Bad outlook for SAP, ServiceNow, Mastercard, and Visa

00:00:00: You know, you hear today a lot of people saying oh AI it's not capable and makes errors.

00:00:04: It hallucinates.

00:00:06: I think the people who are saying that or still looking backwards six months ten months ago when they look at what is possible to them The issue with NOT is not capable enough.

00:00:15: its too capable Its too cheap And now essentially the price of intelligence Is going drop down to the price electricity.

00:00:28: Welcome to the Big Bang Tech Report, where we decode technologies reshaping business power and society.

00:00:37: Is AI high-peaking or are we entering a winner?

00:00:41: takes it all area?

00:00:43: This could be the headline of our conversation.

00:00:46: Alvin great to see you!

00:00:48: Good to see again.

00:00:49: Jens I'm looking forward to your conversation.

00:00:51: as We said earlier there's just so much happening.

00:00:54: So i think uh...we will have A lot to discuss today.

00:00:57: Isn't it incredible?

00:00:58: In only fourteen days, what's happening in the world.

00:01:02: Yeah pretty much every single lab in the World has released a new version and they are all dramatically better than the prior

00:01:08: ones.

00:01:08: And uh...the speed of change?

00:01:10: I think that is the difference i'm worried about.

00:01:15: Let us talk about this.

00:01:16: But first we have had topic number one.

00:01:19: We've got five topics.

00:01:21: The first one is we're having a headline in Germany.

00:01:23: That Warren Buffett Everybody knows Warren Buffet that he has sold Apple and Amazon shares.

00:01:30: So the hype is over?

00:01:34: Well, I mean i think anybody who follows Warren Buffett understand He's a value investor right And value investors buy things That are undervalued.

00:01:44: and The reality Is that you know companies like You Know Amazon or Apple or Nvidia Has gone up dramatically Over the last you know, couple of three years.

00:01:55: So the fact that he's taking some money off the table I think it is actually very natural.

00:01:59: but one interesting thing was added more Google.

00:02:06: so its not an issue of tech.

00:02:08: i think hes doing a rotation to see which ones are valued properly and undervalued.

00:02:13: we discussed google.

00:02:14: remember few episodes ago that

00:02:22: Google is in advance.

00:02:23: Do you see there too, we have so many information about all the models and who's leading?

00:02:29: Who was winning?

00:02:30: this is our topic number two where do you sees a different model right now?

00:02:35: yeah

00:02:35: I mean i think at the end of day it not an issue whether or not google has better model.

00:02:41: they actually are definitely one of top models.

00:02:43: but what thing to say?

00:02:46: And really nobody else right now on the market has that.

00:02:50: They have chips, they have users and data models...and they also have an ecosystem.

00:02:57: So when you have everything…they can work with a very large revenue base so they don't need to borrow money whereas all other labs are either getting investments or loans to sustain their business because they're all money-losing businesses.

00:03:13: So I think that's the difference, it is not about technology but the whole ecosystem around them.

00:03:19: For you Google first?

00:03:22: I think Google in a very safe place.

00:03:25: Companies like OpenAI and even Anthropic are in precarious places as they so dependent on external revenue They losing ten plus billion dollars per quarter right now And that means they need to continue to raise money and announcing you're going to raise fifty or a hundred billion dollars over the next few months.

00:03:45: But again, all depends on investor sentiment and sentiment changes very quickly as we know.

00:03:52: so... And if any time they don't raise what happens?

00:03:56: is that?

00:03:56: They cannot you know build out the type of things That they have planned.

00:04:00: They start to have to cancel orders The trust and their the expectations go down, and then everything else becomes much more difficult.

00:04:09: I've had the headline here in Germany that open AI was buying something.

00:04:12: What was it?

00:04:14: Yeah yeah Open AI bought an open claw.

00:04:16: remember two weeks ago we talked about open claw or you know claw bot.

00:04:22: But the fact that, you know they essentially negotiated a deal where They bought this one person company remember earlier You asked is there a possibility for it?

00:04:30: One-person company to create a billion dollar company?

00:04:32: and in fact You know we're essentially within a few weeks of us talking about It happened then.

00:04:36: There was an agent based company one guy, you Know in Austria who probably sold for somewhere around a billion dollars to open AI.

