The Acquirers Podcast

Nvidia CEO Jensen Huang removed amid governance concerns; speakers debate AI cycle duration, valuation, and competitive threats.

46 min Listen to the original Episode page
NvidiaJensen HuangAI CycleSemiconductorsValuationEnterprise AIAdobeMicrosoft

Summary

Tobias Carlisle and Tim Travis discuss Nvidia's leadership change, with Carlisle reporting Jensen Huang's removal due to personal conduct issues. Travis highlights risks from AMD and Google's TPUs eroding Nvidia's market share, while Carlisle argues AI adoption is transformative and distinct from past bubbles. The speakers analyze hyperscaler cash flows, the potential for AI to commoditize intelligence costs, and the enduring moats of companies like Adobe and Microsoft.

Episode sections

  1. 0:00 – 1:51

    Sponsor Reads

    The episode begins with advertisements for IG Private Wealth, BetMGM Casino, and the Value Options Letter subscription service.

  2. 1:51 – 3:35

    Nvidia Leadership Change and Valuation

    Tobias Carlisle reports that Nvidia CEO Jensen Huang has been removed from his position due to personal conduct issues, coinciding with blowout earnings. Tim Travis expresses concern over the company's reliance on a single leader despite its massive market capitalization.

  3. 3:35 – 4:26

    AI Cycle Duration and Competitive Threats

    Travis discusses the uncertainty of the AI chip cycle's duration, with bulls projecting demand through 2030-2035. He highlights growing competition from AMD and Google's TPUs, which may erode Nvidia's market share and gross margins.

  4. 4:26 – 6:27

    Hyperscaler Cash Flow and Semiconductor Cyclicality

    Carlisle notes that hyperscalers are cash flow negative while raising capital to buy chips, suggesting long-term investment rather than immediate profitability. He contrasts this with the historical boom-and-bust cycle of semiconductor companies.

  5. 6:27 – 8:30

    AI Adoption and Business Models

    Carlisle argues that AI is a transformative technology with clear use cases, unlike crypto or NFTs. He mentions personal usage of Anthropic's Claude and ChatGPT, noting high spending on compute resources.

  6. 8:30 – 13:30

    LLM Competition and Enterprise Integration

    Travis expresses skepticism about the profitability of standalone LLM providers like Anthropic due to stiff competition. He cites Michael Eisner's view that enterprises will build protective layers above base models, similar to Amazon's platform strategy.

  7. 13:30 – 20:51

    AI in Research and Medicine

    Carlisle shares examples of AI identifying errors in physics papers and assisting in medical diagnosis and tax preparation. He argues AI will act as a tool augmenting human expertise rather than replacing it entirely.

  8. 20:51 – 26:32

    AI Acceleration and Singularity

    Carlisle discusses the METR research group's analysis of AI self-improvement rates, noting that AI task completion speeds are accelerating faster than historical predictions, potentially leading to a singularity event.

  9. 26:32 – 28:25

    Tech Revolutions and Value Migration

    Tobias Carlisle discusses historical tech revolutions, citing Carlotta Perez and William Nordhaus, noting that value initially flows to infrastructure (chips, power) but eventually migrates to the consumer layer where most value is captured. He uses the historical cost of light as an analogy for how AI costs will eventually drop to the level of electricity and infrastructure.

  10. 28:25 – 29:25

    Investing in AI: Utility vs. Profitability

    The speakers agree that while AI is impactful, investing requires traditional value principles like competitive advantages and margin of safety. Carlisle expresses skepticism about exponential growth rates in valuations and notes NVIDIA's extraordinary asset returns.

  11. 29:25 – 33:11

    NVIDIA Leadership and Governance

    Speaker_0 questions whether NVIDIA should have fired CEO Jensen Huang, arguing he was essential for creating the 'neo cloud' ecosystem and maintaining relationships with TSMC. Carlisle counters that boards must prioritize risk and compliance, making the decision to remove him inevitable despite the business impact.

  12. 33:11 – 35:58

    Semiconductor Supply Chain Dynamics

    Carlisle cites Elon Musk's view that NVIDIA and Apple are design firms that send digital files to manufacturers like TSMC and ASML. Speaker_0 highlights TSMC's dominant position and lack of sophisticated competition, noting that Intel's attempt to enter the foundry business has been a financial failure.

  13. 35:58 – 41:25

    Microsoft's Longevity and AI Disruption

    Carlisle observes that Microsoft is the only top-10 company from 2000, 2010, and 2020, surviving through cloud and Copilot. They discuss how AI tools like Claude may replace traditional skills like Excel proficiency, potentially disrupting established software suites.

  14. 41:25 – 46:18

    Adobe and Competitive Moats

    Carlisle considers initiating a position in Adobe due to its valuation but remains nervous about AI capabilities. Speaker_0 argues Adobe's moat lies in its content library, user training, and convenience, comparing it to dry cleaners that remain useful despite cheaper alternatives.

Assets discussed

AssetKindStanceContext
NVIDIANVDAequitymixedall right, you have Nvidia down a ton. The market's not down that much, but I mean, so many stocks are down 30, 40 percent.
AppleAAPLequitymixedAnd I wrote an article for Seeking Alpha, just kind of basically saying, you know, neutral to negative on Apple. Yeah. And I can get a good feel for things.
AlphabetGOOGLequityneutraland then Broadcom's designing the chips for Google
AmazonAMZNequityneutraland for and for Amazon. Uh and so you know whatever that dynamic is, you have to keep that relationship strong and it's just, I think we alluded it to it like the other a week ago or something. But I is I remember like investing Bruce Greenwald's book. It's a really freaking good book. And I just
MicrosoftMSFTequityneutralthe top 10 businesses from a decade ago, except for Microsoft, and that's we we can talk about how well Microsoft survives the next decade. Only Microsoft has survived from the decade before then as well. So Microsoft was around in 2000, 2010, 2020.

“Stance” is how the speakers talked about the asset, not a recommendation.

Claims worth verifying

Jensen Huang was removed from Nvidia due to having three affairs (one with a spy, two with subordinates).

Tobias Carlisle ·2:03

How to verify: Check Nvidia's official press releases or SEC filings (Form 8-K) for executive leadership changes and reasons for departure between November 2026 and the episode date. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Primary source

Nvidia reported blowout/blockbuster earnings prior to the leadership change.

Tobias Carlisle ·2:30

How to verify: Review Nvidia's Q3 or Q4 2026 earnings report filed with the SEC on EDGAR. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Primary source

Nvidia's market capitalization was approximately $5 trillion a few days before the episode.

Unidentified speaker (speaker_0) ·3:00

How to verify: Check historical market cap data for NVDA on financial data providers (e.g., Yahoo Finance, Bloomberg) for the period 2026-08-21 to 2026-11-21. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Primary source

Bullish analysts project the AI chip demand cycle to last until 2030, 2031, or even 2035.

Unidentified speaker (speaker_0) ·3:40

How to verify: Search for analyst reports or public statements from Nvidia investors or industry analysts dated between 2026-08-21 and 2026-11-21 regarding AI infrastructure forecasts. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Primary source

Hyperscalers are cash flow negative and raising equity and debt to fund AI infrastructure.

Unidentified speaker (UNKNOWN) ·5:01

How to verify: Review cash flow statements and financing activities in the 10-Q/10-K filings of major hyperscalers (Microsoft, Amazon, Google, Meta) for the relevant quarters. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Primary source

Casey Handmer fed 50,269 physics papers into an LLM and found 872 errors, with half persisting for seven years.

Unidentified speaker (UNKNOWN) ·13:42

How to verify: Search for Casey Handmer's public posts or papers on X (Twitter) or academic repositories detailing this meta-analysis. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

METR (meta-metermetr) analyzed internal Anthropic models and found AI task completion speeds are 2.5x faster than the previous year.

Unidentified speaker (speaker_0) ·21:15

How to verify: Locate METR's published research or blog posts discussing their analysis of AI scaling laws and self-improvement rates. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Apple is trading at 35 times earnings with 6% revenue growth over the last few years.

Unidentified speaker (UNKNOWN) ·25:43

How to verify: Check Apple's (AAPL) current P/E ratio and historical revenue growth rates in recent SEC filings or financial news sources. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Primary source

Tim Travis wrote an article for Seeking Alpha stating he is neutral to negative on Apple.

Unidentified speaker (UNKNOWN) ·25:43

How to verify: Search Seeking Alpha for articles authored by Tim Travis discussing Apple's valuation. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Primary source

Tobias Carlisle looked at 200 years of tech revolutions in a white paper, citing Carlotta Perez and William Nordhaus.

