covering 30 businesses [20:12] and another person with AI engineering support [20:15] downloading all of their IP knowledge [20:17] to build the scoring system, let's say, [20:20] is allowing you to cover many, many more companies. [20:22] And I think more effectively, [20:23] to get better returns across 400 stocks [20:26] than you could get manually across 30. [20:29] Melissa Schilling: Wow, wow. [20:31] So do you think, I mean, [20:34] does this mean that there's gonna be less analysts out there [20:36] or does it mean that all the analysis [20:38] is just gonna get better and deeper and wider? [20:40] And what are your thoughts about that? [20:43] Tom Gardner: I think that I would say it's a little bit of a mix [20:46] and it will roll out in stages [20:48] and I have my own personal visions, [20:52] but I would say that what I'm beginning to realize [20:54] is that for whatever visionary contributions [20:57] I make to our company, [20:58] the timeline, the range, [21:03] how far out can I go with that vision is narrowing quickly. [21:07] So it's an interesting dynamic on every level [21:11] in terms of how to manage a business, [21:13] how to evaluate other companies, [21:15] how to make investment decisions. [21:18] Information is coming so quickly, [21:21] breakthroughs are coming so rapidly [21:22] that it's not entirely clear to me [21:25] what investing will look like four years from now [21:29] relative today. [21:30] But I would say certainly [21:32] there will be fewer analysts on average. [21:34] However, there will be more room for analysts [21:37] that are obsessed with using the new tools. [21:41] And I said before, certain categories, [21:43] well, if you're making shoes and shoelaces as a business, [21:46] you're gonna get efficiencies out of AI, [21:48] but you're not in an intelligence explosion [21:53] that's gonna change your business. [21:54] So I would say, for example, drug discovery, [21:57] this is a place where as an investor, [21:59] I've been making a lot of investments [22:00] because I see this category is immediately [22:02] going to be inclined to use these tools [22:04] and to play offense with them. [22:06] There are a lot of companies that are gonna be reactive, [22:09] trying to fail. [22:10] A lot of software companies [22:11] are watching the intellectual property value [22:13] of their software decline to zero. [22:15] They have sales teams, [22:16] they have contracts that extend many years, [22:19] like they're not gonna fall apart right away, [22:21] but the ice cube is melting for them. [22:23] And the pathway is not as clear for them. [22:26] So I would say, are there gonna be more analysts? [22:28] I think there will be fewer analysts. [22:30] I think there are many disturbing things oncoming [22:37] that we haven't figured out. [22:39] We won't go deep into regulation, [22:42] but there are a lot of unknowns right now. [22:45] And I do think, for example, [22:46] that's why I expect to see more companies [22:48] and more workforce formats being contract. [22:55] And that needs to, [22:58] there needs to be a balance struck [22:59] between those two, right? [23:01] So how those contracts are written, [23:02] how you develop your organization and build trust [23:05] when roles are changing so rapidly [23:07] is a real challenge for a business's employment reputation [23:14] or how they build their company culture. [23:16] But I do think that in general, yes, [23:18] we will have smaller teams producing much more output. [23:22] Therefore, we will have many more overlapping competitors. [23:25] You just think of the frontier models. [23:26] Which one matters more to you, Melissa? [23:28] I'm just curious. [23:29] Grok, Gemini, GPT, or Claude? [23:32] Do you have a personal preference [23:34] and do you have an explanation as to why you prefer that one? [23:37] Melissa Schilling: Well, I use Claude for some things [23:40] because everybody in my community recommends Claude, [23:42] but I actually use Perplexity [23:44] when I'm just having a conversation [23:46] over a coffee in the morning. [23:47] Like Perplexity and I chat back and forth. [23:49] Also, Perplexity is pretty good at gathering up sources [23:52] and searching if I'm doing like a literature review [23:54] because mine feels very anchored on research and stuff. [23:58] But if I were gonna do something maybe more complicated, [24:01] I'd probably use Claude. [24:02] I occasionally use ChatGPT 5.0 too [24:05] if I'm doing a research project. [24:07] Tom Gardner: And so I think that what you've described [24:10] is actually happening for a lot of different individuals [24:13] and then also inside of companies and enterprise licenses [24:16] and full teams working using these tools [24:18] is there's a lot of overlap between them. [24:20] And it's almost like they're in the 400 meter high hurdles. [24:24] And when you watch the European championships [24:26] and you see that the winner wins by 0.12, [24:31] they got their notes right at the last second [24:33] over the finish line. [24:34] I think there's so much overlap in these frontier models. [24:37] It's just a lot that's happening now [24:39] among the largest technology companies [24:41] is the canary in the coal mine [24:42] or is giving us a vision of the future. [24:44] And there's gonna be a lot of overlapping competition [24:47] in IP businesses. [24:48] It's not clear what will be the competitive advantage [24:51] and how long it will last. [24:52] Certainly speed, enthusiasm for discovery and exploration, [24:57] like absolute diminishment of complacency in the workforce, [25:01] a willingness to change your workflows. [25:03] And these things are hard, [25:04] particularly if you have decades of work habits [25:06] that have formed. [25:07] So I think the competition is gonna intensify [25:12] in all intelligence driven businesses [25:15] and we will see fewer people, more small team companies. [25:19] And I think there will be a widening advantage [25:22] for the businesses that are excited to use each new model [25:26] with each upgrade as it comes every two weeks [25:28] or every week or someday every two days. [25:32] Melissa Schilling: You inspired so many questions for me [25:34] and I have like a list of them. [25:35] I'm