00:04:45: so So I think that actually is very consequential because Open Claw to me, it's the deep-seek moment or a chatGDP moment for agents.

00:04:59: And now everybody in the world talks about agents and you know... Everybody's experimenting with it!

00:05:06: Now i think this trajectory of this industry will change Going forward whether or not they become the winner.

00:05:14: It's not really important But I think it's now spawned off hundreds of people building on top Of their code base to make other agents.

00:05:22: but until now we do Not have any impression here especially in Germany that The agents are coming and that agent are more important if you say okay, let's have a look and so on but Chess yeah small as all this generalized ball as everywhere but agents.

00:05:40: so we do not follow right now.

00:05:43: Yeah, I think the difference is Agents are a little bit more complicated from technology perspective to set up and it takes a little training for you take advantage of it And its relatively new.

00:05:57: So that fact hasn't become part of popular discussion.

00:06:03: It's actually less important.

00:06:04: But if talk developers or enterprise IT people they're all talking about agents right now.

00:06:10: And, and Right Now when you create a team of Agents to orchestrate it towards the task It is able to complete The tasks Of You know ten developers in one or two days.

00:06:22: What would take A month for these Developers To do?

00:06:25: So this Is very dramatic.

00:06:28: In fact if we look at last few Days on the stock market Last Two weeks actually All the companies that are losing Value On the Stock Market Because of Agent Technology is because of the fact that people are saying, oh I don't need to build a creative suite with Adobe anymore.

00:06:44: Oh you know yesterday IBM dropped by thirteen percent!

00:06:47: Because clock code says i can now code in Kobo and Kobo used be something that you know.

00:06:54: IBM their mainframes use.

00:06:57: so everybody expected them have it as service revenue for very long time.

00:07:02: but if somebody have an AI agent do the work, then that service revenue goes away.

00:07:09: So there's a lot of real underlying issues that are happening.

00:07:17: People do not know exactly what's going on and you sent me an WhatsApp, I think one day ago.

00:07:25: And I read it as said okay this is a real wake-up call.

00:07:29: i'm talking about the US report Citrini research which is going viral and The headline is that the twenty twenty eight global intelligence crisis.

00:07:39: so yeah This Is A Study about the future and it looks back to the present.

00:07:45: So tell us what's going on there?

00:07:48: I think you have to take everything with a slight grain of salt, uh, with these reports because it is an essentially near-future fiction right?

00:07:57: so What they're doing is forecasting what could happen And then It's a way of telling a story To get people's attention which ii think its actually very valid possible future.

00:08:09: And it really depends on what we as a society, and also what the governments do in the near term to help reduce that shock because what they're predicting is by twenty-twenty eight The economy and the job market collapses Right?

00:08:26: We don't need go into all of these details but there's actually lot of them Because of what just talked about.

00:08:30: The fact that AI agent can so much So capable so cheaply so quickly that all the companies will choose to use agent AI versus humans.

00:08:43: And then, entire economy starts to collapse because now there's no more wages and nobody buying it but the whole cycle becomes a negative reinforcing cycle.

00:08:56: I think this is possible Because you hear today people saying AI isn't capable of making errors or hallucinations.

00:09:05: I think the people who are saying that or still looking backwards six months, ten months ago when they look at what's possible today.

00:09:11: The issue is not it's not capable enough.

00:09:13: It says too capable its to cheap.

00:09:16: and now essentially the price of the price Of intelligence is going to drop down to the price electricity.

00:09:23: And then that changes

00:09:26: again.

00:09:26: We have we after stress it again?

00:09:28: The price of Intelligence Is Going Down To the Price of Electricity.

00:09:34: Wow

00:09:37: Here's a piece of data to think about.

00:09:39: At least in the US, around sixty-five percent are cognitive workers white collar workers.

00:09:46: They account for about seventy or seventy five percent of the payroll.

00:09:50: they account for eighty plus percent of spending.

00:09:55: So if that class is decimated.

00:10:00: Maybe not even a hundred percent, let's say fifty percent or twenty percent how would the economy look?

00:10:06: And it will not look very good at all and according to them its gonna take to twenty-twenty eight before happens.

00:10:15: I actually think this going happen maybe faster If we let the natural forces go, it will go faster.