Unidentified speaker (UNKNOWN) ·26:32

How to verify: Search for Tobias Carlisle's white paper or blog posts referencing '200 years of tech revolutions' and citations of Carlotta Perez and William Nordhaus. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

In the 1800s, the average wage required five hours of work to afford a candle; today, it takes half a minute to turn on house lights.

Tobias Carlisle ·27:47

How to verify: Verify historical wage data and energy costs via historical economic datasets (e.g., MeasuringWorth) and current utility bills. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Jensen Huang created the 'neo cloud' to create competition for hyperscalers and maintain demand for NVIDIA chips.

Unidentified speaker (UNKNOWN) ·29:34

How to verify: Review NVIDIA earnings call transcripts or press releases from 2026-2026 discussing the 'NVIDIA AI Enterprise' or 'Neo Cloud' strategy. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Primary source

Elon Musk tweeted that Apple and NVIDIA send digital files to manufacturers like TSMC and ASML rather than making chips themselves.

Unidentified speaker (UNKNOWN) ·33:11

How to verify: Search Elon Musk's X (Twitter) feed for posts describing NVIDIA and Apple as design firms outsourcing to TSMC/ASML. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Primary source

Intel's foundry business has been a 'huge money loser' compared to its previous dominance.

Unidentified speaker (speaker_0) ·34:38

How to verify: Check Intel's SEC 10-K filings for operating losses in the Foundry Services segment. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Primary source

Microsoft was the only top-10 business from 2000, 2010, and 2020 that survived into the next decade.

Tobias Carlisle ·35:58

How to verify: Compare Fortune 500 or S&P 500 top companies lists from 2026, 2026, and 2026 to verify Microsoft's presence in all three. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Primary source

Carlisle bought Intel stock when it collapsed five years ago and was down on the position before abandoning it.

Unidentified speaker (UNKNOWN) ·35:34

How to verify: Review Tobias Carlisle's public disclosures or interviews from 2026-2026 discussing his Intel investment. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Carlisle is halfway through his 'meta talk' (likely 'meta analysis' or similar) before an update.

Tobias Carlisle ·44:40

How to verify: Search for Tobias Carlisle's blog posts or podcast episodes titled 'Meta Talk' or similar from late 2026. For recent events, verify 2026-08-21 through 2026-11-21. Use another period only when the speaker explicitly names it.

Do your own research

  • Verify Nvidia's Q3 FY2026 earnings report and press release regarding CEO Jensen Huang's departure and the stated reasons.
  • Check Nvidia's latest 10-K and 10-Q filings for details on leadership structure, direct reports, and governance disclosures.
  • Review AMD and Google's latest quarterly results for data on TPU/AI chip revenue growth and market share gains.
  • Analyze hyperscaler (Microsoft, Meta, Google, Amazon) capital expenditure trends and cash flow statements to assess infrastructure investment sustainability.
  • Read Carlotta Perez's 'Technological Revolutions and Financial Capital' for the historical framework on tech value migration.
  • Examine Adobe's Q3 FY2026 earnings call transcript for management commentary on AI-driven competitive threats and moat durability.
  • Search for METR (Measurements of Textual Reasoning) research group publications on AI self-improvement rates and scaling laws.
  • Check Palantir and ServiceNow stock performance and enterprise contract announcements to evaluate the 'integration layer' thesis.
Full transcriptexpand

0:00Unidentified speaker (speaker_0): Okay. When I sell my business, I want the best tax and investment advice. I want to help my kids, and I want to give back to the community. Ooh, then it's the vacation of a lifetime.

0:12Unidentified speaker (UNKNOWN): I

0:12Unidentified speaker (speaker_0): wonder if my out of office has a forever setter.

0:15Unidentified speaker (UNKNOWN): An

0:15Unidentified speaker (speaker_1): IG private wealth advisor creates the clarity you need with plans that harmonize your business, your family, and your dreams. Get financial advice that puts you at the center. Find your advice. advisor at IGPrivatewealth.com.

0:31Unidentified speaker (speaker_2): Legendary has arrived in Alberta with BetMGM Casino. Sign up to unlock access to over 4,000 casino games, including exclusive slots you can't spin anywhere else. Survivors ready? The iconic TV show is now an online casino game where you can tap into island fun. BetMGM has also brought the prices right. Game of Thrones, Family Feud, MGM Millions, and so many more exclusive casino games to Alberta. Get Las Vegas right from your phone with live dealer games streaming from real tables. And every time you play, you can enjoy instant deposit options and fast withdrawals. BetMGM Casino and its legendary library of games are available. Download the BetMGM MGM app. See BetMGM.com for terms, 18 plus Alberta only. Please play responsibly. subject to eligibility requirements. If

1:17Unidentified speaker (speaker_0): gambling is affecting your mental health or well being, 211 Alberta is here to help. Call or text 211 or visit ab.211.ca. BetMGM operates pursuant to an operating agreement with the Alberta iGaming Corporation. Standard message and data rates may apply.

1:30Unidentified speaker (UNKNOWN): Hi,

1:30Tobias Carlisle: I'm Tobias Carlo. This is the Value Options Letter Podcast. I'm joined as always by my co-host co host, Tim Travis. How are you, Tim?

1:37Unidentified speaker (UNKNOWN): I'm

1:38Tobias Carlisle: doing all right. A little hungover, but doing all right. Battling

1:41Unidentified speaker (UNKNOWN): through the

1:42Tobias Carlisle: day. What's the occasion?

1:43Unidentified speaker (UNKNOWN): Uh

1:44Tobias Carlisle: just hanging out. I had some friends in town, so it was a a fun night out in Scottsdale,

1:48Unidentified speaker (speaker_0): but regretting it uh this morning, that's for sure.

1:51Tobias Carlisle: Nice. Uh big news overnight. Well, there's there's lots of news since we spoke last. Nvidia, like massive headline, I guess, for everything. Two things that have happened.

2:02Unidentified speaker (UNKNOWN): They

2:03Tobias Carlisle: gave Jensen the boot. Uh, it wasn't selling chips to the Chinese, it was having three affairs.

2:09Unidentified speaker (UNKNOWN): Not

2:10Tobias Carlisle: allowed to do that. Is it

2:11Unidentified speaker (speaker_0): three? I thought it was two.

2:12Tobias Carlisle: Well, one with a spy and then two with subordinates.

2:15Unidentified speaker (speaker_0): Oh, okay. Wow. I didn't know about the third.

2:17Unidentified speaker (UNKNOWN): Yeah,

2:17Tobias Carlisle: it was another subordinate. And uh, I don't know which is worse, the spies or are the subordinates, but in any case, uh, he's gone. He's gonna pop up somewhere. else, I guess. Um, but they reported blowout earnings, blockbuster earnings.

2:34Unidentified speaker (speaker_0): Yeah.

2:35Tobias Carlisle: What's your take?

2:37Unidentified speaker (UNKNOWN): I

2:38Unidentified speaker (speaker_0): have a few takes. Um one is

2:41Unidentified speaker (UNKNOWN): you

2:41Unidentified speaker (speaker_0): know, Nvidia, it it's almost amazing how dependent they seem to be on him. Like the way that 60 people report to him directly. and uh, you know, no one, a lot of the people that follow Nvidia really closely seems to have a lot of confidence in the board of directors.

2:59Unidentified speaker (UNKNOWN): Uh,

3:00Unidentified speaker (speaker_0): so it's you know, it's amazing to think of a five trillion dollar company, which was the market cap, you know, a few days ago, that that you know doesn't have that deep of a bench. And so I think that that's pretty worrisome. And I mean, I can honestly understand. why there is the movement to keep him with that, you know, being said, honestly. I mean, it's it's, you know, I it's it's a tough situation. You know, is it a spy? I don't know. I don't I don't know the details. Is it good to, you know, have an affair with your sport units? Of course not. Um, but uh, but yeah, it's a it's a challenging situation. Obviously the stock has gotten a lot more reasonable from a valuation standpoint. And and so I've done a a lot of analysis on it the last couple of weeks. And it's really the question is how how long is this AI cycle on chips? That's one question. Uh so I think the bulls would say, like, mostly probably going out to 2030, 2031 would be like the bullish case, maybe longer. I've even heard 2035.

4:01Unidentified speaker (UNKNOWN): Uh,

4:01Unidentified speaker (speaker_0): then the other thing is. that other competitors are starting to pick up market share, Toby. And

4:10Unidentified speaker (UNKNOWN): you

4:10Unidentified speaker (speaker_0): know, how how how quickly can companies like AMD and Google with the TPUs, how quickly can they encroach on Nvidia's market share and maybe sacrifice some of those gross margins? And so I think those are the keys to watch.