not gonna get to all of them, it's clear. [25:37] But one thing you said, I was curious about, [25:40] if we have a lot more people analyzing [25:43] and there's a lot more competition between the analysts [25:45] and they're able to go into this fine detail [25:47] you talked about, like really scrutinize [25:48] in the footnotes, [25:50] really looking for indicators of corruption, [25:52] that it sounds to me like we ought to be more precise [25:55] and accurate in our evaluation of companies. [25:58] And does that mean we would be less likely [26:00] to have bubbles, for example, [26:02] or to have things slightly, you mentioned Enron before, [26:05] like for a lot of people, Enron was a huge surprise, [26:08] a huge bad surprise. [26:09] Like, are we gonna have less of that [26:11] because we'll be better at tracking the information [26:14] all the time? [26:16] I mean, I don't mean when I say we, I don't mean me, [26:17] but I mean like the investment community? [26:21] Tom Gardner: I would guess, I will guess, I am guessing [26:25] that actually in the short term [26:26] you'll have more volatility and uncertainty [26:30] because the challenge of analyzing a single public company [26:36] will begin to widen as the concentric circles [26:42] of information expand and you start to realize, [26:45] wow, I needed to find that little piece of information [26:48] as early as possible [26:49] and draw the correct conclusion about it. [26:52] And that's, so we have some of our AI engineers, [26:56] the full building models that, for example, [26:59] estimate what the AI infrastructure spend will be, [27:03] and we're doing this two years ago, [27:04] and then they take that multi-trillion dollar spend [27:06] over the next five years, seven trillion [27:08] or whatever the number is, [27:09] and they sprinkle it down across all the relevant categories [27:13] and end up seeing which are the small companies [27:16] for whom a scrap off the table of a large tech business [27:20] is a complete game changer to their business [27:24] and their revenue model. [27:26] They may have only 106 million in sales [27:29] as a tiny public company, [27:31] but if they're gonna benefit from the demand [27:35] for electrical equipment or from upgrading the grid [27:38] in even a small way that's gonna take their sales [27:40] from 106 million to 170 million, [27:42] which is nothing in the grand scheme [27:43] of trillions of dollars being spent, [27:44] this company all of a sudden [27:46] might have 70% revenue growth annually [27:49] for the next three years. [27:50] And this stock that said whatever per share, [27:53] but its market cap is 300 million [27:55] is gonna be 4.6 billion in four years. [27:58] So I think this is real. [28:00] We're seeing it as real. [28:01] Obviously, if we were completely expert, [28:05] we would just find these one after another. [28:07] That's the spirit of investing. [28:08] You're trying to find the best opportunities right now. [28:10] And so I think the reason I think it'll be volatile [28:13] is that the use of the models [28:16] is gonna be very uneven and different. [28:17] And when I said concentrically, [28:21] the circles will widen as the information you need, [28:23] all of a sudden you will start to realize, [28:25] wow, I don't just need to evaluate this company [28:27] and get down to the level of its footnotes [28:29] and its 10K over the last three years [28:32] to see what has changed in the footnotes. [28:34] What new disclosure, [28:35] what risk have they started to identify [28:37] that they didn't before? [28:38] Maybe they were hiding it [28:39] or whatever the conclusions we're drawing. [28:41] That's going very deep into a single company, [28:43] but that company exists inside of an industry [28:45] with a number of trends. [28:47] Those trends include headwinds, tailwinds, and crosswinds. [28:51] I have to be able to score what those trends look like. [28:54] And all of a sudden you start to realize, [28:55] wow, as we've seen from great investors throughout time, [28:59] there's more than just the price-earns ratio. [29:02] There's more than just the last quarterly earnings report. [29:04] There's more than just the existing product suite. [29:07] There's more than just this management team [29:09] or the board of directors [29:10] or who the institutional shareholders are. [29:12] There's actually a whole trend, [29:14] set of trends that they exist in [29:15] and what are the scores and direction of those trends? [29:17] And then within those trends, we're in a market system. [29:20] We benefit in the US from the regulatory standard [29:23] that we have for public company stocks. [29:26] That is not the same if you invest in Vietnam. [29:29] You do not have the same regulatory standards. [29:31] You still can have great companies. [29:32] But so I think what's gonna happen is [29:37] the winners are gonna realize [29:39] that they wanna get a holistic view [29:41] of every individual business [29:43] with scoring and all those elements. [29:45] But the moment in time from here to getting to that view [29:49] is a lot of experimentation, [29:51] a lot of different weighting of scores, [29:53] different uses of the tools and different time horizons. [29:56] Remember, there are organizations [29:56] like Jane Street, Point72, DE Shaw, Citadel. [30:00] I mean, they're looking for the next 30 seconds. [30:03] They oftentimes may not even really know [30:05] what the business does. [30:06] They're just in using these systems to trade [30:08] very quickly. [30:09] So I think it's gonna be more volatile than not [30:15] until we get to the level of sophistication [30:19] with the systems where it could begin to bring stability [30:22] as pattern seeking becomes more commonly shared. [30:27] This is pretty speculative as to what will happen, [30:29] but I'll just say I wouldn't be surprised [30:30] if we continue to see higher levels of stock volatility [30:33] because the level of stock volatility in the US today [30:35] is very high. [30:36] The overall market is pretty relatively normal [30:40] in volatility, but you go below [30:41] the largest 10 technology companies [30:43] and you're having stocks like Atlassian [30:45] fall from 400 to 50 