00:10:22: In fact I saw an interview this morning with the author of that Caitriona research paper and he said they wanted to make a far away enough doesn't scare people too much but close enough for them.

00:10:32: pay attention.

00:10:34: They actually think could happen fast or not.

00:10:36: And then i agree With

00:10:38: you know The attention is currency of journalism But once again For everybody.

00:10:45: Companies replace workers with AI, okay?

00:10:48: Check.

00:10:50: As they save costs and buy more AI Okay check which enables more layoffs.

00:10:57: And then displaced worker spend less Demand weakness in the markets and companies automate even more to protect margins and their eye keeps getting better and cheaper like you have told us and the cycle accelerates.

00:11:12: And this is really a dramatic way we are going, right?

00:11:18: Yeah.

00:11:19: I think beyond that it's not just the economic and financial aspect It's the societal stability aspect.

00:11:26: That's the thing they don't talk about.

00:11:28: They talked about the economic impact.

00:11:30: Actually what's more important there will be social instability When you start having twenty-thirty percent of your population becoming unemployed, this is what happened in Germany before the World War II.

00:11:41: This was happening during Great Depression.

00:11:45: that level of unemployment and social fabric breaks.

00:11:51: when it happens we don't know whats going to happen.

00:11:55: That's a scary part.

00:11:57: Well, we hope that it won't come.

00:12:00: But once again to the study they talk about companies too like ServiceNow SAP.

00:12:08: big software companies normally would say oh wow!

00:12:11: They are so stable and fantastic.

00:12:13: but what's going on with these companies when their agents coming?

00:12:18: Well, I mean that this is the issue.

00:12:21: They're all software companies and rather than saying hey if i'm a Business do I spend hundreds of thousands?

00:12:28: A year on This software or millions of dollars a year in The software Or do I just ask my agent to build what I need?

00:12:35: maybe it's not a full function But you know Maybe another few days or week or month I can Build It And they cost me You Know a Few Thousand Dollars or Maybe uh, ten thousand dollars instead of paying a few million dollars.

00:12:49: Right?

00:12:49: And I think a lot of companies when they start to have margin pressure When They Start To See What's Possible With This Technology That Choice Becomes Very Clear.

00:13:00: Yeah and um then We Have The Financial Area Mastercard and Visa.

00:13:06: what's going on with them?

00:13:07: i think they are very secure or not?

00:13:11: Well so this is.

00:13:12: the other issue Is that In the past, people talk about humans had locked in.

00:13:19: They have brand loyalty.

00:13:21: they make decisions based on more other habits than out of rationality.

00:13:27: but what's... What people are talking now is that we will essentially an agent layer to do commerce Whether it's the B to B commerce or be this e-commerce.

00:13:36: if you say hey I have This open claw agent that will now help me buy everything.

00:13:40: I need i just tell a I Need The cheapest toilet paper, the cheapest car Or the cheapest you know rental on my when My travel?

00:13:48: It Will go and find That deal for Me.

00:13:49: it doesn't care what payment system I'm using.

00:13:52: it Doesn't Care which advertiser put What adds On.

00:13:55: it Just Finds the best Deal For me Right.

00:13:57: and so it takes all the friction out of this system, And It Takes All The Loyalty Out Of This System.

00:14:02: So You May Have Direct Transactions.

00:14:06: I Think The Financial Industry Will Be Disrupted.

00:14:11: Its Unclear How Quickly The Payments Sight Because They will Still Probably Use Payments Although A Lot Of People Are Saying Agents Will Probably Use Crypto Payments because its More Directly Accessible To Them.

00:14:23: But I don't know if there's enough of the rails around that yet.

00:14:28: That is also a possible place to go, as well as financial services If you're stockbroker or somebody usually because your accounts are there it's hard to move.

00:14:42: but in future someone says hey broker that can give you services at one third the price, and they can do the same thing.

00:14:52: You say okay agent go make it happen.

00:14:54: so there will be very soon little loyalty in terms of big banks and the big brokages.

00:15:00: So There's a lot things people aren't considering.

00:15:05: And all this does is takes away the entire removed margin out of system.

00:15:09: It makes the system more efficient.

00:15:12: This going to an issue across-the board on all of the industries?

00:15:17: Well, a lot people live very well because of the inefficiency that markets are not efficient.

00:15:25: You know especially why do we need one hundred?