4:26Tobias Carlisle: Yeah, it's been I I've been watching, it's been a little bit of a contest between the hyperscalers on one side and the semis on the other. And the hyperscalers want the chips, so they're prepared to pay any price. And it looks like that's sort of,

4:42Unidentified speaker (UNKNOWN): you

4:42Tobias Carlisle: know, the the best chips are uh selling at a premium, but even old chips now. I I saw this tweet from Musk where he was saying even old

4:50Unidentified speaker (UNKNOWN): chips.

4:51Tobias Carlisle: chips are getting premium rates, the training, which is like five or six generations old, still getting pretty good rates. Um so it seems

5:01Unidentified speaker (UNKNOWN): like

5:01Tobias Carlisle: there's almost this unlimited demand there. But then when I look at where the demand's coming from on the hyperscale side, those guys have gone massively cash flow, negative, raising equity, raising debt,

5:13Unidentified speaker (UNKNOWN): which

5:14Tobias Carlisle: says to me that. that I I think that they're they're probably going to get very good returns. I mean, but you and I are both pretty big believers in AI as a technology. Um, that doesn't mean that you're not going to see this cyclicality, particularly out of the semis, because

5:32Unidentified speaker (UNKNOWN): th

5:32Tobias Carlisle: it's been their historical pattern to have these sort of periods of boom and bust

5:38Unidentified speaker (UNKNOWN): where

5:39Tobias Carlisle: they do earn, they do over earn. And every time they do, everybody points out they trade on low multiples because the entire market knows that they're expected to the market can see that they're over-earning. So the semis are like a lot of commodities, you want to buy them when they're not earning that well and sell them when they're overearning. over earning. But have we now gone into a new world where

6:02Unidentified speaker (UNKNOWN): there's just

6:03Tobias Carlisle: so much demand for compute? And I'm I I I find it

6:08Unidentified speaker (UNKNOWN): a

6:08Tobias Carlisle: little bit hard as a like a pretty traditional value guy.

6:13Unidentified speaker (UNKNOWN): Yeah, I

6:13Tobias Carlisle: don't want to get too carried away with these with these booms.

6:17Unidentified speaker (UNKNOWN): And

6:17Tobias Carlisle: I think that it was easier in dot com one point oh and certainly easier in the meme stock period too. Like I thought NFTs were bullshit. I think crypto's probably bullshit.

6:26Unidentified speaker (speaker_0): Yeah.

6:27Unidentified speaker (UNKNOWN): You

6:27Tobias Carlisle: know, we're 17 years into crypto and there's really no use case. Although maybe

6:32Unidentified speaker (UNKNOWN): something like

6:32Tobias Carlisle: that'll they'll figure

6:33Unidentified speaker (UNKNOWN): something out,

6:33Tobias Carlisle: but not

6:34Unidentified speaker (UNKNOWN): yet. But

6:35Tobias Carlisle: then AI is different. Like I use the tools myself, they're amazing.

6:40Unidentified speaker (UNKNOWN): Life changing.

6:40Unidentified speaker (speaker_0): Literally life changing.

6:42Tobias Carlisle: You can see how every business is gonna just adopt these tools and use them. So then they're gonna going to go through.

6:49Unidentified speaker (UNKNOWN): That

6:49Tobias Carlisle: means anthropic, I think in I don't I don't know about the price. I'm not talking about the price. I don't know about the IPO. I'm just talking about the underlying business. Business, you know, that's going to grow for a very, very long time. Same with ChatGPT. They may compete against each other. But I just noticed my own spending on this stuff. I upgraded to the most to the $200 plan and basically blew through it and I'm spending like a hundred bucks every time I get there because I

7:18Unidentified speaker (UNKNOWN): I

7:19Tobias Carlisle: just need more compute and I I use it all.

7:22Unidentified speaker (UNKNOWN): But

7:22Tobias Carlisle: I do have a favorite.

7:23Unidentified speaker (UNKNOWN): But

7:23Tobias Carlisle: I think that all of that means

7:25Unidentified speaker (UNKNOWN): that the

7:26Tobias Carlisle: investment by the hyperscalers is probably going to pay off.

7:30Unidentified speaker (UNKNOWN): And

7:30Tobias Carlisle: so their equity and debt raises are probably a good idea.

7:34Unidentified speaker (UNKNOWN): The

7:34Tobias Carlisle: fact that they're cash flow negative at the moment is kind of irrelevant. That just means they're investing. But it does mean that it doesn't mean that you can't have that sort of short term overenthusiasm, have a little or a big crash in the interim, and then have a much longer, like 25 year uh period where it's installed. That's the the Carlotta Perez idea that you have the the bubble and the installation and then after that you have the sort of longer period of the plateau where

8:01Unidentified speaker (UNKNOWN): it g

8:02Tobias Carlisle: it just filters out into society. What what do you think?

8:05Unidentified speaker (UNKNOWN): Oh,

8:05Unidentified speaker (speaker_0): I I think you make good points there. Uh and you're right. This is a much more challenging uh paradigm for value investors because you can make the argument that, you know, there's there's that it that it's a value stock. Whereas, you know, the the dot com, it was, you know, the multiples were so outlandish that it was really hard to justify that. Um so

8:27Tobias Carlisle: and profitless too, for the most part.

8:29Unidentified speaker (UNKNOWN): Yeah, yeah, yeah.

8:30Tobias Carlisle: Revenueless.

8:31Unidentified speaker (speaker_0): Right. Right. I think so. I think the LLMs, I don't love that business model. Like, I think that the competition is so stiff, Toby, that,

8:41Unidentified speaker (UNKNOWN): you

8:42Unidentified speaker (speaker_0): know, as as much as we love anthropic and and people like the other ones, I don't think that the difference is going gonna to be that substantial. And

8:52Unidentified speaker (UNKNOWN): can I ask

8:52Tobias Carlisle: which do you use? I

8:53Unidentified speaker (speaker_0): use and I use Claude. Yeah, I use Claude. I all I only really use Claude and I use a little bit of Muse as like personal assist. I don't like it as much, though, nearly as much as Claude. I

9:03Unidentified speaker (UNKNOWN): used I

9:03Tobias Carlisle: just chat a little bit too. It's not as good as Claude, but yeah, I agree with you on Claude. Sorry.

9:07Unidentified speaker (UNKNOWN): Yeah, yeah,

9:08Unidentified speaker (speaker_0): no, no worries. But so

9:09Unidentified speaker (UNKNOWN): I

9:10Unidentified speaker (speaker_0): I watched an interesting I watched Michael Eisman, uh the big short guy. Yeah.

9:14Unidentified speaker (UNKNOWN): Um

9:15Unidentified speaker (speaker_0): he's got a great podcast. And uh he had some guys on. Uh one was Dan Ives, and I don't know who the other gentleman was, but they were talking about how

9:24Unidentified speaker (UNKNOWN): you

9:24Unidentified speaker (speaker_0): know businesses are gonna on brand, they're already doing it, of course, but they're gonna onbrand Claude and Chat GPT and and the LOMs, but they're gonna have a layer above it, which could be like a Palantir. It could be

9:37Unidentified speaker (UNKNOWN): maybe

9:38Unidentified speaker (speaker_0): something like a server. service now, potentially. Um, but uh a layer above it

9:42Unidentified speaker (UNKNOWN): to protect

9:43Unidentified speaker (speaker_0): it. Yeah, because the reality is is even though these LLMs are saying, look, we're not going to use your info against you or to compete against you, because the investment is so substantial, Toby, and the losses are so huge.

9:57Unidentified speaker (UNKNOWN): Yeah, you can't

9:57Tobias Carlisle: lock in.

9:58Unidentified speaker (speaker_0): It's hard to imagine them not doing that, kind of like Amazon. I mean, Amazon's actually a great. uh model for that because they that's what they did. You know, a lot of the products and stuff, they they would compete with them. And now advertising is a bigger part of the business. So I don't know if that happens as much, but that was like a major thing. You put your product on Amazon, and then you know, you get a lot of competition coming up from others and from them. Uh so so but I think with the chips,

10:24Unidentified speaker (UNKNOWN): I

10:25Unidentified speaker (speaker_0): think if it if if you told me Nvidia is gonna keep its market share, then I would honestly be fine owning the stock at the current price, uh, because I am that bullish on the long-term AI spending cycle. But I do think that like the the CUDA, the the software that It was the huge competitive advantage because it's so dominant in the GPU space, uh Um that is easier to, it's it's now it's much easier to train for like the LLMs to train uh people on like the AMD chips or the organizations on the AMD chips. So the switching costs don't seem as high. So I think that that's what creates um a really challenging picture. And, you know, like on the value options layer, Toby, we did a trade

11:14Unidentified speaker (UNKNOWN): uh

11:14Unidentified speaker (speaker_0): yeah. yes yesterday actually. Uh yesterday before earnings. And we sold a 139 put and we collected uh 229 bucks on it. So it's about a 1.7% return on your total cash invested, but in two days.