and up to 190. [30:50] I mean, the stock has gone from 50 to 190 [30:51] in a matter of a few months. [30:53] So there's a lot of uncertainty [30:55] and these tools are bringing new information to bear, [30:57] but it's more of a chaotic exploration right now [31:03] than a cool, smooth system that has emerged. [31:07] Melissa Schilling: Wow, wow. [31:08] Ruby just reminded me to remind the audience [31:11] to put your questions in the Q&A. [31:13] So don't forget to do that. [31:14] I wanna come back to something. [31:17] I'm not sure if we, I'm not sure if, [31:20] well, let me just pose it to you this way. [31:22] If we look at how companies have historically been valued, [31:26] a lot of it's kind of consumer sentiment [31:28] and trends on the stock price. [31:30] And a lot of times that value, [31:32] I talk about this a lot in my strategy class, [31:33] a lot of times the market cap [31:35] doesn't look anything like a net present value [31:38] of an income stream we expect [31:39] to get out of this company, right? [31:41] It's something else entirely. [31:46] Will more detailed analysis like this close that gap [31:50] so that companies that are currently undervalued [31:53] might see a rise in value [31:54] and companies that are currently overvalued [31:56] are gonna see a decline? [31:57] And is that where some of this short-term volatility [31:59] is gonna come from, [32:00] where there's sort of a reset [32:01] in the valuation of companies? [32:04] Do you think? [32:04] Tom Gardner: Well, I think that's a wonderful question. [32:07] I think I'll say that the more information that we have [32:12] and the more rapidly we can accurately assess [32:16] that information, [32:19] we should see a reduction in volatility. [32:22] There should be, back to your original point, [32:25] there should be fewer bubbles. [32:32] And I mean, maybe a counteracting force is that, [32:37] because it's understandable when somebody says, [32:40] well, all the AI systems, [32:41] they will just all integrate into one [32:44] and there will be no market anymore. [32:46] And again, I'm happy to say I'm speculating [32:50] about what's gonna happen, [32:51] but I don't believe that. [32:53] I think that the systems will always be chasing [32:57] a new insight forever. [33:01] So again, that new insight to go back to, [33:04] I'm more of a college basketball fan [33:06] than an NBA basketball fan, [33:07] but I'll just say, [33:08] I don't know at what age recruiting really begins. [33:10] I'm pretty sure there's some rules about it. [33:12] I myself don't, but I know that there are these, [33:14] the summer camp basketball leagues is like, [33:16] look at this 11-year-old. [33:17] And this 11-year-old is now on YouTube [33:21] playing an amazing level of basketball for an 11-year-old. [33:24] And I'm pretty sure a college coach [33:25] is talking to them at some point. [33:27] And I don't know what the path is there, [33:29] but I certainly know that the NBA would be, [33:32] there were periods of time [33:33] where you could not go public, [33:35] you could not join, [33:37] you could not go pro, excuse me, [33:38] until you'd had at least a year, [33:40] at least a couple, [33:40] you had to graduate, all these. [33:42] And then it shrank because the talent pool is so high. [33:46] And then you could recruit players from Lithuania. [33:49] It wasn't just, [33:50] it wasn't local recruiting by a college of incoming freshmen. [33:54] It was like, hey, [33:55] how early can we find these players anywhere in the world [33:58] and bring them in my program [34:00] so we can continue to be a top 25 college basketball team? [34:03] Okay, setting basketball aside, [34:05] similar dynamic in investing. [34:07] How early can I find the advantages that this business has [34:12] and determine how long those advantages will persist? [34:15] And the only other thing I'll say about this [34:17] is one potentially disappointing development [34:21] that we're already seeing happening, [34:23] but the Motley Fool's on the side of the disappointment [34:25] is companies are staying private longer [34:28] because they need to raise less capital. [34:31] They can utilize more technologies. [34:33] A small team, [34:34] I think Instagram was bought out for a billion dollars [34:36] with 13 employees. [34:38] And we've already said, [34:39] and we've already seen, [34:39] you're gonna see six person companies [34:41] bought for $10 billion. [34:43] And obviously what that does to wealth concentration, [34:46] also all of that capital is moving around [34:48] in the private markets [34:49] that individual investors don't have access to. [34:52] So the dynamics of the markets are gonna change. [34:54] I do think by the time you're public [34:56] and you have to distribute public information, [34:59] is it gonna be two times a year or four times a year? [35:01] We know that that's up for decision here in the near future. [35:07] But I think that we should see [35:11] a greater level of market stability. [35:13] However, I'll say it again, [35:16] humans and the models, [35:18] and then the models advancing AI, [35:22] doing AI R&D and teaching itself [35:26] will always be in the search for new information [35:29] and a better way to analyze that information. [35:31] It will happen until the end of time, I believe. [35:34] And that will give advantages to people [35:35] that are able to get the system [35:37] that finds the crucial bit of information [35:40] as early as possible. [35:42] And that delivers the benefit of finding cancer [35:47] very, very early and treating it. [35:48] And how early can we find it? [35:51] And so that race is on and will continue to be on. [35:53] So I think that there will still always be [35:56] some volatility in the market [35:57] because there are unknown factors [35:58] that the systems are trying to identify and assess. [36:03] Melissa Schilling: Okay, so thank you. [36:05] That was amazing, my head is spinning. [36:08] But have there been any hard parts? [36:10] Any things you wish you'd done differently [36:12] or lessons you've learned? [36:14] What have been some of the challenges [36:15] and how did you navigate them? [36:17] Or challenges you would warn other managers about [36:19] as