Apple as a Bank and what AI means for markets and corporates

We discuss a number of toping relating to markets and society. 

• Apple’s Savings Account and is it a new Financial Player
• Future of retail banking
• Can payments ever be costless?
• The weight and cost of financial regulation
• AGI and open source AI tools lead the way
• The power of an independent AI developer scene
• What does AI mean for companies large and small

Andrej Karpathy: Tesla AI, Self-Driving, Optimus, Aliens, and AGI | Lex Fridman Podcast
https://www.youtube.com/watch?v=cdiD-9MMpb0

Transcript

0:00

hello. Good evening, good afternoon, good morning, wherever you are. This is Off the Fence, a transatlantic podcast where three guys in Madrid and in the New York metropolitan area get together and try to put the world to this week we're gonna talk about developments in banking, in US banking and specifically Apple offering a savings account. It's gonna actually pay a competitive interest rate which should be interesting given the way deposits have been moving in the US and and other and other developments. But Luis, this was your topic. Why didn't you, why don't you take it, tell us what you think what the issues are. Thank you Peter. I think, we discussed this topic in previous podcasts and. My impression is that when you're presented with such a fantastic spread between deposits that are site deposit that pay nothing and federal concentrated 5% or 500 basis points spread between what banks pay for deposits and treasuries short data treasuries, there will be somebody somewhere that will take advantage of that to take market share in deposits and do something with it. And we have, we, we talked about it and we couldn't, I couldn't figure out who would be the disruptor. Apple yesterday broke the ice. It is, in my opinion, a much more significant attacker than any of the attackers that we saw in the early 2000, such as I n g Direct or the online brokers. Various reasons. First of all, apple as an, as the. Owner of the operating system in which a lot of people's lives work has an enormous amount of subscribers already many of whom use different functions that the Apple Pay app allows them to use. It is also a big contender in the buy now pay later layaway system of buying for retail purchases. And therefore the fact that they can get funding at a competitive rate for them and maybe not for them, maybe somebody else who does the actual financial management of both the asset and the liabilities of these in, of this book. I think it's quite a threat to some lenders to consu to some consumer lenders in particular. Unsecured consumer loans. One has to believe that it won't be too long before Amazon thinks that maybe this is also their domain, that they should probably get into this. And perhaps other of the platforms that are involved in either payments or consumer Retail might, will also develop their own special, the interesting savings products with this. I think the one belonged before this trend moves from west to east as these trends generally do. And discuss with you guys in previous podcast. I believe that the current excitement with cyclical Companies, specifically banks in, in the Eurozone might be a little bit ahead of itself because, this would be one of the threats that we could think of for that markets. What do you think? I, go ahead Alex, please. No, please, Peter. I'm having a look on online right now. What big banks. Our paying is an average rate, and it's about 0.1%. And most if you look across smaller banks, they're, they get to around 3%. I see you have some special offers. PNC has an account at 4%. But this is a competitive move. Banks are slow to raise deposit rates. We know that they're a lot quicker to raise their lending rates. It's which is natural. Goldman Sachs, who will be servicing this, took a stab at retail banking in a, with a unit called Marcus that hasn't gone so well. It hasn't been unwound, but it's been it's been reduced that back. Apple and Goldman have been working together for a while. And especially on the card side. I, and, but this is a new and interesting move. It's interesting also cause Apple has none of the capital or I believe liquidity concerns that other banks might have in, in backing up a savings account. And this has become an issue in the last few, we, few months as because Yeah. What sovereign Silicon Valley Bank was was liquidity. The liquidity can move incredibly rapidly. And this is a huge concern for regulators and supervisors. Supervisors right now, I don't think, I don't think Apple offers a deposit cause Apple doesn't own a bank. I think the deposit have to look into the small print is probably deposit at Goldman Sachs or some third party bank. Apple just is the facilitator. I think. The marketer. Yeah. But that's the whole point, right? So few things. That's right. So the re what I'm trying to say, Alex, is that the regulatory constraints on the deposits will be with the department taking bank will be with Goldman. That, that's the whole question. The funny part is that Marcus offers today 3.9% versus four and change percent. You're not gonna your own product are you? Yeah, of course. Exactly. You're gonna love that. And it shows you the power of the thing. So a few things you mentioned at the beginning, the impetus for this and it makes sense is, the fact that the spreads have widen significantly and that the banks being slow to raise rates, as Peter mentioned. Given that we've had, a good 10 years of customers being used to not earning anything in their deposit accounts. This is, marketing wise kind of an important move for a. National player as opposed to a regional player. Regional recently certainly have had to fight for the deposits by paying higher rates. It's a move that makes a whole lot of sense. And then actually like similar to SVB in some ways, what you worry about is with an inflow of significant amount of savings in all one swoop in a fairly big marketing push by a very important player that has a lot of customers, you really wonder where, where