11:33Unidentified speaker (UNKNOWN): Uh

11:33Unidentified speaker (speaker_0): so the annualize is like 220% or something crazy like that. So so that's why I like having that tool in the toolkit to where you can take advantage of volatility and manufacture a price where, hey, you know, if we end up owning it at this price, like I I feel like there's an adequate margin of safety there. Um so it's kind of a chickens way of doing it, but uh, but it works.

11:56Tobias Carlisle: Okay. When I sell my business, I want the best tax and investment advice. I want to help my kids and I want to get give back to the community. Ooh, then it's the vacation of a lifetime. I wonder if my out of office has a forever setter.

12:12Unidentified speaker (speaker_1): An IG Private Wealth Advisor creates the clarity you need with plans that harmonize your business, your family, and your dreams. Get financial advice that It puts you at the center. Find your advisor at IGPrivatewealth.com.

12:27Unidentified speaker (speaker_2): Legendary has arrived in Alberta with BetMGM Casino. Sign up to unlock access to over 4,000 casino games, including exclusive slots you can't spin anywhere else. Survivors ready? The iconic TV show is now an online casino game where you can tap into island fun. BetMGM has also brought the prices right: Game of Thrones, Family Feud, MGM Millions, and so many more exclusive casino games to Alberta. Get Las Vegas right from your phone with live dealer games streaming from real tables. And every time you play, you can enjoy instant deposit options and fast withdrawals. BetMGM Casino and its legendary library of games are available. Download the BetMGM app. See BetMGM.com for terms. 18 plus Alberta only. Please play responsibly, subject to eligibility requirements. If

13:13Unidentified speaker (speaker_0): gambling is affecting your mental health, or well-being, 211 Alberta is here to help. Call or text 211 or visit ab.211.ca. BetMGM operates pursuant to an operating agreement with the Alberta iGaming Corporation. Standard message and data rates may apply.

13:28Tobias Carlisle: Uh slightly changing subjects.

13:30Unidentified speaker (UNKNOWN): I

13:30Tobias Carlisle: saw there was a good um

13:33Unidentified speaker (UNKNOWN): a

13:33Tobias Carlisle: couple of interesting papers. One, uh

13:37Unidentified speaker (UNKNOWN): I saw

13:38Tobias Carlisle: this. tweet, Casey Handmer is his name.

13:42Unidentified speaker (UNKNOWN): And

13:42Tobias Carlisle: he took fifty thousand two hundred and sixty-nine physics papers

13:47Unidentified speaker (UNKNOWN): and

13:48Tobias Carlisle: fed them into an LLM.

13:50Unidentified speaker (UNKNOWN): And

13:50Tobias Carlisle: he found eight hundred and seventy-two errors. And he said half of the copied errors were still spreading seven years later.

13:58Unidentified speaker (UNKNOWN): And

13:59Tobias Carlisle: so then he shared this um

14:01Unidentified speaker (UNKNOWN): he

14:01Tobias Carlisle: shared this chart. but it's kind of it's kind of amazing. He's asking for a hundred thousand dollars in tokens from one of the big uh LLM providers because he wants to be able to run this through. But basically somebody in the comments was saying, um, would you, you know

14:18Unidentified speaker (UNKNOWN): well

14:19Tobias Carlisle: I wish they would go through and like verify all this data and figure it out. And he's like, it's it's unnecessary now. You just feed all of this all of these papers in. like, and it figures out stuff that's behind paywalls and all that sort of stuff and runs these meta-analyses. Basically, he's very, very bullish on the ability of the LLMs now to pass because he's got this now massive database of every paper that's been published for like 50 years. The LLMs to be able to like do this meta-analysis of all of these things and make some unique breakthroughs

14:48Unidentified speaker (UNKNOWN): uh

14:49Tobias Carlisle: just by reviewing this stuff.

14:51Unidentified speaker (speaker_0): Incredible.

14:52Tobias Carlisle: The use cases for for for AI are just unbelievable.

14:56Unidentified speaker (speaker_0): Another one, like I I know people that are using like AI like with Google Earth to find like ancient civilizations

15:04Unidentified speaker (UNKNOWN): in South America.

15:05Tobias Carlisle: It's great.

15:07Unidentified speaker (speaker_0): It's it's it's wild what you can do. And and I mean I think like another like in the medical field, like I I mean like my brother uh is having health issues and and I mean I would love to

15:20Unidentified speaker (UNKNOWN): I

15:20Unidentified speaker (speaker_0): mean obviously there's HIPAA and everything but I would love to have like you know all of the information that the doctors have and plug it into AI and and see what they can come up with because they haven't really been able to resolve Solve some of the issues. Uh, I think that like in medicine in in new developing new drugs, it's incredible. And just from a practical standpoint, I mean, like I have a tax guy. My CPA was doing my taxes, and I plugged in his calculations to uh AI, and like literally it was 60 to 80,000. I don't know what the difference is. It saved me that much money. And then I like sent it back to my tax guy, and there was like no argument. It was just, oh, it was like completely right, you know? it's just,

16:03Unidentified speaker (UNKNOWN): I

16:03Unidentified speaker (speaker_0): mean, it's I it's it's gonna hurt a ton of people, it's gonna benefit a ton of people. It's just like this, you know, phenomenon. Uh, so it's it's hard to tell. Like you said, I mean, look, like, like on your on your new website, like the last investment, it's it's it's a totally different paradigm, and you just have to do the best you can with the limited information and the multiple possibilities that can occur.

16:29Unidentified speaker (UNKNOWN): Yeah,

16:30Tobias Carlisle: I read this white paper, uh, and when I say I wrote it, I wrote it with Claude. I disclosed that at the bottom. Um,

16:35Unidentified speaker (UNKNOWN): which

16:35Tobias Carlisle: means that basically I found a whole lot of stuff that I find interesting. Like I had this idea that I think there are two things going on. I think I see a lot of people who are traditional uh value guys and careful investors uh just writing it off in. entirely saying it's it's a an obvious bubble

16:53Unidentified speaker (UNKNOWN): and

16:54Tobias Carlisle: there's nothing of value there. And then on the other hand I see these guys who are just like buy this stuff at any price so you get left behind. If you you know anthropic will be the only business around in whatever three years, ten years, twenty five years. There'll be anthropic motors and anthropic pharmaceuticals. vehicles And if you don't own it, then you're left behind. And I don't think obviously those two extremes, neither of those are correct. I think probably what happens is it just becomes another tool that that many people use. I do think that that's that's a that's a that's a likely scenario. And the example I'd give for that, I've got a friend who's a um specialist doctor who reads the uh

17:32Unidentified speaker (UNKNOWN): no

17:32Tobias Carlisle: I can't think radio radi ready

17:34Unidentified speaker (UNKNOWN): not a

17:35Tobias Carlisle: radiographer radiologist works in Australia doing the off hours in London. So they do the midnight to 8 a.m. shift in London from Australia because you can send the the files digitally

17:47Unidentified speaker (UNKNOWN): to

17:47Tobias Carlisle: Australia.

17:48Unidentified speaker (speaker_0): Okay.

17:49Tobias Carlisle: And so um the AI has for years AI has been doing the analysis, the initial analysis there. And I I saw the this statistic of the day, there's there are more radiologists now than there were 10 years ago. So they're clearly like and that's one area where AI has been penetrating for a very long period of time. I don't know if you've they've been good at finding uh breast cancer and other sort of cancers that are very difficult to detect with a human eye, but the AI can can find them as a trivial sort of matter. And so I do think that that's the most likely outcome that

18:25Unidentified speaker (UNKNOWN): we

18:25Tobias Carlisle: all just use it as a tool. and ultimately there's still a human being there on top making those decisions. But I would be very worried if you had a purely internet email job that the AI will probably be able to do it in the very close future.

18:40Unidentified speaker (speaker_0): Oh yeah. Call whole centers, I mean there's so many, there's so many different elements. But yes, yeah, you have to have a a human to oversee it. And and there is a lot of slop. Like I mean, I mean you do see it's getting less and less. Like they're getting so much better, but there is a lot of slop. And it's like you said, like it like if you're writing a white paper, the reality is that the research that you can do faster. Like I love, I love like I'll I'll do comparisons to this uh market versus like the tech bust. And like it's so easy to quickly pull up, okay, what were the value? valuations? What happened to um the revenue and the margins for like the intels and stuff and how quick, how quickly did it change for them? You know, and and what you can accomplish is is you're putting out better products, better, better material. And it's just something where maybe you'd have a research assistant in the past and the guy

19:35Unidentified speaker (UNKNOWN): would

19:35Unidentified speaker (speaker_0): be slaving away at the library or something

19:37Unidentified speaker (UNKNOWN): like

19:37Unidentified speaker (speaker_0): that. Or

19:38Unidentified speaker (UNKNOWN): yeah it's

19:39Unidentified speaker (speaker_0): and and you I mean it's it's it's incredible. So I don't think that I I think you have to be very bullish. The crazy thing with like that I think about with like anthropic and these other ones is like all right like you know this guy's putting in all these PhD papers or whatever, what if physics papers uh and and but they have access to all of this information. So while while you know you and I both educate our own LLMs and you know they're pretty sophisticated, imagine on Claude's side how much info that they have and like what they could do with that. I mean, imagine uh a nefarious CEO and and like what power they would have with that. And it seems like it's gonna be so hard to regulate because I mean the reality is is China's gonna eat everybody's lunch. if if you do.

