E168: Can Google save itself? Abolish HR, AI takes over Customer Support, Reddit IPO teardown

Yes, Klarna replaced 700 people and they saved $40 million of OPEX, but Teleperformance, while they were just doing their everyday work, lost $1.8 billion of their market cap at the exact same moment. ... But I just wanted to show you that the destruction was quite quick and it was pretty severe. And if two or three other big companies launch these kinds of tweets after real measurable results, Teleperformance will be a $1 billion company in short order.

So, what Klarna seems to be talking about are email-based customer support cases. I think where this is going to go next is to phone. ... I think where it goes next is you'll call up the call center and you'll get a voice that sounds like a human, just talk to you. And you won't even necessarily realize that you're talking to an AI, because there are already these AI companies that can do generative voices, any language, any accent.

So there may be a moment here where Google stock, which currently is trading at just 17 times 2025 consensus earnings, which is cheaper than all the other big tech companies by far and is still growing. ... So the bull case is now is a great time to buy because it's so cheap and there could be this moment where you see some of the changes that are needed internally to get the AI products to where they need to be to maintain the lead that is inherent because of search.

Do I think that Larry and Sergey are going to come in and pull an Elon and go deep and figure out which 50% or 20% of the company is actually good in doing their jobs? Probably not. But is it possible that they could make a leadership change? Yeah, it's possible. Probable? I don't know. I mean, I've heard that the company is the way it is because they like the way it is.

All Google needs to see is 300, 500 basis points of change. ... And the market cap of this company is going to get cut in half. ... Okay, because there is only one way to go when you have 92% share of a market and that is down. ... But in all of these choices, what I'm telling you on the dispassionate market side is if you see perplexity or anybody else clip off 50 basis points or 100 basis points of share and search, this thing is going straight down by 50%.

Yeah, I think it's just that memo is a tip off that there is a 20,000, 30,000 person rift and a reorganization and they're just going to keep cutting the DI group.

I think it's inevitable that human knowledge labor, where the job of the human is simply the ingestion of data and then communicate an output of data, seems like it will eventually be replaced by computing somehow. And this is happening now in an accelerated way with these LLMs.

Well, it sounds like they're able to eliminate a lot of front line customer support roles by using AI, which is what I would expect. ... And I think the AI will do a really good job eliminating level one. It'll start to eating to level two, but you're probably going to need humans to deal with the more complex cases.

is it going to improve 10% a year or 10% a month? If it's improving 10% a month, we're going to get to 98% of queries done this year. If it's doing 10% a year, okay, we're going to get to 99 or 98% of queries in four years. In other words, this is happening, folks, and it's happening at a blistering pace.

A one-person company should be able to spend less than a few hundred grand to get to product market fit in the next few years.

if they've been given the job DEI and they've been given the job to do trust and safety, they're looking to fill their time and make an impact. You got to, I think, maybe have less of these people. I honestly think it's about cutting people.

I, there is no company where I have majority control, where I have an HR department. ... And what is left over is the very dark part of HR in most organizations, which is the police person, the policeman, right? What is Sacks called? The commissar. That is why everybody hates HR. I have never met a company where that is a successful role over long periods of time.

I think a lot of founders don't understand that DEI is not something they have to do. They don't have to have a DEI organization. This has somehow become a thing. It's not required. I think people are realizing, why would you do that? Why would you create this large bureaucracy in the company that undercuts the meritocracy, that adds a lot of costs, and that slows you down?

And so I'm not a huge fan of race-based metrics or gender-based metrics. I'm generally more oriented around being blind to those variables and focus much more on the variables that can actually influence the outcome of the organization.

I think like for you to make an investment at a $5 billion valuation here, you've really got to believe that the growth continues at this rate and it doesn't revert back to the mean growth rate of the last couple of years of basically 5%, which is roughly flatlined. ... I mean, you could argue it's probably worth, in the best case, in the $2 to $3 billion kind of valuation range

It's kind of like in the mid, you know, kind of two, three, four billion dollar ish range. The big problem is the ARPU, because these are not users that represent sort of Facebook's bread and butter.

But you can see that as a challenge or an opportunity, because if there are ARPU is only 10% of Facebook's, there's a lot of headroom there to grow it.

The issue with this user base is they're incredibly sophisticated internet users who don't click on ads and are kind of anti ads as opposed to, you know, the general population on a Facebook or a generic service.

So if you look at everybody else, for example, here's Teleperformance, which is a French company that runs call centers. ... They lost $1.7 billion of market cap when that tweak went up, about 20% of their market cap. So this is the real practical implication. ... Because at the limit, if every single company is able to implement something that is as economically efficient as what Klarna did, Teleperformance doesn't exist and there's $10 billion and $335,000 employees that will not have a job.

And I just said that we should call this TAC 2.0, except now what Google is doing is, instead of paying for search, they're actually paying for your data and saying, give it to me so that I can train my models and make it better. ... So if you're an entrepreneur building a website or building an app that has really unique training data or really unique data, you'll be able to license and sell that. And that'll be an incremental revenue stream to everything you do in the near future.

And I'm not sure how you get paid a continuous licensing stream for that content. Once you've trained the model, the content gets old, it gets stale at some point in a lot of cases, like news. And then eventually, if you don't have a high quality, continuous stream of content, it's not worth as much anymore. ... And so every year, all the old data becomes worth even less.

Reddit, Quora, Stack Overflow, they're going to just get taken out. I think this is going to be the new model. ... I think they're going to get taken out. I think these businesses will become too valuable because they do have ongoing content that just keeps getting generated.

There's just a huge number of vendors of content. And so, models will need to buy some, but as long as they can get some, they don't need to have all. And therefore, it's basically highly competitive among suppliers, and there's a very limited number of buyers. So that tends to be the buyers.

You took the other side of this. But now that you see the market based solution starting to emerge, what do you think of this market based solution? You think it's got legs?

Well, so what Zuck said on the last Meta call is the reason we open source everything is because we don't directly sell AI. We create products that AI makes better. So by open sourcing this, we allow the community to advance the ball and we get to reincorporate those changes. So it's a very smart strategy for companies that aren't directly selling the AI.

By the way, did you guys see there was a meta demo, which I thought was really cool, which was it was run on Lama 70B, but it was a real-time translation tool, where the person was speaking in Hokkien Chinese and the other person was speaking in English and they were able to understand each other.

The entire Zendesk workflow could be replaced by a handful of these open-source agents, where all of a sudden people can eliminate a lot of op-acks. ... But the point of these AI agents and bots and workflows is that it'll reintroduce the concept of cost savings, of this idea that you can have cheaper, faster and better. And the more that that stuff is open source, my gosh, I think it just makes it very hard for companies that have point products to survive.

By the way, developers, by the way, on Groq, just this past week, in the queue, the wait list tripled. Now there's almost 10,000 developers. ... I think the most important North Star metric for these developer platforms is basically that. As goes the developers, so goes the platform.

I understand. I'm just saying it doesn't make a lot of sense that 2,000 employees that were specialized in building a car, all of a sudden now become the AI team.

The speculation is that they're going all in on AI. They just see that as a much better future.

Well, I think it's more core to what they do. The car never seemed that core.