00:15:28: in Germany We have more than one hundred insurances Because nobody knows which is the best and there's a lot of things that are different.

00:15:37: And so you trust on or rely maybe on the best agent.

00:15:44: The friction of switching, right?

00:15:48: You have to get different validations.

00:15:51: now all That can be done.

00:15:52: for you just say give me the best price Give me the most best rating.

00:15:59: I'm really worried because every part when all those efficiencies are taking out, that means all of those wages you're taking out.

00:16:06: And that means All Of Those Jobs Are Going Away right?

00:16:09: So That's The Big Issue and I don't think our economies is in our societies ready for it and Our Governments Are Definitely Not Preparing For It.

00:16:19: And so we have AI compresses both value chains and purchasing power.

00:16:26: The real disruption path is automation, weaker demand and more automations.

00:16:30: So there's something going on like a wheel right?

00:16:36: I think the thing that you need to start getting governments to wake up.

00:16:40: They can step in help slow it down.

00:16:44: I'm not saying slow down technical progress, but slow down the transition because natural economic forces will force us to be very increasingly accelerate.

00:16:55: Because of what that flywheel you just described.

00:16:58: so we need make it harder for people to be fired and where move from a five-day work week or four day work week to three days' work week even if they get more efficient still have some people around.

00:17:10: So economy has time.

00:17:13: Right, and we talked about this before the prior industrial revolutions took eighty sixty forty years to play out And an economy can adapt to that when it has time.

00:17:23: Mm-hmm There will be new type of rose.

00:17:26: they not may not be paying jobs But there'll be new rows made for people because We will I think become more and more a service oriented More and more human in human relationship oriented which is fine Which is actually a good thing?

00:17:39: Yeah but you need to give The social safety net so that the people who lose their jobs has a soft landing.

00:17:47: Yeah, oh That's very very important and I think well the politician have to learn about it And they have to do something To build parachutes or something like that.

00:17:56: But right now?

00:17:58: I think i have the impression but the technology is entering the political scene too Especially in the u.s.. In China.

00:18:05: and then you have the impressions if you are an advance If you have A superior system, a superior AI ecosystem.

00:18:14: You can do everything by yourself and you don't need any partnerships or the NATO and all this stuff.

00:18:22: Do your share of these impressions?

00:18:27: There are people in America who believe that but I think they're mistaken And they don't realize how interconnected the world is, whether from a supply chain perspective or financial perspective.

00:18:45: I think their overplaying current position.

00:18:49: America has definitely right now in lead overall on AI but it's also so dependent everything else that happening around to continue And it is also dependent on a stable society, stable financial markets for that progress to continue.

00:19:10: Because if you think about all of this... If we look at the last year, America's... Eighty percent growth in America's GDP was based on AI companies building data centers.

00:19:25: So I mean, this is the problem.

00:19:28: When one sector accounts for so much of the growth essentially without that it's no-growth.

00:19:37: and if people stop investing... If companies and investors stopped putting hundreds of billions or trillions of dollars into space That building stops construction stops, people then say how are they going to make more money?

00:19:56: Because the whole story has been about you build more data centers.

00:19:59: You make more tokens.

00:20:00: You sell more tokens and get more revenue.

00:20:02: And what I think a lot of these labs who were telling this stories that we're not also telling is that as You're not creating a growing economy, you are creating more efficient economy.

00:20:19: A productive economy but actually destroying GDP.

00:20:24: Removing value from the market.

00:20:26: because removing inefficiencies and efficiency today that accounts for GDP.

00:20:32: it's wages The living capability of all those people is what creates GDP.

00:20:42: And when you stop hiring, when you stopped spending on wages there's no cycle.

00:20:49: There is no cycle of consumerism.

00:20:54: and the building a data center only lasts few months.

00:20:59: maybe year after that may be.

00:21:02: a dozen people runs at datacenter.

00:21:04: No more jobs to build it but temporary thing.

00:21:09: so what happens?

00:21:10: to the, you know a million workers that were displaced by data center.

00:21:14: Right?

00:21:14: Those are kind of longer term issues these lab heads aren't saying and VCs who're backing them when they go back to Washington DC.

00:21:25: And I think it's travesty actually.

00:21:30: So what do we have solution for us now?