20:28Tobias Carlisle: I don't think we can take the foot off the accelerator.

20:31Unidentified speaker (speaker_0): I wouldn't.

20:31Unidentified speaker (UNKNOWN): I

20:31Unidentified speaker (speaker_0): wouldn't. Yeah. You

20:33Unidentified speaker (UNKNOWN): want data

20:33Tobias Carlisle: centers powered by nuclear with the fastest chips in the world running the best LLMs. And that's gonna be the you you just need that at a national level.

20:43Unidentified speaker (UNKNOWN): And

20:43Unidentified speaker (speaker_0): doors out of the box, Toby. Yeah.

20:45Unidentified speaker (UNKNOWN): You're

20:45Unidentified speaker (speaker_0): not putting it back in.

20:47Unidentified speaker (UNKNOWN): That's

20:47Tobias Carlisle: right. One of one of the really interesting charts that I saw. So I put this in the last. investment.

20:51Unidentified speaker (UNKNOWN): I

20:51Tobias Carlisle: put a couple of these charts in the last investment. This little email goes out every day.

20:56Unidentified speaker (UNKNOWN): And

20:56Tobias Carlisle: I just basically flick it um interesting stuff that I see on Twitter and it puts it together into an email. And one of the ones that I've seen on this is one that I've AI stuff moves so fast, it's crazy. Like last week there was this group in San Francisco, Meta metermetr, M E-T-R, it's uh uh.org. dot org.

21:14Unidentified speaker (UNKNOWN): Okay.

21:15Unidentified speaker (speaker_0): And

21:15Tobias Carlisle: they they're like uh watching the AI guys, you know, whatever that's going to do. But they're they had access to anthropics internal models.

21:23Unidentified speaker (UNKNOWN): And

21:23Tobias Carlisle: the question that they wanted to ask uh answer, which is the one that I sort of, you know, the the reason I called it the last investment is there's this science fiction guy called the the supercomputer, the superintelligence, the last invention, with the idea that once you make the superintelligence, the superintelligence. makes everything else. And so I sort of thought, well,

21:42Unidentified speaker (UNKNOWN): once

21:43Tobias Carlisle: you've built that thing, like what other capital do you need? Like this thing just sort of figures it out itself. So that's why it's the last investment. But this meter analysis, they said, what we've always watched, like Moore's Law, which is kind of amazing, is this It's uh scaling where it's the number of transistors you can fill film in on a chip, but the point is basically computing power has been growing, has been 10xing on a regular schedule for like four years or something, for 50 years now. And they've they every time it looks like it's not going to happen again, they figure out some new way to keep on working on that

22:16Unidentified speaker (UNKNOWN): at

22:16Tobias Carlisle: that level of growth. And what that has allowed is all of these other technologies that now sit on top. And so AI is the technology that sits on top and takes advantage of those things.

22:26Unidentified speaker (UNKNOWN): And the

22:26Tobias Carlisle: question that the science fiction writers always ask was once you get an AI that can improve itself, that's recursively learning on computer time, indefatigable, just works all week long, puts in so many more hours, and you can have 10,000 of these things going at once instead of one.

22:42Unidentified speaker (UNKNOWN): So

22:42Tobias Carlisle: they will make much faster breakthroughs in their own AI. They will make super intelligent AIs faster than humans can do them. So it will at some point in the not too distant future, we won't understand how the AI is building new AIs. It'll just be too far advanced for us. So MEDA is interested in whether the internal AIs are doing this, and they've looked at these curves.

23:07Unidentified speaker (UNKNOWN): And

23:07Tobias Carlisle: so the usual, you know, the log curve, you take the you take the y axis, and instead of it being one, two, three, four, five, it goes one, ten, one hundred, and then you fit a flat line to that.

23:18Unidentified speaker (UNKNOWN): And

23:19Tobias Carlisle: so the new AIs are completing tasks and they're they're turning up on that curve, which means they're going faster than any rates that we've seen historically, faster than we had predicted. And every time we look at it again, it looks like it's accelerated again. It's gone up it's two and a half times faster over the last twelve months than it was in

23:38Unidentified speaker (UNKNOWN): the

23:38Tobias Carlisle: twelve months before then. So

23:40Unidentified speaker (UNKNOWN): we're

23:41Tobias Carlisle: in this very, very like this is why we've been talking about the singularity a little bit. Like I really do think we are at some sort of inflection point. We're at that singularity point where you think about where LLMs were a year ago. They're like okay, but they needed to be watched. And now we're like they're pretty good at just about everything that I want them to do. A year from now, like, where are they going to be?

24:04Unidentified speaker (speaker_0): Right. It's uh exactly. I mean, and that's why it's hard, it's hard not to be bullish on it. I mean, think of a company like Amazon or Google. I mean, they're just you have these huge organizations with the smartest people and all this fine, all these financial resources, and they have the the power of AI too now, and and they're building their own models. It's it's it's it's totally transparent. transformational i mean realistically it might be more transformational than the internet you know yeah so so no it's it's crazy to think of you know how that's gonna impact things uh you know it's you know so many so many different levels and from a stock pickers um standpoint i think you've just got to realize that there's going to be a lot of change. Um, I saw that a lot of capital was going, uh, I not not a ton of capital. I think someone may raise like 50 million or 100 million, but they're setting up like a CRM. So it's people that used to work for one of the CRM companies, and they're starting up a new

25:06CRM, uh, you know, utilizing AI capabilities. So, and and I know that there's that guy who recreated a lot of Adobe's uh products and stuff like that. So,

25:16Unidentified speaker (UNKNOWN): you

25:17Unidentified speaker (speaker_0): know, you you need a margin of safety, that's for sure. And

25:20Unidentified speaker (UNKNOWN): one

25:20Unidentified speaker (speaker_0): thing I was thinking about, Toby, is like, all right, you have Nvidia down a ton. The market's not down that much, but I mean, so many stocks are down 30, 40 percent.

25:30Unidentified speaker (UNKNOWN): I

25:30Unidentified speaker (speaker_0): feel like this market is just being held up by Apple. You know what I mean? It's like you got Apple here trading at like 35 times earnings, it's growing grown revenue. by like six percent over the last few years. And it's just,

25:43Unidentified speaker (UNKNOWN): you

25:43Unidentified speaker (speaker_0): know, it's just such an outlier. And I wrote an article for Seeking Alpha, just kind of basically saying, you know, neutral to negative on Apple. Yeah. And I can get a good feel for things. Like whenever everyone hates it, you know, I always feel good about the prediction. The same thing happened. I wrote about Micron when it. hit a high over a thousand, like eleven hundred dollars or something like that. So

26:06Unidentified speaker (UNKNOWN): I was just getting ridiculed man.

26:09Unidentified speaker (speaker_0): Even clients. Like clients were calling me and just like like, no, no, you don't get it. They're sold out. Like you don't get this. Like this is you know, and it's like, no, I mean this is kind of the history of it. Um so I think if Apple goes, you know, the market goes.

26:25Unidentified speaker (UNKNOWN): Yeah,

26:25Tobias Carlisle: that's kind of interesting. I I I think there's there's there's two things there. That's one of the things that I considered in this little white paper.

26:32Unidentified speaker (UNKNOWN): One of

26:32Tobias Carlisle: them was I looked at 200 years of tech revolutions and there's all of these um Clayton Christensen, Carlotta Perez, who I mentioned another guy, William Nordhorse, Nordhaus, something like that.

