Financial Crisis, AI Bubbles and Big Tech - Three Podcasts

Three Podcasts - The 2008 Financial Crisis, The AI Bubble and Big Technology

Recently I have been trying to soak up as much information and learning as possible on the financial, technological, philosophical and sociological impact of Artificial Intelligence. As technologists its easy to get sucked into the tech side of the use of AI or the tech side of building AI, both of which are fascinating in their own right, but I think its very important to understand this current revolution across as many vectors as possible.

Being a ravenous consumer of podcasts I am of course hovering up as much podcast content on these topic as I can squeeze into each week. A week or so back I listened to some excellent episodes of different podcasts that really got me thinking. One podcast is not directly AI related at all, the other two are but all are linked.

Anyone as interested in these things as I am will, I am sure, find these equally as interesting so I’d like to share them,

Podcast One - Against the Rules: The Big Short Companion Podcast with Michael Lewis

It’s 15 years since Michael Lewis published his excellent book on the 2008 financial crisis, The Big Short. To mark the occasion he has released a new audio book version narrated by Michael Lewis himself. It often strikes me as unfair that so much of more than one talent finds a vessel in single person. But he is also a fantastic speaker and ‘reader of audiobooks’. Such is the case with Michael Lewis in that not only is his writing and narrative style very engaging, but he is also a fantastic speaker and ‘reader of audiobooks’. The latter of which is a skill other authors who have tied to do the same have failed miserably at. Lewis always finds entertaining ways to turn esoteric and niche stories into compelling tales.

As he did with the release of the new audio book version of his first book, Liars Poker (which he also narrates), Lewis has released a series of podcast episodes talking to characters from the book. The people involved with the movie of the book and people involved in the financial services industry at the time of the crisis. The series, Against the Rules: The Big Short Companion Podcast with Michael Lewis, is indeed the perfect companion to the audiobook. It’s of course compelling and interesting and very insightful to hear the perspectives of the actual people involved.

There are two particular episodes I want to call out. Not because they are any better than the others, but they triggered some thoughts. The first episode is Episode 6. Michael is speaking with Bloomberg writer and former Goldman Sachs investment banker, Matt Levine about the impact the 2008 financial crisis had on the industry. He has some super interesting observations and things to say, but one thing struck me. At one point Levine says something along the lines of: “When I was at Goldman, I witnessed the financial crisis from inside, but I didn’t understand it because I was working my job”. I can identify with this.

In 2008 I was working at the French bank BNP Paribas in fixed income. Not as an investment banker, but in IT. I worked on the capital markets IT team looking after the production trading platforms for the FX and Interest Rates trading desks. During that time we all knew there was a big, historic event happening in the financial systems. But I for one didn’t understand the full length and depth of it within my domain. I distinctly remember running on the treadmill (I was also getting married in 2008 so was working hard to fit into my suit!) in the gym at the London BNP Paribas building in Marylebone watching the news on the TV and seeing reporting on Lehman Brothers. It was the first time I thought: “Ok this is big”. I texted a friend that worked at Lehman asking if it was as bad as it sounded. His reply was something like: “Yes. But worse”. In my memory, Lehman fell that afternoon but it may well have been the following day. In any case we were dragged into work super early the next day after Lehman failed to make 100% sure all the systems were working and could run the platform out of hours so people could work on figuring out the banks exposure to all the over-the-counter derivatives that were exploding. Another memory is that on that day I remember asking someone dealing with these complex financial instruments: “How are WE doing?” and his answer was “We don’t have a clue yet. Could be fine, could be not.”. He meant it. I believe many banks were racing to figure out how exposed they were and were they the next Lehman.

This is all to say, we were in the middle of something major, a shift was going to happen. But in the moment I was just riding the wave and keeping the lights on. It’s only with hindsight that I can look back and say “Wow! Wasn’t that at thing!”. Which made me think. We are in the middle of another thing now: The AI revolution. It’s going to have as big of an impact on as many industries as the 2008 crisis did. Perhaps I should take a beat and comprehend more deeply what is going on. What do I need to understand, while I am “working my job”?

The other episode of the series I’d like to call out is the final episode, which ties into the AI theme. Episode8, Michael Burry Speaks. Michael Burry is the central protagonist of the Big Short, he is the guy that discovers craziness going on with mortgage backed securities and found a way of using Credit Default Swaps (CDS) (famously explained in the forth wall breaking fashion, by Margot Robbie in a bath tub in the Big Short movie) to bet against the market. It’s great to hear him speak about that period of time. But towards the end of the podcast he talks about his current short position in the AI market, betting it’s a bubble. In particular he has placing bets against Nvidia and Palantir I found his analysis of both companies to be interesting. But his view on Palantir is really eye opening and one to watch!

Which leads me nicely to my second podcast pick.

Podcast Two - Big Technology Podcast

In my search for news, interviews and analysis on AI in the tech industry, I quickly came across Alex Kantrowitz’ Big Technology podcast. I like Alex’s (self proclaimed, but fair) level headed analysis. He speaks to many of the industry heavy hitters on his show. I enjoy his take and his interview style.

The episode I want to call out is one from a few weeks back. Titled Inside The AI Bubble: Debt, Depreciation, and Losses — With Gil Luria. Gil Luria is the head of technology research at the financial services firm D.A. Davidson. As the title suggests, the episode is about the possibility of a market bubble centred around AI.