you bring more AI into your organization? [36:23] Tom Gardner: Well, I mentioned that the large technology companies [36:25] are sort of the canary in the coal mine [36:27] for what's to come. [36:29] And one of the factors that we've seen [36:31] is that the companies with the largest amount [36:34] of cash flow per year [36:35] and the strongest balance sheets in human history, [36:39] the most powerful, most dominant, [36:41] there's no period in time where you can go back [36:43] and say monopolistically, this handful of companies [36:47] was beyond what we're looking at today [36:49] with let's say the top 10 technology businesses. [36:53] And what you're seeing across those companies [36:55] has been a reduction in the workforce. [36:58] So you wouldn't expect to see, [37:00] I mean, layoffs 15 years ago, [37:02] you wouldn't be moving forward [37:03] because reputationally it would signal, [37:06] let's say to institutional investors, [37:07] there might be something wrong inside the organization. [37:09] Now, they're seeing something, [37:11] management is making these moves [37:12] because they see problems ahead. [37:14] But instead the layoffs are happening opportunistically [37:17] because the large tech companies have the tools [37:21] before everyone else, they're using them at scale [37:23] and they're identifying, [37:25] we don't need a marketing department this large. [37:27] It's a worthwhile short article to dig around [37:30] and find of how small Anthropx marketing team was [37:34] en route to creating essentially a trillion dollar company [37:37] in less than six years. [37:38] I mean, all those words in that sentence should jock us. [37:44] But so that's the first one is, [37:46] what's the proper size of the workforce [37:47] for the business that you're in? [37:48] And financial services generally number two [37:51] after IT for disruption. [37:52] So you would then look to the Motley Fool [37:54] and you would say, wow, do you have too many people? [37:57] And do you have the right mix of talent [37:58] for this new toolkit? [38:00] And again, Melissa, you came in [38:01] when we were migrating to AWS [38:03] and gave us some indications, [38:04] hey, it's gonna be hard [38:05] and you should have been upscaling earlier. [38:07] That was a mistake you made. [38:08] I do remember you saying that. [38:09] I thank you for saying that and I agree with it. [38:11] But against that now, [38:12] unfortunately you're at a time period [38:13] where you can't go slow [38:15] because if you have a security breach getting into AWS [38:18] because you're about your internal culture first, [38:21] you're gonna lose a lot of member personal information [38:26] because you didn't advance your security team [38:28] rapidly enough to be ready through the transition [38:31] and to properly run your organization on AWS. [38:35] So that was kind of a crucial moment there [38:38] when we looked at our security team [38:40] and said, we're going to AWS. [38:42] We love our team, but we have to make changes here [38:45] and it's not gonna be able to happen through upscaling. [38:48] So from that, I became CEO of our company in 2007. [38:53] Again, we were founded in 1993. [38:54] We had significant layoffs to save our company in 2001. [38:59] We had raised a lot of capital. [39:00] We weren't a company throwing a lot of parties. [39:02] There wasn't a lot of reckless behavior. [39:03] We just didn't have a business model that worked. [39:05] I won't bore us with what that was, [39:06] but overall we realized, my gosh, we've got it all wrong. [39:09] And now to save our company, we have to do this. [39:11] That was a heartbreaking time. [39:13] One of the most, it was a beautiful heartbreaking time [39:14] because out of everyone who lost their job, [39:16] only one person got upset. [39:17] And about an hour and a half later [39:18] called me on my home phone and said, [39:20] I'm so sorry I reacted that way. [39:22] And I was like, this is a devastating period for us. [39:25] So from that, we concluded, we're not gonna do layoffs. [39:28] We don't wanna be a layoff company. [39:30] And we studied a lot too [39:31] because that's a strong statement to make. [39:33] And we got a lot of advice not to say that. [39:36] But we basically made the statement [39:37] because a layoff to me means surprise, [39:40] you're doing a great job. [39:41] I'm taking it to the extreme scenario [39:42] where somebody loses their job as a layoff. [39:44] I'm not saying this is what all layoffs are about, [39:46] but surprise, you're doing a great job. [39:49] The project you're working on is no longer important to us. [39:52] Your skills are not transferable in the organization. [39:55] We're laying you off. [39:56] So we decided that once that happens to one person, [39:59] then it spreads in the organization. [40:00] Everyone's like, hey, it's gonna be kind of random, [40:02] you know, based upon financial metrics [40:04] or whatever the factors are. [40:05] It doesn't matter how hard you're trying. [40:06] It doesn't matter how much you care. [40:07] It doesn't matter what value you're able to create. [40:09] It's just like, you're on a project that's closed. [40:11] So we made a decision that we're private. [40:14] We bought back all our VCs. [40:15] We have a lot of ability to kind of buy time [40:18] and transitions a little bit. [40:20] And so we're a no layoff company. [40:22] And that meant, but you know, we've got, [40:25] at that point we had 640 employees [40:27] where like these tools are coming fast. [40:30] The capabilities of using these tools. [40:31] Also, understandably, there's resistance. [40:35] We all know, perhaps we all know, [40:36] there are a lot of young people who are like, [40:37] I hate what is happening right now [40:39] and I don't even wanna use the tools. [40:41] And we'll get a lot of throwback developments in society. [40:43] And I celebrate that. [40:45] I celebrate vinyl for people who wanna go for vinyl. [40:47] I want a varied market. [40:48] I don't want everyone to feel like [40:49] they have to force march in one direction. [40:51] But when you're allocating capital in a business [40:53] and you have members that are relying [40:55] on your investment advice and