26:43Unidentified speaker (UNKNOWN): And

26:43Tobias Carlisle: they looked

26:44Unidentified speaker (UNKNOWN): at who

26:45Tobias Carlisle: keeps the money. So Carlotta Perez looks

26:48Unidentified speaker (UNKNOWN): at

26:49Tobias Carlisle: the different stages of the uh of the build out in terms of time. Clayton Christensen looks at where it migrates through the bottlenecks and so on. And it certainly uh it seems like it goes to where there is a bottleneck. So it goes to whatever the the money flows to and initially it's this infrastructure layer. And so the infrastructure layer is going to be data centers, chips, power, all of the physical stuff

27:13Unidentified speaker (UNKNOWN): that

27:13Tobias Carlisle: helps us sort of run the computers. And you want to be careful because that's the very end of the ball whip that sort of moves the most, but

27:23Unidentified speaker (UNKNOWN): the

27:23Tobias Carlisle: money flows towards it. We do eventually solve whatever that problem. It could take a decade. That's not to say it happens really quickly. It could take a long period of time. And then you have to look at the builder layer on top of

27:34Unidentified speaker (UNKNOWN): that. And

27:34Tobias Carlisle: so these are these are things, and then there's a a consumer layer on top of that. And it ends

27:38Unidentified speaker (UNKNOWN): up

27:38Tobias Carlisle: the consumer. consumers tend to collect most of the value. And the example, the really great example in the book is in the in the white paper is light. So it took

27:46Unidentified speaker (UNKNOWN): I

27:47Tobias Carlisle: looked at like in the 1800s how long you had to work to get some light and it was like a candle.

27:52Unidentified speaker (UNKNOWN): And

27:52Tobias Carlisle: you had to work five hours to afford a candle, the average wage.

27:56Unidentified speaker (speaker_0): And

27:56Unidentified speaker (UNKNOWN): now the average

27:57Tobias Carlisle: wage the average wage or the the median wage whatever the the the lower estimate is

28:04Unidentified speaker (UNKNOWN): You

28:04Tobias Carlisle: work like half a minute

28:06Unidentified speaker (UNKNOWN): to

28:06Tobias Carlisle: turn all the lights on in your house. Like it's just the or or five seconds for the intel so the cost of light has come down. And I think that intelligence will do the same thing to the

28:15Unidentified speaker (UNKNOWN): or

28:15Tobias Carlisle: or we're going to do the same thing to intelligence. So there's no reason why it won't cost more than the cost of uh electricity, the chips and all the infrastructure it's going gonna to be built on, it'll drive towards that cost.

28:25Unidentified speaker (UNKNOWN): And

28:25Tobias Carlisle: so in valuing this stuff,

28:27Unidentified speaker (UNKNOWN): I

28:27Tobias Carlisle: couldn't agree more. You the question of whether something is going to be useful and impactful is entirely different from whether you can make money investing in it.

28:36Unidentified speaker (speaker_0): Yes. And

28:37Tobias Carlisle: you need the same things you've always needed, which is a competitive advantage, uh a discount to the, you know, a wide enough margin of safety so you don't get blown

28:47Unidentified speaker (UNKNOWN): up.

28:48Tobias Carlisle: And then sort of some skepticism too about the growth rates. because I I think that when you see those exponential growth rates, it's it's it's easy to get carried away with where you're going to be in the future. It's hard to build a model out

29:01Unidentified speaker (UNKNOWN): that

29:01Tobias Carlisle: truly encapsulates the you know the most likely or the the median outcome there because that you get that top line bending up like that all the time. It sort of starts looking it pulls everything up. So you need some skepticism. too in those valuations. I like you, I look at NVIDIA and I look at the the amount that it earns on its assets is kind of extraordinary.

29:22Unidentified speaker (speaker_0): Incredible, incredible. The cash flow.

29:25Unidentified speaker (UNKNOWN): Yeah,

29:25Unidentified speaker (speaker_0): I mean, I think that the one potentially negative argument is like, all right, with Jensen And gone,

29:34Unidentified speaker (UNKNOWN): you

29:34Unidentified speaker (speaker_0): know, are the neo clouds still gonna going to get the funding that they need? I mean, he basically created the neo cloud. clouds to some extent to create competition for the hyperscalers so that people would still have demand for the chips. And

29:46Unidentified speaker (UNKNOWN): let

29:46Unidentified speaker (speaker_0): me ask you something. I'm just curious about your opinion. It could be controversial either way, but what would you do you think they should have kicked out Jensen?

29:56Unidentified speaker (UNKNOWN): Do

29:56Unidentified speaker (speaker_0): you think they had to?

29:58Tobias Carlisle: I think a board.

29:59Unidentified speaker (UNKNOWN): you

30:00Tobias Carlisle: Yeah, know, as a

30:01Unidentified speaker (UNKNOWN): as a

30:01Tobias Carlisle: former lawyer, I used who to. advise boards on these sort of issues, there's just it's black and white. There's just no question for something like that. Because it just exposes the company too much risk.

30:12Unidentified speaker (speaker_0): Right.

30:12Unidentified speaker (UNKNOWN): And

30:12Tobias Carlisle: so when people say, you know, the board is doing so like the board is uh boards are boards are risk first. Like that's they're but they just can't This be anything other than risk and compliance and that's what they do. So they there's no question they had to make that dis that was sort of that decision makes itself. You can't have you can't do anything else there.

30:31Unidentified speaker (speaker_0): Also

30:32Unidentified speaker (UNKNOWN): the

30:32Unidentified speaker (speaker_0): government. You know, what is the government's real take on it? I know Scott Bessant really wants change there, and then Trump allegedly said that, you know, he'd be okay with them say I don't know. You know, but yeah, there is a lot there's a lot of pressure. The thing I always think about, and I see this in football, like a a a player will, you know, have some stupid crime. you know, and and it's like there's a lot of pressure to cut him, and then the next team just signs him and he's awesome, you know. And right

30:57Unidentified speaker (UNKNOWN): the

30:57Unidentified speaker (speaker_0): reality is, you know, if if if Jensen joined, let's say, SpaceX, or

31:03Tobias Carlisle: yeah,

31:03Unidentified speaker (speaker_0): you know, uh, I mean

31:05Unidentified speaker (UNKNOWN): I think that's

31:05Tobias Carlisle: entirely possible.

31:06Unidentified speaker (speaker_0): Or yeah, or any like any of those companies, that stock's going up 15%, you know, in in a day, probably. I mean, he's he's he's that much. of kind of a phenom uh with his relationships. And you also have the variable where, like,

31:20Unidentified speaker (UNKNOWN): I

31:21Unidentified speaker (speaker_0): mean, Taiwan semiconductors are a fascinating company because they make all of this stuff pretty much. There's really not competition for it that's anywhere near as sophisticated. And from my understanding, I'm not an expert in this field, but you know, Jensen has the real you know, strong relationships there, that obviously. other people at Nvidia do too. But like that's something you kind of have to massage. Um,

31:44Unidentified speaker (UNKNOWN): I

31:44Unidentified speaker (speaker_0): don't know. I mean, I look, I'm obviously there's a moral standpoint, and then there's a business standpoint. Like, for me,

31:50Unidentified speaker (UNKNOWN): there's a national

31:51Tobias Carlisle: there's a national security issue as well,

31:53Unidentified speaker (speaker_0): security exactly. And with all those said, that's why personally I think I think they should have kept him. Um, because because you need him to steer the damn ship. I mean. i mean they they should use this window as an opportunity to build a bench you know and and to have succession in line and whatever

32:13Unidentified speaker (UNKNOWN): rings

32:13Tobias Carlisle: it's amazing that they didn't

32:15Unidentified speaker (speaker_0): i can't believe i can't i can't believe and i haven't heard the conference call yet but i a lot of people were very uninspired inspired on the cfo and but i mean i get it you have a it's really hard to replace a charismatic leader

32:26Tobias Carlisle: like

32:27Unidentified speaker (speaker_0): Jensen. I mean imagine replacing Elon Musk. I

32:29Unidentified speaker (UNKNOWN): mean you don't want

32:30Tobias Carlisle: a charismatic CFO.

32:31Unidentified speaker (UNKNOWN): That's

32:32Tobias Carlisle: just

32:32Unidentified speaker (UNKNOWN): yeah you don't that's

32:33Tobias Carlisle: one thing that I've heard.

32:34Unidentified speaker (speaker_0): That's actually a great point. Yeah you don't you don't want that. Uh I think maybe like the world comms and the Enron's probably had that but yeah that's a good

32:43Unidentified speaker (UNKNOWN): point.

32:43Tobias Carlisle: That's right.

32:44Unidentified speaker (speaker_0): Yeah.