Luria calls out The Big Short in his analysis as well as another high profile corporate failure: Enron. Like with learning what Credit Default Swaps were during the 2008 crisis, we all learnt what Special Purpose Vehicles (SPV) are during the Enron scandal. So it’s interesting. Is history echoing?

With both Meta and xAI using SPV’s to move the debt used to build out AI capabilities off their balance sheets. Some folk are getting a bad feeling. However where Enron used SPV’s to hide massive fraud, Meta and xAI are doing this completely above board and in the open.

Luria also mentions that there is limited systemic risk when we are talking about ten’s of billions of dollars (pretty soon we’ll be talking about real money! Right?! Right?!) in debt, but when you start talking about hundreds of billions that’s when systemic risk creeps in. In that light, it feels like OpenAI have read the “to big to fail” playbook and are in fact making themselves to big to fail as a strategy. OpenAI has made a +$1 Trillion dollar debt financed commitment to spend already, which if Luria and others observations are correct, does indeed represent a significant risk to the US economy. As reported in many places, OpenAI’s CFO Sarah Friar already suggested that “the government could provide “the backstop, the guarantee that allows the financing to happen”.

What struck me is how the current market situation looks a lot like the players are learning a playbook from the 2008 crisis while trying to time-the-market around something that looks a lot like another financial event I lived through, the late 90s/early 2000’s dot-com bubble.

There are a few ways to look at the AI bubble and how it is different to the dot-com bubble. Namely the stakes here are very different and there is a winner takes all mentality in play. Which looks to me like how the frontier labs like OpenAI and Anthropic are treating it. Which is different to how the old school companies such as Microsoft, Google, Amazon and Apple are playing it.

Microsoft, Google, Amazon and Apple are largely not financing their AI initiatives with debt, rather they are using cashflow to finance. Which isolates them from much of the systemic risk should a 2008 situation arise. It also allows them to survive the bubble bursting. Again, learning from history, ultimately the exuberance of the dot-com bubble created the infrastructure we needed to drive the mega engine of the internet we have today. While many start-ups perished in 2000, Amazon and eventually AWS was quite literally enabled by it and are now a trillion dollar winner. Like a clever hedge fund, the big players seem to be positioning themselves to take advantage no matter what the future market conditions are. In fact they will probably win big if the bubble does burst.

Two interesting outliers here. Meta who are going big on debt financing could find themselves on the losing side. And Apple who have taken a beating from consumers for their poor AI offering. However unlike Meta, Apple haven’t tried to buy their way in. I wonder if Apple believes there is a bubble in play and are waiting for it to burst to scoop up the cheap talent and infra in the wash up? Apple’s only mistakes being that they gambled and over promised at WWDC 2024 with AI products they failed to deliver and their inability to improve Siri.

How Microsoft are approaching the current market and the AI boom is what leads me to my last podcast recommendation. An interview with Satya Nadella.

Podcast Three - Dwarkesh Podcast

Mentioned in the episode of Big Technology above by Alex Kantrowitz, in this episode of the Dwarkesh Podcast, Dwarkesh Patel and Dylan Patel talk to Microsoft CEO Satya Nadella about build out data centres for AI workloads and Microsoft position on AI in general.

One of the early comments was in relation to GPU depreciation that Gil Luria spoke about on Big Technology. Right out of the gate Nadella is addressing this with regard to Microsoft’s data centre build out. Saying they are building for “scaling over time”. Very interesting. Another early comment was on how Microsoft are linking these huge new data centres together to maximise the scaling across multiple venues. Clearly Microsoft are thinking ahead on these things and its not just over exuberance for AI. Microsoft are using this bubble to build and pay for their data centre infrastructure such that they can capitalise on it beyond and in spite of any bubble. Much like the survivors did with the infrastructure built during the dot-com bubble.

The really interesting thing about this podcast to me is not the tech talk so much, but rather how Satya Nadella himself speaks about it. Ok so this Podcast is clearly a safe place for Nadella with no risk of uncomfortable questions. But he clearly enjoyed talking to Dwarkesh and Dylan and revelled in the discussion. Ok maybe he is extremely well briefed, but he also clearly knows his stuff and can talk extemporaneously and eloquently about everything. Right from the core economics of the market and Microsoft’s position in it down to the tech that is in the servers in his data centres. I don’t think he is faking the enthusiasm. It’s very impressive. The way Nadella talks, AI has reinvigorated him and his company and both are relishing it.

It’s therefore hard to bet against Microsoft being a winner in this game. The same is true of Amazon. The outlier here for me is Google. Despite squandering a lead, they have really accelerated and are fast becoming the leader. Google have the data centre and network infrastructure much like Microsoft, but they also have their own very strong frontier models, very strong AI research lab and their own well established silicon.

As this shakes out in 2026 its going to be interesting to watch who the winners and losers will be. because there has to be some of both right?

Anyway! These three podcasts are a worthy listen. With Dwarkesh and Big Technology rapidly becoming my goto’s for analysis of the AI space. I hope you enjoy the content as much as I do.

A Note on Links to Amazon Products

I have started using Amazon Affiliate links for products I link to on Amazon as way to get a small return on writing this blog. This comes at no cost to you, dear reader, it simply tells Amazon that something you saw here compelled you to purchase it on Amazon, for which Amazon will send me a small commission. If you find this not to your liking, feel free to avoid making purchases via the links supplied above.