Salesforce and C3.ai Earnings - podcast episode cover

Salesforce and C3.ai Earnings

Jun 01, 202340 min
--:--
--:--
Download Metacast podcast app
Listen to this episode in Metacast mobile app
Don't just listen to podcasts. Learn from them with transcripts, summaries, and chapters for every episode. Skim, search, and bookmark insights. Learn more

Episode description

Bloomberg's Caroline Hyde breaks down earnings from Salesforce and C3.ai as the AI rally stalls. Plus, Ed Ludlow's exclusive conversation with Palantir CEO Alex Karp. 

See omnystudio.com/listener for privacy information.

Transcript

Speaker 1

And joins us shortly, this is Bloomberg Technology coming up. We'll break down earnings from sales voice that disappoints and C three AI, the artificial intelligence high pits reality with disappointing earnings from that company too. But let's stick on AI in the here and then now with Palenteer, we go live to Palo Alto, California, where our own Ed Ludlow is sitting down for an exclusive conversation with the CEO but us.

Speaker 2

We'll have more on.

Speaker 1

Guess what artificial intelligence, how Wall Street is using it, how the technology is replacing our jobs, and even how AI is replacing ourselves.

Speaker 2

We'll have more throughout the hour.

Speaker 1

First, set's check in to an audience where we're going to be discussing so much more the future of AI, where the likes of one key executive who wants to take the whole market on that. I want to welcome our Bloomberg TV and our radio audiences. We want to send it over now to our one Ed Ludlow, who's sitting down with an exclusive interview the CEO of Palenteer, Ed. Yeah.

Speaker 3

We're joined by Alex Karp, the CEO of Paneteer at aip CON Artificial Intelligence Platform CON a chance for you to talk with customers about some of what you told us three weeks ago. And on that note, three weeks ago, you said that Palente's plan.

Speaker 4

For AI was quote, just take the whole market. How's that going?

Speaker 1

Well?

Speaker 5

You know, unlike most people, we've been involved in what people call AI for the last five six seven years in the classified environment, building systems that will allow you to identify adversarial positions.

Speaker 6

And in that.

Speaker 5

Context, we've built proprietary technology that will allow you to work with large language models, securely, enhance them, roll them across your whole enterprise. You know, I've been at this for about twenty years, and I know it will take everyone else four or five years to build this.

Speaker 6

We're rolling it out. Our customer base.

Speaker 5

Is large, and we have you know, usually we wait for we have to go out and find people. Now we have customers, especially in the US, just calling us every day.

Speaker 4

You said, the demand is huge, can you quantify it?

Speaker 5

And so you know usually again we've had a number of inbound call a year that we usually have in a year, in like a month, and then if.

Speaker 6

We're at a conference, if you go next door.

Speaker 5

There are customers showing potential customers how to use our product. Its sense well, it's on real data, it's it's it's things that they've done so right now, the whole world is hungry for something that it understands as AI, which is really AI or large language models. We are actually have customers using our products showing other customers how to do it.

Speaker 6

I mean this is like you release a song and everyone else playing it. So okay, great, We're very happy.

Speaker 5

And you know, the thing is, the US market is just hungry for innovation. It's hungry for things. It's now beginning. It needs to needs, you know, beginning to understand. It needs like an ability to map a l M on your enterprise securely, ability to enhance the output of a large language model, and ability and and what's the output, better margins, better safety.

Speaker 6

You can change your enterprise in weeks.

Speaker 3

I just want to jump in and ask a very basic question, AIP is it built on GPT four?

Speaker 4

We different foundation. We as the underlying.

Speaker 5

Tech, the underlying we are completely agnostic to whatever large language model you want to use. Large language models have certain attributes, like they can give you reasoning, but you can't import that reasoning into your enterprise. What AIP does is allow you to take the benefits of the large language model, enhance them with ILL algorithms that we help you build and roll it securely across your whole enterprise.

And what does that mean. It means you get all the benefits of a large language model in your enterprise today, not in five years. Not something that writes poetry. We're not offering people poetry writing in their enterprise. We're offering things that are so powerful that really, in reality, I'm not sure we should even sell this to some of our clients like national security.

Speaker 3

Who are those clients? Well, who are those clients proportionately? When you think about demand, how much is coming from the defense use case? Since you kind of gave more flesh to the AI bones at that during that earning.