00:21:34: I know well, I mean II wrote about this You know in December when I released that Stanford paper.

00:21:40: Yeah beyond rivalry.

00:21:41: and it talks About what to do?

00:21:42: It talks about how do we How do we work together?

00:21:45: so they're instead of building a hundred different data center at Hundred Different models on you know thousands of data centers Let's start to work Together more because This technology is the public but it Is being shared it as Being commoditized by open source.

00:21:57: So The value is not real.

00:22:02: Open AI's going to supposedly go public this year at one-to-one and a half trillion dollars based on ten to twenty billion dollars of revenue, it's not real!

00:22:13: It's insane

00:22:14: isn't

00:22:15: it?

00:22:16: And based on losing something like forty to fifty billion dollars so that the thing Just just the economics of that doesn't work.

00:22:26: and you know companies like Anthropic is raising I think fifty billion dollars on around four hundred billion dollars evaluation, right?

00:22:34: And You know it's this all these companies.

00:22:38: They are just being pushed by The financial sector because everybody wants a new round to be hired in the last round.

00:22:46: if any time you start having A down-round around where the next round.

00:22:49: It's actually worth less than tire house of cards will collapse.

00:22:54: Okay, and I know it from the study we talked about.

00:22:58: It should happen in autumn twenty-seven.

00:23:01: they wrote that a crash will come.

00:23:05: but okay We have plenty of time to look at this.

00:23:10: Yeah, yeah.

00:23:11: And I think going back to my paper is that we start need the government to come in.

00:23:16: they need to create a social safety net what i call a GI bill for DIH so that they can create a South landing for the workers.

00:23:23: We also need to created a Marshall plan for AI.

00:23:25: So then we started creating more equality of resources To global south and developed countries.

00:23:32: so everybody rises together because you're not looking.

00:23:35: You don't need trillions of dollars maybe tens or hundreds of billion dollars globally, and everybody has enough.

00:23:41: Because right now why don't we have enough?

00:23:43: It's because we are all competing over the same resources when we start to share resources.

00:23:47: We already have five thousand data centers in America about a hundred something In India you know probably five six hundred in Europe.

00:23:58: So there actually a lot of data centers A lot of compute resources And that they're actually one Of them.

00:24:03: most important things we didn't talk About is at All these Things.

00:24:06: All of the models are getting more and more efficient, smaller.

00:24:10: So that very soon they will run on your laptop or phone And it'll be good enough for most people.

00:24:17: They don't need to all go into cloud.

00:24:20: Not everybody is trying solve cancer or create quantum physics.

00:24:25: I feel like we have to change narrative

00:24:31: But thats difficult because especially in the financial market The winner takes a toll.

00:24:36: Do you believe that one day, like chat GBT a model was coming said oh I have the best way for agents and then this is a winner.

00:24:45: And the other ones they will lose immediately?

00:24:49: No no i think it's just the false narrative because every single VC in the valley thinks that their investment is going to be winner takes all.

00:24:58: So they're betting, I hear things like this.

00:25:01: They say oh the global economy.

00:25:02: a hundred and ten trillion dollars.

00:25:04: you know we are only investing one or two trillion and will be able to capture forty-to fifty trillion of that hundred trillion dollars right?

00:25:13: It's delusional because first of all there actually destroy economic value so that that hundred and And then the value goes down to what we said earlier, electricity.

00:25:25: So when you're paying a trillion dollars and an electricity is actually only about ten percent of the cost of operating data centers so your investment becomes worth around one hundred billion dollars right?

00:25:41: There's no return on that investment today.

00:25:45: Companies like Apple will do better than most because they are not wasting the hundred billion dollars that companies like Amazon or Google, or Microsoft are spending.

00:25:58: They're spending a fraction on these data sensors because they want to keep cash and I think when-when the market crashes... ...they will have a lot of cash.

00:26:06: go buy all those resources for pennies on the dollar.

00:26:10: just what happened after the bubble crash for internet for all those data centers.

00:26:17: So this is something

00:26:18: that people don't talk about,

00:26:19: That all these models were the cum commodities?

00:26:23: They already are commodities.

00:26:24: I think if you look at what happened over last two weeks All of the models came out in China.

00:26:29: open source.

00:26:30: one-tenths cost to operate and they're open sourced and free.