those deposits are going to go and who's taking the gap book decision. You'd hope that it's not Apple, as you mentioned, since they don't have expertise, credibility, or even a structure to do it. So as you say, most likely somebody else is doing it. And if it's Goldman, fantastic. It's Goldman Sachs, I mean at the Apple. There you go. Starting today, apple Card users can choose to grow their daily cash rewards with savings account from Goldman Sachs, which offers high yield annual percent in yield, annual percentage yield of four point 15% rate. That's one than 10 times the national average. No fees, no minimum deposits. Yeah. It's a perfect marketing thing, but the argument that I'd like to make, just to move the conversation elsewhere is that for sure we have an opportunity right now because we've had this, the rates back up and expectations have been very low and it takes a while for customers. Customers will look at this as being a good deal. That's great, but my argument is a different one, which is that Apple should. Itself, long term B, in the transaction processing business, because one of the main factors that keeps the wholesale rate of transaction processing in the US at 2% right is fraud. And if there's one thing that the marriage of software and hardware in your hand, has been able to handle is biometric verification and really getting an end-to-end between the receiver and the sender of a if not trusted, but. Verified identity ecosystem, which is quite strong. The apple has both the physical hardware and the software to run transactions. Again, forget crypto, forget, how it gets handled. But the idea is for them to be a payment processor, given that they can ensure what is probably the largest piece. Of the puzzle in terms of the cost of wholesaling, transaction processing I think long term ease and attractive is an attractive business. It could very well be that they wanna grow their marketing shops at this point, which makes sense, and turn that into over time, building a real financial business where they do transaction process. Not take, necessarily back and balance sheet risk, but just be there to make sure that you have the. Funds in your account and that you're making an accurate payment. That to me, I think is a long term play That would be a significant source of growth for Apple. And what does it mean for banks? I don't know if you can hear me. Can you hear me? We can. Yes. And just because of that, I'm gonna answer it. The the problem with banks, I think is that the physical existence of a bank branch may have outlived its relevance. And as such, you have a lot of legacy costs associated in a cost structure that makes it fairly difficult to, to compete. Those kind of banks are gonna be, this is not a new theme. People have been talking about it for 20 years and it hasn't happened, of course. But I think more and more my sense is particularly as you get to alternative ways of processing transaction, which are almost costless, it's gonna be very difficult for people to accept the fees associated with traditional banking. That's, and that I agree completely. And the, and place where fees are most obvious is in payments. Yep. And for years, banks basically had a monopoly on payments. They no longer do. Yet there's still competitive, there's still room for new competition in payments and and Apple is part of it. Others are in the field and and banks are feeling it and bank, especially in Europe. Agreed, yep. I think if I may say something very obvious the big difference between. A bank like i n g direct that in the early nineties started collecting deposits by paying a competitive rate and then didn't know what to do with those deposits. And the current situation is that Apple also originates an asset product, which is the, these layaway loans. And they are the platform where you can do both. Whoever they partner with that runs the both portfolios. We'll be very keen to see that, this works as, as well as it does into the advantage of Apple and the financial partner they have. And I think that if they continue on this vein and their opportunity for other products, it's probably different in various parts of the world, but, I am going to go to meet some people from Apple because, there is an enormous opportunity for a brand like Apple to develop a mass affluent independent financial advisor business in Europe, which is probably one of the areas of the market where there's the most fat in the world. Yes. And and this would be terrific for the vast majority of the. Population of the European Union or and other Europeans. But what I would also remind you of is that a few years ago, Michael Milken, in one of his appearances at one of the conferences with, asked to provide advice to some of these platforms. And what he told them is, don't get into finance because the regulation will bug you down. I talked this morning I should correct myself. I exchanged WhatsApps this morning with one of the most senior people that I have access to WhatsApp at eight o'clock in the morning in finance. He was the chairman of a bank in Spain. He was the vice chairman of one of the top four investment banks for Europe. And I asked him what, where he thoughts about this, and one of the things he said is, this company was doing very well, UN, until or has done very well. Let's hope they don't want to get into the MOAs of being regulated as a financial entity. And I, and my, what I understand is great about this deal with Goldman Sachs is that Goldman Sachs already pays all the fixed costs of being regulated as a financial entity and if they can get the additional business, Of being Apple's joint venture partner in this, everybody's better off and including the public probably. So I think it's a, it's an incredible, incredibly I think it is as material to the banking to the retail banking world as the cash management account in 1973 was to the retail banking world when Marin h introduced that. I think you're right. And there's something else about it, which is in recent years I have heard an