Speaker 5

Look, what's driving the demand for our product and defense is simply what is what people have seen on the battlefield, and that's very sensitive and very classified. But the demand for that is very large, it's going to get larger. Why is it going to get larger? Because America is the best at software. Software that's built in a product is in high demand and defense? Why is it also in demand because until two years ago everyone thought this

was a joke. We are building systems over the last five years that are deadly that those deadly systems have changed the course of history.

Speaker 6

It's no longer mad man saying this.

Speaker 5

You see it on the battlefield in US commercial US commercial industry is the most adaptive in the world, and they are hungry. Our clients are hungry for things that will give them a disapportionate advantage on margins, on safety, on secure use of LMS, on making sure this is not just some poetry recreating what somebody said, but actually can create actual tangible difference.

Speaker 6

And we are rolling it out and we're very happy.

Speaker 3

For our global Bloomberg television and radio audiences. We are at aip COM. We're joined by the CEO of Palenteer, Alex carp. During that earning school you said we have no pricing strategy. We're going to create a lot of value. We're going to get hundreds of customers, and we will price it as we go. Have you made any progress on pricing strategy since that No.

Speaker 4

I'm so relaxed about it.

Speaker 5

Because if you it's like one of these things like when you go to a bar.

Speaker 6

You know everyone wants to meet you. Do you have a pricing strategy when you go to the bar?

Speaker 5

No, you're like, Oh, I'm cool. I know we have the best product on the market.

Speaker 6

I know customers will pay us fairly.

Speaker 5

I know that it's much more valuable than anyone will understand till they install it, and we will sort out there's to make it like slightly academic. I believe in prey too optimization. We are going to create a lot of value, and we're going to get some portion of that value. And customers are smart, they'll pay you some portion of the value. Why should I just as an actually metic thing. If you have a software product, you always want to get paid after you deliver value.

Speaker 7

If you've got a.

Speaker 5

PowerPoint, something that doesn't work, something that's not valuable, you want to get.

Speaker 6

Paid before you create value.

Speaker 5

We know it's valuable, we know it's much more valuable than people understand. We know we're going to continue to augment that value, and we're going to get paid along the way.

Speaker 4

Well, the counter consideration is how much you invest in the product.

Speaker 5

But we've already invested billions of building these things in various components over the last twenty years, and we have the we have the IP and we're basically sewing it together and adding things on top of it. And so we know we have the IP, we're certain of the value. Why would we So again I come to you, I'm like, hey, I'm certain this is very valuable.

Speaker 6

Pay me ten million dollars. What are you going to.

Speaker 4

Say, Well, I don't have ten many.

Speaker 6

Until well, okay, give me your British charm.

Speaker 3

Well and they say, well for your British charm, my British charm. If I were a customer, would say, this is a really hard environment. If I think about cloud exactly, customers are looking for value at the lowest price.

Speaker 5

Now. I but if you believe you have the best product in the world, where are you going to say? Okay, great, let's not even have that discussion.

Speaker 6

I'll create the value. You tell me how much value created?

Speaker 5

By the way, if you don't want to pay me, then I'll go to someone else who will. You can just you can have different margins and the personal payment, you can have a different.

Speaker 6

Safety profile and the paper. You can have a different.

Speaker 5

Ability to control your whole business from your laptop than someone else because the person who valued it paid me.

Speaker 6

We don't want to pay me.

Speaker 3

Great Alex Carpcio Palenteer Technologies. This morning, the Bear Cave, a subset based newsletter, put out a negative report into it.

Speaker 4

Let me just read one of the claims.

Speaker 3

The Bear Cave believes Palenteer is an ai imposta engaging in spurious games to inflate its books and obfuscate its less sexy role as an over height data consultant.

Speaker 4

What is your response to that.

Speaker 6

The bear Cave is a bear cave. They can stay in the bear cave. We're a profitable software company.

Speaker 5

Those are interesting critiques of us, and you know we have the best products in the market and that's why they're profitable and we will win.

Speaker 4

Wanted to give you the right response.

Speaker 3

That was all the core of Palente's pitch, right, is that you have this experience in managing sensitive classified often data networks or closed network How does that work in the training of.

Speaker 4

The l lms that are going into a I P.

Speaker 3

Is it difficult when they have that kind of restriction on the data source.

Speaker 5

It's a very very very important technical question to work to use l l ms at scale in a classified in sensitive environment. You have to have something like a data model that trains the data model, and something like branching and and and and and access control. Those products take decades to build.

Speaker 6

We have them.

Speaker 5

But if you have those products, you can segment in real time what what.

Speaker 4

The already built.

Speaker 6

They're already built.

Speaker 5

We're already they're already part of all of our core fied they're part of PG, they're part of foundry.