00:26:35: You can download it And run anywhere.

00:26:38: Why do i have use expensive models running with cloud or OpenAI?

00:26:46: In fact, if you look at the data I just looked yesterday on Open Router which is AI companies who decide what models to use right now.

00:26:54: Four out of the top five models for traffic or open router are Chinese open source models because they're so much cheaper than Western closed models.

00:27:07: And so we are, we're heading to the topic number five.

00:27:12: What's on your mind?

00:27:13: We have it in every session here.

00:27:16: what is on our minds?

00:27:17: you've explained so much things.

00:27:19: if there left something that you want share with us

00:27:22: Well I think we been talking about this whole time but i think whats really on my mind?

00:27:31: how do get governments around the world wake up?

00:27:34: That one thing Your audience should be thinking about and they should be doing something about it because we have a small window of probably six months to twelve months To start doing something before the impact of this technology Gets hold to.

00:27:53: apart that where?

00:27:54: It'll be hard to return Hard-to-fix right.

00:27:57: We it is still fixable, but there's still.

00:27:59: we're still able to to change the trajectory of where the economy goes and how quickly this happens.

00:28:06: I'm not saying slow down the technology, I'm saying slowdown the economic and social impact of this technology so that the economy and society can adapt to it.

00:28:18: That's key.

00:28:19: So How Can The Government Get Smarter?

00:28:22: How Can They Start To Put In Preparations And Regulations.

00:28:29: More expensive is the problem.

00:28:31: It's not as too expensive.

00:28:32: The promise has to cheap and when it's too deep decision to use AI instead of using a human will happen too quickly.

00:28:39: so we need to be able To Make AI expensive enough that the decision was slowed down.

00:28:45: We'll need to make then the decision of Reducing workforce expensive.

00:28:50: Enough, but it's actually worth it for the humans For the leaders of these companies two to make slower decisions on reduction.

00:28:59: So I'm just finishing a new report for Stanford.

00:29:02: I've been working on it for five months and one of the interesting findings is that even with last year's technology, what CEOs decide to do?

00:29:14: Forty-two percent at a time is to cut workforce when they get productivity gains.

00:29:19: And if you think about what's going to happen next year in this year Yeah With the new technologies and what is capable now That number will only go up.

00:29:28: And if you see the adoption of this technology go up, the way that we are expecting.

00:29:35: Forty-two percent reduction in workforce is going to decimate the economy.

00:29:42: But normally if we have to change something like a crisis otherwise... We wouldn't change it!

00:29:52: You need a heart attack, and then you will have to do sport.

00:29:55: And all this stuff... Yeah!

00:29:56: You need something that can change your behavior?

00:30:00: Do we need like a crash or crisis in order for us to rethink our behaviors?

00:30:08: Sadly I think you're right but there are other people who would change their minds from more psychological experiences.

00:30:19: They go to Burning Man and have an enlightenment, or maybe they take ayahuasca.

00:30:24: And then start to see the light right.

00:30:26: there are different ways too change your mind.

00:30:29: It doesn't always result in a heart because of a heart attack or because you had a life threatening situation but we'll see what's possible?

00:30:41: I hope that our leaders wake up before We Have A Crisis We Cannot Return From.

00:30:48: So I mean the time is over and we have three for me core messages.

00:30:53: AI isn't slowing competition.

00:30:55: It's intensifying.

00:30:57: Technology is entering in our political systems, and the third one is autonomous Agents remain the next frontier.

00:31:06: that Is are they message from this session?

00:31:09: Yeah you want to add something on

00:31:11: it only accelerating every day its getting faster.

00:31:15: That's what people don't understand is that it's already amazingly quick, but its only gonna get faster.

00:31:21: So we have to act.

00:31:23: We have the start preparing.

00:31:24: thats the key.

00:31:26: so thank you very much Alvin.

00:31:28: fasten your seatbelt Because of whats going on.

00:31:32: and

00:31:33: Yes Well hopefully we will have some better... Hopefully we'll also have some good news next time.

00:31:40: Please, please send us better news.

00:31:41: We need it because spring is coming in Germany I think and the US too And we need some hope for everything For you guys subscribe to Big Bang Tech Report where we track the forces shaping future.

00:31:59: See you next time

00:32:01: with good news!