argument in banking, which goes something like this. A bank does three things. It takes deposits, it lends money, and it makes payments. It affects payments. The payments business it's going it's, if it's not gone, it's going. Others have entered. You don't really need a banking license to do it. They're companies that can do it without the legacy. The lending business is also is also going private equity and others are now lend money. Amazon lends money to small and medium enterprises and securitize to to fund it. There's a lot going on in the lending space. Buy now, pay later is done by a non-bank, bank financial institution or can be done. So there's a lot going on in that space. But deposits, taking deposits is where the regulation kicks in when you're a deposit taking institution, and thus others weren't gonna go into it. The, they may try to create Anna, the instruments that look like deposits, feel like deposits, but they're not gonna be able to offer customers guaranteed deposits because they don't wanna be regulated. As Luis has just pointed out. This seems to be a way in. You, this partnership with Goldman Sachs to to get a big tech into that space, into the deposit taking space. And these will be deposits because I'm, they'll be guaranteed by the F D I C. So I, I agree. It's, it this is potentially a big move. I agree. I think what's interesting about here is you have two best of breeds, honestly, in, in my mind, between Apple being knowledgeable on the software hardware side, consumer loyalty, privacy brand and obviously Goldman Sachs on probably regulatory management and all different financial capabilities. So you have two best of breeds, which for the moment early are, have compatible. And corresponding strengths in trying to address this. My, my question would be long term, I think to both your points, how forward integrated into this, does Zapp want to be long term? Right now it's easy. It's a marketing deal. That's no problem. Longer term, it's gonna be a business for them. But if it is, it's a completely different business than what they do and would be, enormously risky. And I think for the investment base, for Warren Buffet, for everybody else around would be viewed as something that to due to gingerly. So it's gonna be very interesting to see. And agreed. I think Warren Buffet will probably not be around long enough to see that, but cuz just statistically speaking but. He does like banks and he does like insurance products. And imagine Apple being hub for banking and insurance products and Warren Buffett sitting at the helm. I cannot think of a more powerful platform to steward retail products ever in the history of mankind. Yes, agreed. Completely agree. Again and probably fairer, I think even for everybody around, interestingly enough. That's one would hope. But I can, let me tell you what's happened in banking in a small backward countries such as Spain, 20 years ago, there were, Peter would probably know the numbers better, but probably about 80, 90 banks of national, of them with branches in Madrid. Yeah. Barcelona and some of the largest cities. And banking services were competitively priced because there was a significant amount of competition. Nowadays we're down to a handful of banks that could, with large market shares. I think the top five banks in Spain probably have, I don't know, 50, 60% market share. Maybe probably know the numbers better. I'd say more. More so what's happened as a result of the great financial crisis and the European. And Nike Crisis is that with an enormous consolidation. And as a result of that consolidation, people expected that regardless of where the ECBs set the deposit rate they would have Sorry. They would, they would be able to pay 0% for deposits. This is the argument that nine, nine out of 10 bank analyst had to buy European banks, that the jaws would work exceptionally well in this cycle, that there would be no increase in the funding costs from deposits, and then you would get all the benefits of. With loans and the reason they could be sand about such a strange prediction was that there's so much less competition. And they tested that in the UK market, which is a very concentrated market. And the, until the hiccups of October and even after that, there hasn't been a lot of these termination of bank deposits in spite of everything. I think that the svb and the other two banks with the plus the pretty sweet situation may have changed that, sense of, safety of having a bank deposit for many people. And then this message from Apple, and if I'm right, and Amazon comes out within the next few weeks and has its own deposit product with some other financial partner I, we are going to have another financial crisis in the making within a couple of years I and focused on deposits. I think that's one scenario. I think you're right that there are fewer competitors, so less competition. Another element of what's going on in Europe is the quantitative quantitative easing, which is still going on. The money supply is the market is still a wash in liquidity, and it's not until June that the European Central Bank is going to actually go into quantitative tightening mode. Let's see if that, let's see if that changes the dynamic on deposits. May I just say something on that, which is that part of the qualitative quantitative typing in Europe has. It's too complicated for a normal person to understand, which is the reduction in tlt r o Yeah. Funding and that is going quite fast, right? So Yeah. And then it was one of the key programs for providing liquidity to the banking system. So maybe we can shift to a less financial topic. Let's look at ai. Oh, fun. What, Alex you wanna give us an update on how we're going and getting to artificial general intelligence? Agi, so in a godlike manner, so enormously a rapid sub update. I think everybody's seen, obviously stable diffusion six months ago DLI and so forth. On the image side, everybody a couple of months ago on the G P T three, the three five, then four side. And it's important to understand more or