financial services, [40:57] demands the best information analyzed [40:59] as quickly as possible, as thoughtfully as possible. [41:01] All of a sudden that pressure [41:03] of what it means to your member become, [41:05] again, the priority of like, [41:06] don't have a breach as you transition into AWS. [41:10] Don't miss an analysis of this company [41:13] because you didn't use the most recent upgrade. [41:18] You know, you're not using Opus 5.5, right? [41:20] When it comes out, but if you were, [41:22] you would have been able to find [41:23] that company's a little bit more at risk. [41:24] You would have been able to get that guidance out to members, [41:26] but you weren't using the most advanced tooling [41:28] that you could. [41:29] So for our mission, it is, [41:31] we're gonna use every tool we possibly can [41:32] to help our members succeed. [41:33] We have to be the world leader in AI [41:35] for investors competitively and also to serve our members. [41:38] And so a big thing for us has been exit offers [41:41] and buyout offers. [41:43] And you know, anytime you're gonna talk about this, [41:44] there are various legal parameters [41:46] and not wanting to misfire or say something wrong, [41:47] but I'll just say that we have basically consistently said, [41:51] this is the new set of requirements. [41:55] And I would say maybe one step down from requirement, [41:59] these are the new standards. [42:00] This is what we have to work towards. [42:02] If you don't wanna go for that, here's our offer. [42:05] If you do wanna go for it, let's do it, [42:08] but it is now a performance factor. [42:11] These new tools are. [42:12] So that dynamic has worked well. [42:14] I'm not saying everyone has been excited about [42:16] at every stage that we've gone through, [42:18] but that has been the hardest thing for us [42:21] because we have long tenured fools that we love. [42:25] And we also have the pressure of getting the guidance right [42:28] for our members. [42:29] And those two things have collided in the last three years [42:32] more than any time in our company history by far. [42:35] And so that's been. [42:37] And you said, you know, [42:38] where do you think you made a mistake? [42:40] I think that the mistake is, it's gonna sound Orwellian. [42:46] Orwellian. [42:48] The mistake is not to have been firmer earlier on [42:52] because actually if somebody gets the wake-up call in 2023 [42:56] that you are seeing and you're talking to, [42:59] everything's being confirmed for you in your conversations [43:01] with people who have made this the focus [43:04] of their professional life and you're seeing it, [43:08] but you think that the right thing to do is to be gradual [43:12] rather than more demanding and firm right away [43:16] because we probably disadvantaged some people [43:19] in the market for employment [43:24] by not getting an earlier wake-up call [43:26] that's more substantial to realize. [43:27] Like there should be almost, [43:29] there should be no feeling of vindication [43:32] or justification for decisions we made [43:34] because look how it's played out. [43:36] It's like, damn, look what is happening right now. [43:39] And people are, end of civilization, [43:41] all these conversations, [43:42] we were talking and thinking about this in 2023, [43:44] but we acted in a more staged way. [43:47] And I think that some people who exited in 2024 [43:51] have gotten the wake-up. [43:52] It was a surprise to me. [43:54] I know I'm going long with no punctuation in this answer. [43:56] I apologize, but it's an emotional thing for me. [43:58] It was a surprise to me to have somebody [44:00] who was relatively resistant to the change, [44:02] then move and have a stumbling time for let's say [44:06] three months trying to find a job or six months. [44:08] And then they went to work [44:10] for the most advanced AI practitioner in the field. [44:13] And you were like, oh my God, they woke up. [44:16] They got it. [44:16] But now we need to translate that for everyone internally. [44:20] We can't have this be a cozy place where you feel like, [44:27] you can kind of relaxedly improve your understanding of AI [44:31] because that's not what's happening in the market. [44:33] And the last thing for all these companies is, [44:35] of course, the newcomers that will be AI, AGI native, [44:38] let's say, and they're hiring 11 people [44:42] and replicating everything that a mid-cap software company [44:46] has from an IP standpoint with just 11 people, [44:49] what are you going to do with your 1800 person workforce? [44:51] So I think all the workforce issues and questions [44:54] and trying to make sure that you're serving your customer [44:57] and that you're strengthening your business [45:00] and that you're a true partner to the people [45:03] that work at your company and that have left your company. [45:06] Melissa Schilling: Wow. [45:07] So I think you're saying move faster than you think. [45:10] Don't move slow at all. [45:12] And that you're actually doing your workforce a favor [45:14] if you move earlier versus later. [45:16] Is that what you're saying? [45:17] In kind of at a high level? [45:19] Tom Gardner: That is what I'm saying at a high level, [45:21] particularly looking back. [45:23] So, but I assume it actually just applies [45:25] in the same way today. [45:27] And it's a bewildering time for employment. [45:31] And I think it's going to become more so. [45:34] And most of the people that I've talked to, [45:36] whether they're at our company or working technically [45:39] at other companies, do have question marks [45:42] moving around their head about what level, [45:44] what sort of dislocation will happen in employment. [45:46] We're all asking and thinking about these questions, [45:48] but there's a lot of uncertainty. [45:53] And, but the biggest threat would be [45:59] that Claude decides to category kill your offerings [46:04] or that a startup with 12 expert employees [46:10] or AI engineers and 12 deep domain experts [46:14] have just chosen, let's do this with 24 people. [46:16] What somebody else is doing with 1500 people, [46:18] let's raise a hundred million dollars. [46:20] And we can basically replicate [46:21] their intelligence model