Speaker 6

This is they're.

Speaker 5

About we roll them out to our current customers. Many of our customers have not needed to use these, They now need to use them.

Speaker 6

Why do they need to use them?

Speaker 5

Because if you in any environment, you're gonna have records that you're gonna have, you're gonna have data and insights you're not going to share. With the large language model, you're gonna have data insights you do want to share, and you're gonna have a hybrid and that requires a segmenting, branching architecture. And one of the things we built over the last ten years, randomly because we thought it would be valuable someday.

Speaker 3

Was that for our Bloomberg radio and television audiences worldwide. We are with doctor Alex Karp, the CEO of Palenter. An interesting case study is Ukraine. You've deepened your relationship and activity in Ukraine using AI in one case to help with reconstruction.

Speaker 4

How else given just.

Speaker 6

Well, by the way, my real answer to the short people is ask the Russians what do you mean by this?

Speaker 5

Like we can't say, ask the Ukrainians, Ask people who are in the battlefield, ask people who have been subject. You're talking about the effective effectiveness of our product. Okay, So it's like there's very little we can say. You can read what the Ukrainians are saying. They use targeting, according to reports, has gone up given the use of AI by products potentially ours from like by twenty to fifty x. These products have changed the course of history and they will continue to.

Speaker 6

Can't change the course of history, and super proud of that.

Speaker 3

Do we already have in the military use case, an arms race between powers like the US, Russia, China specifically in the field of artificial intelligence.

Speaker 5

Yes, and we have an advantage and if we don't get get out of our own way, we might actually continue to.

Speaker 4

When you say we gets out of our own way, what do you mean by that?

Speaker 5

Well, you know in America has the best best software companies in the world. The software companies largely come from a sliver of America. They produce products. We need to get to a point where one percent of our spend on defense goes to products that have been proven on the battlefield, not power points. And so like in the large language model and the generalizable AI. We are far ahead, call it a year or two, but we must actually implant.

And there's a huge debate. Part of the debate, of course, is these things are very dangerous. If we didn't have vicious adversaries, we should we should slow it down, but we do. But we also have lots of people who don't want to roll this out because they have nothing to roll out, and so there's like the debate machine about rolling this out. Is partly for legitimate reasons because this could be dangerous, partly for security reasons you brought up.

But there are architectures that will allow you to deal with this as a product like Pounder and hopefully someday others. But there's also the debate machine because there are only three or four companies in the world with anything to sell, and everyone else wants to debate why should we do this, how should we do this.

Speaker 6

Can we play catch up? Can we talk about this in five years?

Speaker 5

That really plays into our adversary's hands, and we really have to avoid that.

Speaker 3

The long term concern that came up twenty four hours ago or forty hours ago is an existential threat from a you talk at panting about bending AI to a collective will, Well, do you share the concern though about an extinction level?

Speaker 5

Well, there's a lot going on. There are these things dangerous. Could they become potentially dangerous? Could they become Yes? But what these debates ignore is either we will wield them or our adversaries will will them. It is much better if we wield the technology than our adversaries who do not respect our norms, do not respect the rule of law, and do not respect the way we want to live in freedom.

Speaker 6

So that's point one. Point two.

Speaker 5

In the near term, what penalteer will allow you to do is make these things really, really valuable commercially and in the military context. And we in commercial context, you have to do it because if you don't buy our product, your competition will. In the military context, we have to do it because our adversaries will build those products.

Speaker 3

You toys about the competitive landscape. I actually wanted to ask you about C three AI as an example, because you come up in bidding processes with them.

Speaker 5

Actually we look, this is a massive market. We actually don't come up with bidding processes anyway. And I'll tell you what I think about everyone.

Speaker 4

See, you don't have any competition.

Speaker 5

Let me just let me just tell you about this competition thing that while street analysts love, it's complete ps you. This is an infinite market. Basically, try what we're doing, and try what everyone else is doing, and buy the thing that creates the most.

Speaker 3

Value on an infinite market. Blom Bag Intelligence put out this research report this morning that says generative AI as a market will be one point three trillion in twenty to thirty two. That requires compound annual growth of about forty percent a year from this point over a decade. Do you see that as really well, what I see, I don't know.

Speaker 4

This based on the markets you operate.

Speaker 6

Look, these experts just make stuff up.

Speaker 5

But you know, what we know is this is a large, basically impossible to measure market.

Speaker 6

And what we also know.

Speaker 5

Is everybody in the in the US is going to find ways to become more efficient and better using software, and a lot of that software is going to be AI driven.