32:44Unidentified speaker (UNKNOWN): Yeah

32:45Tobias Carlisle: I I I kinda can't I I I do agree with you like it it does seem like he was very important to NVIDIA, but

32:52Unidentified speaker (UNKNOWN): there's

32:52Tobias Carlisle: that great Charles de Gaulle quote which has been making the rounds for the last few days that the work the graveyards are filled with indispensable men. And I do wonder, like, really, how indispensable is he to to a business that is basically uh making I actually I'm d I'm I'm not entirely clear on whether they make the chips or not, because I saw this

33:11Unidentified speaker (UNKNOWN): I

33:11Tobias Carlisle: saw this actually I don't think they do, because it Musk put this tweet out where he said, really what Apple and NVIDIA and he listed a few other companies that I just can't I'm I'm I can't remember who they are off the top of my head. But

33:24Unidentified speaker (UNKNOWN): he

33:25Tobias Carlisle: said all they do is they sell they send these digital files

33:29Unidentified speaker (UNKNOWN): to the

33:30Tobias Carlisle: people who actually make this stuff. And it's TSMC, as you point out.

33:33Unidentified speaker (speaker_0): TSM, yeah, Taiwan Semiconductor, exactly. They make up.

33:36Unidentified speaker (UNKNOWN): ASML.

33:36Tobias Carlisle: ASML

33:37Unidentified speaker (UNKNOWN): was is the other one.

33:38Unidentified speaker (speaker_0): ASML. creates the lithography machines that Taiwan Semiconductor buys. Um, so ASML is just another amazing business. But yeah, I mean, I I researched, I did this too, like this last week. I really did a deep dive on Taiwan semiconductor and and I looked at like their competitors, and the competitors are really far behind. Like Samsung's probably the best competitor. Uh but but yeah, I mean, I mean, so that's the thing. So they could they could outsource more of like Nvidia's market share to like uh Broadcom and then Broadcom's designing the chips for Google and for and for Amazon. Uh and so you know whatever that dynamic is, you have to keep that relationship strong and it's just, I think we alluded it to it like the other a week ago or something. But I is I remember like investing Bruce Greenwald's book. It's a really freaking good book. And I just

34:38remember one of the case studies, though, was how Intel is so dominant. Right. It's almost impossible for AMD or Taiwan Semiconductor to compete with them. You know, the foundry business is crap competitive. compared to what Intel does. And then Intel started trying to create foundries, and they still are, and it's just been a huge money loser.

34:58Tobias Carlisle: One

34:58Unidentified speaker (UNKNOWN): of the

34:58Tobias Carlisle: more shocking things. Like I I couldn't agree more.

35:00Unidentified speaker (UNKNOWN): Because

35:01Tobias Carlisle: as I've as I like, we all grew up on stories of don't be IBM because you don't want to be packaging with Wintel machines. Windows and Intel got all the value there. And then the the Groves book uh only the Paranoid. Survive, which was about Intel and how they outcompeted everybody else. And just every every book that uses competitive advantage, like the illustration of competitive advantage, is Intel. And somehow they squandered it. And I

35:29Unidentified speaker (UNKNOWN): made the

35:29Tobias Carlisle: mistake early on, like what they well not early on, but when they collapsed

35:34Unidentified speaker (UNKNOWN): a

35:34Tobias Carlisle: few years ago now, five

35:38Unidentified speaker (UNKNOWN): plus

35:38Tobias Carlisle: years ago. when they sort of had that initial collapse, and they w they were off quite a lot. I bought some because they were off quite a lot, because in my mind I had this is one of the great competitive advent competitively advantaged businesses in the world, available at a pretty good price, and I bought it and I was down shockingly on that position. And then I abandoned it. And the only I think that it sort of came back to life when uh Trump put some government money into it. But I don't even know if it's I don't know if it's up on that. But that's

36:05Unidentified speaker (UNKNOWN): that's it's

36:05Tobias Carlisle: a

36:06Unidentified speaker (UNKNOWN): it's just a good thing

36:07Tobias Carlisle: to bear in mind that the top 10 businesses from a decade ago, except for Microsoft, and that's we we can talk about how well Microsoft survives the next decade. Only Microsoft has survived from the decade before then as well. So Microsoft was around in 2000, 2010, 2020.

36:25Unidentified speaker (UNKNOWN): But

36:26Tobias Carlisle: now like

36:27Unidentified speaker (UNKNOWN): I

36:27Tobias Carlisle: guess that they've got I know they've got other businesses there. They've got they've got a cloud business and they've got uh copilot and all that sort of those tools as well that I that are very widely spread throughout business. But how often would you open up Word or Excel these days?

36:45Unidentified speaker (speaker_0): No, that's such a good

36:46Unidentified speaker (UNKNOWN): point. You

36:47Tobias Carlisle: go through claw cool.

36:48Unidentified speaker (UNKNOWN): yeah.

36:49Unidentified speaker (speaker_0): Yeah ex X Excel has always been like my biggest weakness. I would have to have, you know, my brother do them for me because I just have never learned Excel. I

36:57Unidentified speaker (UNKNOWN): remember

36:58Unidentified speaker (speaker_0): I remember doing a job interview and I it was like five interviews and I was killing it. I thought I was like this superstar gonna get this job. And then the very last part was like an Excel test. And man, I could not do I didn't I could not do anything. And I never developed that skill. And now you don't have to. It's just absolutely extraordinary, you know, what you can do. So no, very much agree. I Microsoft's in great condition. I think all I really like all the hyperscalers. I just think

37:25Unidentified speaker (UNKNOWN): that

37:25Unidentified speaker (speaker_0): that's like the window you want to be. And I don't think the valuations are are are that bad either. And and so yeah, curious to see how it plays out.

37:35Unidentified speaker (UNKNOWN): Yeah,

37:35Tobias Carlisle: the the Excel one was, like you know, all of the Excel jockeys, all of the guys who came out of the investment banks, like they've pride of themselves on their ability to use Excel without a mouse.

37:43Unidentified speaker (UNKNOWN): Like

37:43Tobias Carlisle: that was the big flex. If you

37:44Unidentified speaker (UNKNOWN): could like

37:45Tobias Carlisle: hotkey through the whole document, you were

37:47Unidentified speaker (speaker_0): yeah,

37:48Tobias Carlisle: you you were the man. Yeah.

37:49Unidentified speaker (UNKNOWN): Uh

37:50Tobias Carlisle: no more. Now you're gonna be able to now you talk to Claude, Claude goes and does it, and it might not do it in Excel, it does it in you know Google Sheets or something else like

37:57Unidentified speaker (UNKNOWN): that. There's

37:58Unidentified speaker (speaker_0): a lot of yeah, like a lot of the technologies that I had never heard of, like honestly. I I mean this is pretty embarrassing. I didn't even know what like GitHub was, you know, for example. Or like or like there's render, which is like an email service, I think. Uh and like it also does something for websites. Like you use these softwares and I pay for these subscriptions now. Uh

38:18Unidentified speaker (UNKNOWN): and

38:19Unidentified speaker (speaker_0): so I mean, you know, I I don't know. No, I I think there's a lot of opportunity. And I don't think like even if you can recreate like the Adobe products,

38:28Unidentified speaker (UNKNOWN): I

38:28Unidentified speaker (speaker_0): don't know. I mean, it's not that expensive. I mean, maybe if you're using like the most advanced one, but like but I can understand companies paying, you know, to have it done and to have it serviced, you know. I mean, that's the problem, is if you create like your own thing, you know, you have to service it and this stuff gets glitchy, you know. So

38:46Tobias Carlisle: okay. When I sell my business, I want the best tax and investment advice. I want to help my kids, and I want to give back

38:52Unidentified speaker (UNKNOWN): to the

38:53Tobias Carlisle: community. Ooh, then it's the vacation

38:56Unidentified speaker (speaker_2): of a lifetime.

38:58Unidentified speaker (UNKNOWN): I

38:58Tobias Carlisle: wonder if

38:59Unidentified speaker (speaker_0): my head of office has a forever setter.

39:01Unidentified speaker (speaker_1): An IG private wealth advisor creates the clarity you need with plans that harmonize your business, your family, and your dreams. Get financial advice that puts you at the center. Find your advisor at IG IGPrivatewealth.com. PrivateWealth.com.

39:17Unidentified speaker (speaker_2): Legendary has arrived in Alberta with BetMGM Casino. Sign up to unlock access to over 4,000 casino games, including exclusive slots you can't spin anywhere else. Survivor's ready? The iconic TV show is now an online casino game where you can tap into Island Fun. BetMGM has also brought the prices right: Game of Thrones, Family Feud, MGM millions, and so many. more exclusive casino games to Alberta. Get Las Vegas right from your phone with live dealer games streaming from real tables. And every time you play, you can enjoy instant deposit options and fast withdrawals. BetMGM Casino and its legendary library of games are available. Download the BetMGM app. See BetMGM.com MGM.com for terms. 18 plus Alberta only. Please play responsibly, subject to eligibility requirements. If

40:03Unidentified speaker (speaker_0): gambling is affecting your mental health or well-being, 211 Alberta is here to help. Call or text 211 or visit ab.211.ca. BetMGM operates pursuant to an operating agreement with the Alberta iGaming Corporation. Standard message and data rates may apply.