entirely [46:24] and then see how well we can do on the marketing, [46:26] which Claude proved you can do with a small team. [46:29] So we're going to see a lot of competitive momentum [46:36] from the private markets. [46:39] Melissa Schilling: Okay, we have a bunch of questions. [46:40] We'll try to get to a couple before we run out of time. [46:44] The first one I have here is, [46:46] how do leaders like you, Tom, [46:48] think regulators and lawyers can be brought in [46:50] as informed partners early [46:52] rather than reacting out of fear after the fact? [47:00] Tom Gardner: Well, I don't know that I'll provide the best answer [47:03] and I want to be giving a succinct answer to these [47:06] so we can get through all of them. [47:09] I would say that I've mentioned [47:13] our sort of Motley Fool model of full-time [47:17] and then contract and anywhere in the world. [47:20] And I talked about contract, [47:22] but I would say another layer for us [47:23] is just straight up advisor. [47:26] And it turns out, it depends on your domain, [47:30] your company, the business you're in, et cetera. [47:33] So I can only speak for our business. [47:34] It turns out there are a lot of people [47:36] that are willing to get on a call with us for two hours. [47:39] A lot of people, a lot of executives, [47:42] retired executives, people in technology departments. [47:44] I think there's some curiosity about us [47:45] always because of our brand [47:47] and our 30-year history and other factors. [47:49] But I would say I would be looking to bring [47:53] legal guidance and regulatory guidance in as an advisor [47:56] as quickly as possible [47:57] and taking a discrete challenge and laying it out [48:01] rather than thinking one fell swoop of transformation [48:05] in that area is the way to go. [48:06] I don't know if that helps, [48:07] but that's how we're moving forward. [48:09] Melissa Schilling: Definitely, and I love how concrete that is. [48:12] The next question is Enron was a case of fraud [48:15] or hiding the issues, [48:16] which was done with collaboration with their auditors. [48:19] Would AI be able to warn the investors [48:22] of potential disasters if the data provided [48:24] were limited and not fully disclosed? [48:27] So I think they're asking would AI have been able [48:29] to penetrate that kind of intentional obfuscation [48:32] that might've happened at Enron [48:34] and detected the fraud earlier? [48:38] Tom Gardner: I would say if that question is about [48:44] whether 100% of the time it can hit it perfectly, [48:47] then we'll both agree immediately, no chance. [48:50] Does it move us faster to finding details and patterns, [48:55] even patterns across auditors? [48:57] I remember reading a wonderful essay about a money manager, [49:00] it's to be like 25 years ago, [49:02] who spent a tremendous amount of their time [49:04] just following board networks. [49:06] Who's joining what board? [49:07] And that you can find a lot of patterns. [49:10] And once you find, [49:11] since we're talking about the negative side of this, [49:12] because really the real joy of it [49:14] is to find the patterns of directors [49:15] and institutional investors, let's say, [49:17] that are surrounding particular companies [49:19] that are set up to win. [49:20] And that's how we're gonna make our money. [49:22] We're not gonna be able to make that much money [49:24] generally as investors by finding the big losers, [49:28] because most of us aren't gonna short, [49:30] and even then you have 100% max gain, right? [49:32] So really the tools are like, [49:33] how do we find the opposite of Enron as early as possible? [49:36] But in the case of Enron-like businesses, [49:38] I think it will be easier to find certain patterns. [49:42] As I said, board directors that are moving, [49:44] people have moved from one low-grade [49:47] promotional business to the next. [49:48] I mean, the simplest place to do it [49:50] would be following some of the movement of talent [49:53] around small penny stock companies. [49:56] Again, it's not gonna be a very productive use of time. [49:58] Nobody can short those companies, [49:58] but you better believe that there are patterns [50:01] of people moving through companies [50:02] capitalized under $100 million, [50:03] where they're finding ways just to extract salary [50:06] and wages from shell businesses [50:09] that have no operating future whatsoever. [50:11] So I do think it'll be easier and easier to get there. [50:15] And the one thing I would say to the questioner is, [50:18] to Vincent is, it actually might come a little slower [50:22] because of the reason that I shared. [50:23] You can't really make money by finding the negatives. [50:26] So most people are out searching for the big wins. [50:28] That is how bubbles can form. [50:30] But I do think for teams that wanna point the tools [50:34] towards problems, you're gonna find them earlier. [50:37] Melissa Schilling: Wow, cool. [50:39] That's neat. [50:40] The next question I see here is, [50:42] do you worry about an over-reliance [50:44] on AI models for insights? [50:46] And do you have any security concerns [50:48] on its use in your industry? [50:51] Tom Gardner: I don't have many worries about over-reliance [50:54] on AI models for insights. [50:56] I think that, I mean, that's too simple of an answer. [51:01] I wouldn't get on an airplane flown by GPT-3, right? [51:06] So if we're gonna put our organization, [51:09] our analysis, our decision-making [51:10] all in the hands of a system of intelligence [51:13] with no human checks, I'm not interested in that at all. [51:17] Pretty much, I'm not interested in almost any field. [51:19] But obviously, we'll start to see that change [51:21] as we already are with self-driving Waymo, [51:25] tens and tens of millions of miles automated. [51:28] We're putting it all in the hands of a system, [51:31] our lives in the backseat of the car. [51:33] So we're obviously entertaining this in different areas, [51:38] but to get to that level of scale, [51:40] you have to have so many checks in the system. [51:43] So I think I would say that for us right now, [51:46] I like the balance that we have, [51:48] but I do believe that we need to advance