less, what happened. The most important thing to think about, I think, and understanding landscape is that for years people have been publishing papers, working very hard at trying to understand how in a lab, how to use artificial intelligence. And at the Genesis, this is back 20 15, 20 16, OpenAI was built to try to do this independently from the larger companies. So again, Google, Facebook Microsoft many people, one after. And Google has two dedicated teams on ai. Very powerful teams both going after it. And the idea when Dali was first released, which was a shot across the bow to all the larger companies. Back in September and then when G P T three was released and quickly upgraded. What happened was you now started to have an open source, a set of tools that people could use. The pricing of the a p i of open AI was so low or could remain so low that it provides for an enormous amount of capability for people to go and try things and to basically offer. AI ish or AI products, which are really just the GPT APIs, meaning that the ability to feed onto G P T, whatever it is that you're doing. So you've seen all kinds of other companies larger companies come out with tools, those that were working on AI before. So things like Adobe and obviously Google and everybody else has come up with stuff. And then those that have just resold their bulk access or wholesaling access to to open ai. The fact that it's open source is enormously powerful because what it does is that it takes away from people in lab coats the control that they had in some very interesting interviews of Eric Schmidt from a couple years ago, who was adamant that all of his AI stuff, for all the reasons you know, that ethicist will, will bring up, should be kept very controlled and inside the lab and so on, so forth. That cat is. That Legion of cat is completely out of bag by now. So what you're seeing is that you're the main limita. You're seeing a lot of development in all kinds of different ways. Most of the activity on GitHub, which is the place where developers particularly open source of developers, share code. It's now all ai all the time, 24 hours a day which is quite interesting. So you're starting to see a lot of things being developed. The thing to think about right now, I think just to summarize it and I'll open it up, is there are three things I think there are worth thinking about. The first thing is that these models have been trained a certain particular with certain particular set of data. Some of it cooperated again with a cutoff date like about a year ago. I. But that's not the future of models. Models are going to be nightly builds. It's going to be something in which, Morgan Stanley or any organization where they have an internal one or an external one facing customers are going to have the, to retrain or to cont continuously train on new information. These master models that they're gonna have that're gonna make available to customers or make available to employees in order to be able to find pretty much anything that you have as institutional knowledge in your company. And that's a gonna be a. Fairly large business that's gonna require a lot of effort and handholding because again, you're democrat democratizing a a technology. The second thing is that right now there's this concept of how many tokens, or let's say how many words you can feed. Into a thread. So Che g p t, for example, tells you that they are threads and people and that the threads will remember what you said. That's not entirely true. Basically, the way to think about it is anywhere between, in the beginning 2000 words, which are really tokens, but let's just say words for simplicity. 2000 words is its memory, meaning that if you give it. 10 times 200 words. So if you ask it a question for 200 words, then get a response and ask it again, another question, or remember the first 200 words and actually the answer they gave you. So it has a little bit of memory for a short period of time, but one of the biggest problems out there is to try to give, its more short. Term or long-term memory. And so you've seen this development of what they call vector databases, which are basically ways in which you can add some memory to the model with having to retrain it. If you train the model, you add permanently the data to the model. That's a good thing, but very expensive, very complicated. But the idea is that you want to have kind of a memory buffer. The way to think about it is like if you have 10 years of income statements or 50 product PDFs or anything like that you really want those to be available, not, to, to j to your AI to share G B T. In order to be able to answer correctly, again, because these generalized models, can answer generalized things, not specific things about your knowledge. And then the third piece, which is fascinating is, which has just started the last couple of weeks, is this concept. There's a very famous project called Auto G P T, which has been Having a lot of activity and recently, but basically is the idea of user agents. So the idea is that if you prompt a model meaning you write it a job description saying you are an expert at marketing and you know everything there is to do about the four Ps and you know everything about all this stuff, and you give it a persona, then it's going to answer in that particular way. You can replicate that in a bunch of different ways where you spin up multiple models and you give them different personas. So you can create yourself a c e O persona, which is, or a team leader persona who is in whose job it is to achieve certain goals. And then you have it interact with other models who are specialized in finance and operations and marketing and HR and whatever it is, or whatever the components of your project is. And then you just, let it loose, give it a, a goal, and you just watch, while they have all kinds of conversations about trying to accomplish the goal at hand, again, it's very early. But this gives you an insight as to how you can get really deep insight and thoughtful insight