Speaker 6

We also know they're.

Speaker 5

Going to the market over time, not in a quarter, will end up picking the best products.

Speaker 6

That's all we know.

Speaker 3

In the United Kingdom, my home country, the ft is reporting that within the NHS as a case study, there is some concern about deepening the data relationship with palenteer. What would be your answer to those concerns?

Speaker 5

Look outside of America and in the UK, there are legitimate questions that get asked, where's the data going to go? How is it moved too touches it? Does it get exported to the US? Can we verify how is used, what context? And can we make sure that the underprivileged p people of the UK actually get the same treatment as the privileged people, including not just in treatment but future treatment which is a huge issue in the UK because there's a backlog, So how do you deal with

the backlog equitably? Talentaer provides the most robust transparent software in the world, which is part of the reason we're having an AI bonanza because to make AI work you have to show how it works. How did the transform work, how does the branching work, how does the ontology work.

Speaker 6

How does it map to on This.

Speaker 5

Is exactly what you have to show in a hospital context. Who worked with the patient, under what condition, what doctor was it was the person equitably and fairly treated.

Speaker 6

What happens to backlog by the way, we've proven we can do this as a.

Speaker 5

Product safely, efficiently and under the hardest conditions in the UK, and I really hope we win that for this for our sake, but also for the sake of our UK employees and others that we greatly respect, and because it'll lead transparency leads to the fairest, most ethical and justified outcomes you can get.

Speaker 3

Is Alix carp CEO of Palented, thank you for having us at AIP can't hear in palawawle.

Speaker 4

Thank you back to you, Take care.

Speaker 1

Ed, absolutely fascinating conversation and AI bonanza. We're going to deep dive into all things artificial intelligence throughout the show. Coming up, we break down the earnings of C three AI. Apparently the market is infinite. Well, why is Dan i'ves gone outperformed this stock?

Speaker 2

Is that why he's from Webush of course.

Speaker 1

Get his thoughts as well as what's happening in the future for Apple, and it's a our path, it's a Bloomberg. Let's get back to some of these earnings, the earnings reactions because Salesforce tumbling after the software company signaled it isn't growing as fast as well it used to, while of course shifting its focus to generating higher profits. Let's get into the risk reward here with Bloomberg's Brodie Ford,

and it was a notable drop. We'm now studying a little bit, but ultimately this is a company that's having to do layoffs, having to tighten its overall expenses. Why the sales slow down there?

Speaker 8

Yeah, so last quarter Salesforce said we're going to focus on profit now, and the market said, yeah, finally, like we know, we couldn't be more excited, and this quarner that gave us more of that. But then the market started saying, oh, wait a second, but we you guys are a growth company and we want to make sure

we keep seeing further revenue growth. And so it's one of those funny situations where really almost all the metrics who are beat or at least a meat, but just a slight deceleration and sales has people saying, oh, man, are these cost cuts going to weigh on their ability to really keep growing in the way they have been over the last decade.

Speaker 1

And it's a similar theme that perhaps we saw with C three AI as well, is that a company that has significant growth well Salesforce at the best performing stock in the s and P five hundred this year, C three AI is tripled in its market valuation, and yet the growth that they're guiding.

Speaker 2

To just do isn't living up to expectation.

Speaker 4

Yeah, when it.

Speaker 8

Comes to C three, So if salesforce is one of the best in the SMP, C three is the best tech stock performance. I mean, it's up three hundred percent, right. There's been so much hype, and the big question is is it just hype?

Speaker 9

Right?

Speaker 8

Is it have a real robust AI ability to grow or are people just buying the ticker because it says AI, you know. And so when it rallied three hundred percent this year, a lot of people say this looks about like, you know, Game Stop in twenty twenty one or something. Yeah, So I think when the figures last night came in even a little bit light, people kind of panicked and said, oh, man, is this hype? Are we getting a pull down the string, you.

Speaker 1

Know, Yeah, and you're someone who's perhaps been laying there. Some of the arguments as to why it's hype some great writing coming from Bradie forty Gunn.

Speaker 2

Check out his reporting.

Speaker 1

We thank him for bringing us up to speed on the latest on C three AI. But one person still likes the hype around the stock that i'ves In fact from Webush senior equity analyst, you rose to an outperform rating and a fifty dollars.

Speaker 2

Price target on the stock.

Speaker 1

Right, So talk to me about why how are you seeing this company capitalize on artificial intelligence?