more [51:50] with AI models to provide more insights at scale [51:53] and greater depth, not less. [51:57] And then do I have security concerns for use in our, [52:00] I mean, I guess if I were to be succinct, [52:02] I would say I have security concerns [52:04] about every aspect of AI advancement, [52:05] just because right now, [52:08] the cybersecurity public companies [52:10] are seeing a significant rise in their valuation [52:14] because Mythos and the Glasswing project [52:17] and sort of getting in to see behind the curtain [52:20] of what's happening and starting to power up [52:22] U.S. cybersecurity companies with the goods [52:25] of what's really being developed in these models [52:26] has strengthened their case. [52:28] But if you actually scratch one level [52:30] below the surface of that, [52:33] these cybersecurity companies, if they have a breach, [52:36] it could be very, I mean, CrowdStrike got through, [52:40] it's solved for upgrade problems. [52:42] So, but I think there's a lot of uncertainty about security [52:44] and I do have concerns about it. [52:48] And I'm also not a cybersecurity expert. [52:49] So I'm just out there as a layperson CEO reading [52:52] and talking to our technical teams. [52:53] And I would say that all of us see demand, [52:56] there will be dramatic and problems will be significant. [53:00] Melissa Schilling: Wow. [53:01] Okay, so I'm gonna exert some moderator privilege [53:03] and ask a question that I have. [53:05] And feel free to pass on this one [53:07] if this isn't something you've been following. [53:09] But I read an article in the New York Times the other day [53:11] where somebody speculated that there could actually be [53:16] the funding of the AI data centers and their leases [53:19] could be creating a situation that is analogous [53:23] to what happened with mortgages [53:26] during the 2008, 2009 debacle that we had in that. [53:31] And that if the leases, if the revenues [53:32] or the leases collapse, [53:34] that we could have a really wide crash. [53:38] Is this anything you're following? [53:40] What are your thoughts on that? [53:44] Tom Gardner: Well, I'd say the first thing I look at [53:47] is what is the major long-term trend? [53:50] Because that, and I'm not saying that's easy [53:54] to make final determinations on, [53:56] but at this stage, I would say [53:57] we probably are all looking and thinking AI, [53:59] when most people are looking at AI is much more powerful, [54:01] much more quickly than I thought it was gonna be. [54:04] Right, for the most expert mind [54:05] to even relative newcomers to the tools. [54:08] Once you use them, of course, [54:09] there's good reasons to be skeptical, [54:11] there are misfires, but in the very beginning [54:13] of chat, GPT was making an error, [54:15] let's say every 10 seconds. [54:16] Now it's making an error, [54:17] I'm gonna exaggerate every 10 days. [54:18] It's not that far out, [54:19] but like it's maybe every 24 hours. [54:22] So we're seeing the significant improvements [54:25] of these models very rapidly, [54:27] which would suggest that they're gonna transform [54:29] and they're gonna be monetized in dramatic ways. [54:33] But what's the timing of that monetization? [54:35] So every pause or hiccup on that side, [54:38] I think stocks will move relatively wildly, [54:41] but then I do think the market will become wise [54:43] to the fact that like that's a speed bump, [54:46] that's not a dead end. [54:48] And so the speed bump maybe shouldn't take the market [54:50] down 29%, maybe it should only take it down 11%. [54:53] And then it'll start to even smooth out more [54:55] and you'll start to realize, [54:56] wow, every single successful company in the world [54:58] effectively is connected to the internet [55:00] and using it in some way. [55:01] But we used to call them internet stocks, [55:03] 28 years ago, but like everything is, [55:05] we were saying back there, [55:06] everything's gonna be an internet company. [55:08] It's not even gonna be a distinction that matters. [55:10] Same thing will happen here. [55:11] It'll just happen more quickly [55:12] because the infrastructure's build out [55:14] is being funded profitably in many cases. [55:20] Now, last point, sorry, [55:21] is that the use of debt that you start seeing, [55:24] okay, there's gonna be more scrutiny on these spends [55:29] because there's a lender now. [55:30] It's not just coming out of the cash flows, [55:32] there's a lender and customers, [55:34] you're gonna find situations where you're seeing [55:36] revenue quality diminishes [55:38] because the seller is also providing stock warrants [55:43] or some other incentive for you to buy from them [55:46] to move into their solution set of chips, let's say. [55:51] And also you're getting stock warrants [55:52] to also be a customer. [55:54] That's not a pure dollar. [55:55] A dollar of revenue growth without incentives [55:58] is not a dollar of revenue growth with incentives, right? [56:01] The latter is weaker. [56:02] So we're definitely starting to see it weaken. [56:04] I don't think it's in collapse mode. [56:05] I think there will be speed bumps though [56:07] and the markets will probably overreact to them, [56:10] but I think the long-term trend [56:11] is just inevitable dramatic monetization, [56:15] very high operating gross [56:18] and operating margins at businesses, [56:20] higher rates of return on invested capital, [56:22] and a lot of valuation gets attached to capital efficiency [56:24] when we think it's about revenue growth. [56:26] So I think for the most part, it'll be overrated, [56:28] but that doesn't mean the market couldn't go down 25% [56:32] if there's a CapEx shift [56:36] for any number of geopolitical reasons. [56:39] There's all of a sudden a CapEx shift [56:40] that could hit the market pretty hard. [56:42] Melissa Schilling: Yeah, I mean, during the internet bubble [56:45] and then we knew that there were gonna be [56:47] all these new internet commerce companies [56:49] and that overall as a whole on average, [56:52] there was gonna be a ton of growth, [56:53] but there were a lot of players individually that shook out. [56:55] And I think like Cosmo.com or Grocery Van [56:58] or