from, from the systems, soon enough. So to close it out, this is enormously early. It is impressive. It has made people realize this is not 10 years from, 10, 10 years away, or 20 years away. And it is just as exciting as it is. Scary. That's all. I'll say fascinating time. Fascinating time. What do you think about the petition to Pause? So I think it is a admirable view. There's a couple of interesting interviews. Lex Frigman has a couple interviews, which are interesting, if any, was interested on the reasons for why you should take the time to pause it. It is laudable. It is never gonna happen. Okay. I feel like I am the ludite, but I talked to Chad GBT often and I have a subscription. And so far from its own account, it tells me that, not in the exact same words, but that it's a very methodical librarian and it doesn't have the ability to do any original thinking beyond what it can find in the library and what I thought would happen. And what is happening with chat GBTs. Two different things I thought that we were gonna go. So this reminds me a little bit of what my friends were doing out of engineering school in the late eighties, early nineties, that they were doing artificial intelligence as an ency effort to a lot of information into databases so that a processor could use the databases to come out with answers that were stock answers to stock question. I thought when Anna and I were roommates in New York and he introduced me to a game called Sin City, that apparently he developed some organizational skills from having small now I forget the words, but this is, goes back to the early work of the Yes. Mathematic. Yep. Which is that you have, from similar initial conditions, these, how do you call these agents? I forget. Alex, these agents that would find different paths to do different things. You can call'em agents. Agents. Agents, yeah. And I thought that was far more interesting cause they would come up with their own solution by playing the game. And when you get to. The game of chess. For instance, there were agents that came up with a way to play chess that was systematically unbeatable. Yep. Some people said, okay chess is a, is a simple game because there's always an optimal solution to any position. Let's try a different game. And they try and go, which is a Japanese game of occupation, of a territory from the enemy. You might be committed with that game has white beads and black beads and it is, they nobody thought that it would take, that it would be easy to program a computer to play Go. And I think it was within like 36 hours a program was able to be the world. Yeah. And I thought for the, let me interrupt you just for one second. Just to give you one piece of insight, which is helpful cuz I remember we, we financed a bunch of companies that were trying to do this kind of stuff about 20 years ago, and it was a mess and it was very much rule-based and so on, so forth. But I would pause at the fall, the following thing to try to reconcile the two views. Your brain essentially learns the same way as a language learning model. Let's simplify, right? So a deep learning model, and the reason I say that is because you as a child, experience certain things. So for example, if you put your hand. Onto a hot stove, you will associate or existing nerve endings that were built earlier, even in your life about moving your hand. And then you will make a connection with the fact that touching the stove, obviously creates heat. And I, if you were to be able to live as many lives as. A language learning model does, which is millions and millions in lives in parallel, right? So it's touching millions of stoves and in certain particular ways, in every possible way. What's happening is, which is interesting, is, and this is your point, I think Compared to the way people thought it would be done, which is by, making sure that we understand all the rules, that we're a good librarian, we know what section of the library to go and look for whatever, where the knowledge is stored. It turns out that language these models have are matrices of numbers and weights. That's all they are. They're literally the equivalent of these connections that if this happens there, then I should go this way, not that way. Kind of thing. Super simple in some ways, but at a level that computing has only been able to make available in the last couple of years. Let's just stuff that would've been massively too big to do. So the argument I would make just in finishing is that we have in some ways replicated with the way the human mind. Develops at the early age and even later stage in matrices in a mathematical algorithm, train them. And so I would argue that it is able to learn. At least as well as a human can. And I know that's a very powerful statement. We may not see it today but I think the argument is that, that, we're, the way that it comes up with the way it thinks is very similar to the way the brain, I think. I think also if you, Luis, you're probably using it a lot for our, the, for the types of issues and topics that we get into and are curious about, and. And talk about and discuss. If you ask it to write a short poem in the style of Elliot about the restaurant across the street, I think you'd find you'd get a, you'd get blown away by by what it can do stylistically. And and, do I say it creatively? Yeah. Yeah. Bloomberg trained the model and when you look at the paper, they just did, released it last week. And what's funny about it is that the Bloomberg data, their propriety data the Crown rules, those things are the most important thing. Only counter for 78 basis points for less than 1% of the training data they put into it. So there's a wall to go before you see something, but when that model gets trained with actual Bloomberg data, it'll be a site to see. I think Think that's, I think we're gonna, we're gonna call it a day and call it and say thanks to everyone. Thanks very much. Thanks, Luis. And take care, Alex. Thanks for organizing this. Yes, thank you very much. Thank you. Bye. Talk soon. Bye. Take care.