40:17Unidentified speaker (UNKNOWN): I

40:17Tobias Carlisle: guess the question is, you know, do you not no longer need the tool if you can just explain it to the like the general AI? The image output from general AI is like this is an image output from a general AI, which is kind of like

40:27Unidentified speaker (UNKNOWN): it

40:27Tobias Carlisle: took I will say that the two things are going to survive. You have to have reasonable taste. I'm not saying that this I'm have great taste or anything like that. I'm just saying that you have to have some eye for what you're trying to produce. Otherwise. Because you get it it does take a a few versions. And I've when I write I ask it to do some writing and it likes to go off in its own direction sometime. It reaches

40:48Unidentified speaker (UNKNOWN): a different

40:48Tobias Carlisle: conclusion to the one that I want it to reach. And so I go back a few times to get it to to go in the right direction. And I think part of that is you have to feed it the right ideas to to get it going. And then you have to be a little bit careful with that. But there's some it's not perfect yet, but I guess it eventually it gets to that point. Um but it's been interesting to get an introduction. I too have had a recent introduction to GitHub and all of that other stuff that

41:12Unidentified speaker (UNKNOWN): powers

41:12Tobias Carlisle: the back end. The world

41:14Unidentified speaker (UNKNOWN): is going

41:14Tobias Carlisle: to be remade.

41:15Unidentified speaker (speaker_0): Yes. Is

41:16Tobias Carlisle: is my point. And I

41:17Unidentified speaker (UNKNOWN): yeah

41:18Tobias Carlisle: I it's too early, I think. It's going to be very, very hard to pick the winners. But I think that there are going to be some very big losers. And I think I know we we we sort of I I equally think that Adobe looks reasonably cheap. Adobe's sort of hovering around in my in my screens as well.

41:34Unidentified speaker (UNKNOWN): I

41:35Unidentified speaker (speaker_0): bet it is.

41:35Tobias Carlisle: And I could I could easily see initiating a position in Adobe.

41:38Unidentified speaker (UNKNOWN): Yeah.

41:39Tobias Carlisle: But it's one of the ones that I'm pretty nervous about because I do see what the uh what the general what the LLMs are capable of now.

41:47Unidentified speaker (UNKNOWN): Yeah.

41:47Tobias Carlisle: I just don't know if it's maybe it's because it's only we're using I'm using smaller images. I'm not using particular not doing anything particularly sophisticated. Maybe you

41:55Unidentified speaker (UNKNOWN): need

41:55Tobias Carlisle: something more impressive if you get it, maybe you need the full Adobe suite. What what kind of a compet competitive advantage do you

42:02Unidentified speaker (UNKNOWN): do you

42:02Tobias Carlisle: see them having where you don't think that you need to hold it for long enough to worry about it?

42:06Unidentified speaker (UNKNOWN): I mean I

42:06Unidentified speaker (speaker_0): think for people that are doing like if you're creating like ads, you know, um their their content library, their ability to uh you know source things, uh

42:17Unidentified speaker (UNKNOWN): you

42:17Unidentified speaker (speaker_0): know, I just think that you also have companies that everyone knows how to use Adobe. It's like it's like everyone's trained on it, you know. Whereas if you have the them all kind of just jerking around on their own LLMs, you know, duplicating something or it could be a little different. So and it just what is the cost? You know, it's like you're willing to,

42:37Unidentified speaker (UNKNOWN): you

42:37Unidentified speaker (speaker_0): know, you could you could dry clean your clothes, you know, but but yeah, yeah, but but but uh but but you know we still use dry cleaners. beaners. You know it's it's it's a convenience. So I agree. I I don't think that the competitive moat is huge there. Uh but at some point, you know, the valuation gets compelling and to where you know it's it's attractive. Um so it's

43:00Unidentified speaker (UNKNOWN): all I mean it's yeah.

43:01Tobias Carlisle: But it's also true that and I think we should talk about this a little bit like the the rules of investing haven't gone away completely either.

43:08Unidentified speaker (speaker_0): Not at all.

43:09Tobias Carlisle: You you want you want high returns on invested capital. Uh you know you want and you and you don't want but you don't want to pay for the them. You want to you want to pay for

43:17Unidentified speaker (UNKNOWN): you

43:17Tobias Carlisle: wanna get a big discount, a big margin of safety on on those returns, knowing that there's probably going to be some compression. And then also knowing that you're going to have some false positives, you're gonna buy some value traps that the things that they look cheap optically, but for whatever reason the business has just competed away. And I don't know, there's a lot of candidates out there. H and R block is one you talked about tax

43:38Unidentified speaker (UNKNOWN): before. Oh

43:38Unidentified speaker (speaker_0): yeah.

43:39Tobias Carlisle: Is there a better use case for AI than tax? Like I can't imagine. It's hard for

43:42Unidentified speaker (UNKNOWN): a human to remember.

43:43Unidentified speaker (speaker_0): I'm not I'm doing my own taxes from now on. It's

43:46Unidentified speaker (UNKNOWN): just

43:46Unidentified speaker (speaker_0): there's no sense in in not doing it. I had some audits and stuff. And so that's why I really was like, all right, I want to put distance on this is like a nightmare. Um but yeah I mean you can do it so much better now and just linking all that up. So yeah that's I mean it's I mean look whatever financial advice. You know, there's there's so many industries that are gonna be impacted, but I think it's critical to have like that human element because much of financial advice is psychology.

44:16Unidentified speaker (UNKNOWN): Well,

44:16Tobias Carlisle: behavior,

44:16Unidentified speaker (UNKNOWN): yeah, I couldn't

44:17Tobias Carlisle: agree more.

44:18Unidentified speaker (speaker_0): It's that's a huge part of it is talking people off the ledge, like helping people with their decision making, complicated cases, you know, where you have like an inherited IRA and stuff like that. And just, you know, telling people like, look, you don't want all this risk. And so, yeah. So, no, we'll, we'll, we'll be here for it and we'll keep talking about it as the weeks progress.

44:40Tobias Carlisle: Well, the way that the the way that the world changes, week by week, it really is uh different. And I I was gonna say just that I I'm halfway through my metat meta talk, my meter talk before. Somebody updated that afterwards and they said it it looks like you know a big part of the gain to be made by the amount of money that's been thrown at it. So the question is are the gains that we're seeing as a result of money or are that they the result of the the LLMs training themselves or creating better LLMs? That that question's still up in the air. So we're gonna keep on watching that. Uh we'll be back next week.

45:12Unidentified speaker (speaker_0): Yep. Sounds good, Toby. Appreciate it, man. Looking forward

45:14Unidentified speaker (UNKNOWN): to it.

45:15Tobias Carlisle: Uh thanks, folks.

45:17Unidentified speaker (speaker_1): My teammate says he plays pickup basketball,

45:19Unidentified speaker (speaker_0): quote, for the cardio. Well, there's more to it than that, Sean. Like with Bet365, I get paid the second my team goes up big with early payout and can cash out at any time before the game ends. Sean's getting his steps, but I'm getting paid. Bet365, the home of winning early. 18 plus. please play responsibly. Alberta only terms, conditions apply. If

45:40Unidentified speaker (speaker_1): gambling is affecting your mental health or well being, 211 Alberta is here to help. Call or text 211 or visit ab.211.ca. Standard message and data rates may apply.

45:47Unidentified speaker (speaker_2): So your decaf oat milk, extra but not piping hot, four pumps of caramel, no sugar, small latte, and a large cup. That will be 10 minutes.

45:54Unidentified speaker (speaker_1): No worries. I'm having fun on Play Ocho.

45:57Unidentified speaker (UNKNOWN): And

45:57Unidentified speaker (speaker_1): I keep it fun with all the latest slot and live casino games. Plus, I get instant payouts with no minimal. minimum withdrawals. Oh, and no wagering requirements ever.

46:06Unidentified speaker (speaker_2): Here's your decaf oat milk extra, but not piping hot.

46:08Unidentified speaker (speaker_1): Ah, thanks. Playtime's over.

46:10Unidentified speaker (UNKNOWN): Mmm,

46:11Unidentified speaker (speaker_1): lovely. Feel

46:12Unidentified speaker (UNKNOWN): the fun. Play O Joe.

46:14Unidentified speaker (speaker_0): 18 plus Alberta only. Please play responsibly. Full term supply.

46:17Tobias Carlisle: We'll see you then.

This page was produced automatically: the audio was transcribed with nvidia/parakeet-tdt-0.6b-v3+sortformer-v2.1 and interpreted by Qwen3.6-35B-A3B-UD-IQ3_S.gguf. Transcription and interpretation can contain errors; always confirm against the original episode. Nothing here is investment advice.
All insights