something like that. [57:00] I'm sure, are you anticipating the same thing here too? [57:02] Like we're seeing an explosion of AI-based companies, [57:05] but presumably a bunch of those at some point [57:07] are gonna start to shake out, right? [57:08] Tom Gardner: Yes, but the only thing I would say is [57:11] I think ultimately the big shock [57:13] is how devastated the incumbents are. [57:16] So I think we see like a lot of little mini boats [57:18] go out there, right? [57:19] But we're not like big battleships dropping [57:21] because of the U-boat, whatever. [57:23] I'll try to find a better analogy, [57:24] but I think that's gonna be the real shock. [57:27] I remember having a member event [57:29] where a member came up to us, [57:31] this would be maybe 1998, [57:33] and said, you know, one thing's for sure, [57:35] like I'm gonna keep holding Kmart as my largest position [57:38] because that's an American brand that's not gonna, [57:40] you know, I still remember that. [57:41] There are a lot of cinematic moments [57:42] looking back 33 years in your life. [57:44] But so I think the big shock is gonna be [57:46] the incumbents that didn't transform. [57:50] And essentially, one of the things I love [57:52] that Toby Lucas said at Shopify is, [57:55] they probably still do this, [57:55] but they used to do things where you have to, [57:57] like every for two weeks, [57:59] you're gonna have to, [57:59] everyone's gonna have to stand in meetings [58:01] or for these two weeks, [58:02] you got to move your mouse [58:03] to the other side of your keyboard [58:05] or whatever it was, [58:06] we're gonna shake up the habit that you're in. [58:09] So the biggest thing for a company [58:11] is just right down to the level of detail [58:13] on how workflows change. [58:15] Because the workflows did not change [58:17] at a lot of magazines and a lot of newspapers [58:19] and a lot of cable TV channels. [58:21] They were resistant to the workflow change. [58:23] Why am I, why would we get the kids to run the website [58:25] and just pay them something? [58:26] But they're not gonna be my boss. [58:27] I'm sure you have that. [58:29] They don't have any idea. [58:30] I have 20 years of history at this organization, [58:32] but it's the ability to change workflows at scale [58:36] in detailed and significant ways [58:39] that's gonna make the big difference. [58:40] So I actually think there will be some fun stories [58:43] about drcoop.com or pets.com going under, [58:47] but actually the big long-term stories will be like, [58:49] damn, there was something like Blockbuster [58:51] had 10s of thousands of employees, [58:53] and it just got, [58:55] so those will be the big long-term tales from this time. [58:59] Melissa Schilling: Barnes and Noble, Blockbuster, those guys. [59:02] Tom Gardner: Yeah. [59:04] Melissa Schilling: Wow. [59:05] All right, well, we are basically out of time, [59:08] but I wanna thank you so much for this. [59:10] I have so many notes here. [59:12] It was like drinking from the fire hose talking with you. [59:15] So I just wanna say thank you so much for joining us, [59:18] and we would love to have you back sometime. [59:21] This was really terrific. [59:23] And I also wanna remind everybody that we have, [59:26] our next series event is going to be on October 22nd [59:32] when we talk to Yin Afeyana Fongs, [59:35] who's the head of Applied AI at NYU Langone Health, [59:38] and they're gonna talk about [59:39] how they're applying AI to healthcare. [59:41] So that's gonna be very exciting too. [59:43] I hope you'll join us for that. [59:45] And I hope you'll come back, Tom. [59:47] That would be really terrific. [59:49] Tom Gardner: Absolutely. [59:49] Melissa, can I do something [59:50] that's not advised in any public speaking setting? [59:55] I wanna take one last question that I see in the queue [59:57] and just give a very rapid answer, [59:59] because I love it, and I think it's, [60:00] Carlos asks, [60:01] I like the idea that we can cover [60:02] 200 companies easier nowadays, [60:04] but what strategies or prompts or methods [60:06] or methods would you recommend [60:08] to actually do this successfully? [60:10] So I will just say, here's a fun one. [60:14] Take 10 insurance companies [60:16] and ask whichever frontier model you're using [60:21] to use the body of work [60:23] of Warren Buffett's professional life [60:25] to assess those insurance companies [60:27] and feed the financial filings of each into it [60:29] and just have Buffett as your agent, [60:33] the world's foremost expert on insurance ever [60:36] in human history, [60:37] and look at through 10 tiny insurers [60:40] and score them one to 100 [60:42] and ask it to focus on current data only. [60:46] So you're gonna have to keep working [60:47] around the angles of it, [60:48] but I think you'll start to see, [60:50] wow, I just learned a lot about 10 insurance companies [60:53] I'd never heard of. [60:53] And even particularly, [60:55] if you don't know insurance, Carlos, [60:56] you're like, wow, I didn't even know [60:57] that's how to think about insurance companies. [60:59] So that would be a way [61:00] is just to get your hands in the dirt [61:01] and start looking at groups of 10 companies [61:04] in categories you know nothing about [61:06] or that you're very, [61:06] usually best to start with the ones [61:08] you're most interested in, [61:09] see how much more you can learn. [61:11] Melissa Schilling: Yeah, and I'm just gonna remind everybody [61:12] that The Motley Fool is doing thousands [61:15] of different permutations of this kind of thing [61:18] about companies you never heard of. [61:19] So you could also subscribe to their newsletter [61:21] and learn about stuff there, [61:23] if that's fair to say. [61:25] Tom Gardner: Absolutely, Melissa, thank you. [61:26] Melissa Schilling: All right, thank you again, Tom. [61:28] It was so great to have you. [61:30] And thanks to everybody who joined. [61:33] We'll send out the web link [61:34] whenever they get the recording, [61:36] if you wanna share it with people [61:37] when they get that edited [61:38] with the captions and everything. [61:40] And that's it for today. [61:41] So thank you so much. [61:47] Bye. (gentle music)