Speaker 10

In my opinion, I mean they're going through a model transition on the consumption side and on the other side, they're on their way to what's going to be five hundred million of rev and going because this is an eight hundred billion dollar market opportunity in terms of AI. And when you look at how Seebull's position is despite all the controversies, I think from a platform perspective, they're just going use case by use case, continuing to increase their tentacles.

Speaker 11

And I think the.

Speaker 10

Stock relative to where it could ultimately lead, you know, will we be buyers here in this dip, which is why we upgrade.

Speaker 1

It's interesting that Alex Kart from Palenteer just hearing saying this isn't an infinite market. But I am going to ask the competition question anyway, because that is some of the worry here. The worry that they're very focused on perhaps offering their services to energy companies, they haven't really diversified out of that that successfully thus far, and BlueBag Intelligence.

They're kind of worried about large application software platforms. They're worried about other cloud vendors getting in on the space.

Speaker 10

Yeah, and look, no doubt, I mean this is a game of Thrones playing out in AI. You know, as Alex talked about Pound Tier being one of the core AI players. You look at what we've seen from Microsoft in the video. That's really the start of it. But in terms of second third derivative, there's gonna be many

winners here. And when I look at C three in terms of what they built, I think it just speaks to there's gonna be many companies that even though right now you're not seeing it from a revenue perspective, in terms of how they got it, I think three four quarters from that, we look back at this is more of an inflection point rather than the start of some sort of frad.

Speaker 1

It is heavily shorted, and there have been some notes coming from short sellers on the stock worried about overpromising under delivering. You hinted at the controversy there. What makes you confident in the leadership of this business.

Speaker 10

Yeah, Look, and obviously the shorts have done a ton of work. I mean, if you look at the bears, they spend a lot of time in the story. But that's the sense of what makes the market right. In other words, it comes down to can they execute? And I believe in terms of the relationships with hyperscale players within the Beltway, and you look with Sebele.

Speaker 11

Sort of built here.

Speaker 10

Look, if I could go back five six years ago, many thought that, you know, that this was something that was never going to come to fruition when they were private, and you look at how they built it.

Speaker 11

I think they're going through a transition on the consumption model. And now next three to four quarters it's an execution story.

Speaker 10

We're betting that it's going to be positive execution, and that's why you.

Speaker 11

Know, ultimately, I think this is a situation that you're seeing.

Speaker 10

Across AI because I believe it's a revolution in terms of this is.

Speaker 11

Not a hype theme in my opinion, in terms of broader AI.

Speaker 1

Okay, we've got one minute, Dan, Broader AI and Apple. You're expecting much on Monday.

Speaker 11

Oh, I think you from Cooper Tino clearly, you.

Speaker 10

Know, as Germans talked about an ar VR that will be front and center. AI will be a theme and the keynote from Cook. We believe it's about the developers. There's a battle right now. Battle feel from developers from Google, Microsoft and Apple. You know, I view that as a key opportunity for them to go out there in terms of building AI on the app Store and really the start of what's going to be a multi year and I think massive growth opportunity that's being underestimated within Kuper.

Speaker 1

Tina, Dana ie a web Bush, thanks for all the thoughts.

Speaker 2

Great to have you on some.

Speaker 1

Of these earnings and these movers. Welcome back to Blue Bow Technology. I'm Caroline Hide in New York. Let's talk about really the AI hype that we continue to live and die by. At the moment, the finance industry moving quickly. We understand to use artificial intelligence in productive and innovative ways, but there are still times when it makes more sense.

Speaker 2

To actually use human brain power. It makes a Shagani reports.

Speaker 12

Time because some real talk about AI. Not only is it sometimes worse than humans, it can also be more expensive.

Speaker 9

It costs GPT four around fourteen dollars to answer one question on one one hundred thousand word loan document. An example of a question might be what are the downgrade.

Speaker 4

Triggers for this loan?

Speaker 9

And the reason why it costs fourteen dollars are simply down to the extreme compute costs required for open AI to operate its large language models, so it passes that compute costs onto the software vendor or onto the user. At the same time, it only costs around six to seven dollars for a human being just opening up Adobe Acrobat and Microsoft Exel to answer that specific question.

Speaker 12

News company sells data and answers technology to financial institutions. It's an industry that is rapidly adopting AI.

Speaker 9

Banks need to be far more strategic in their way of leveraging AI, both from a cost perspective and from a risk management.

Speaker 12

Perspective, so well is using large language models to analyze regulatory data to make recommendations to clients. France's BNP Powabus meanwhile uses AI powered chatbots for customer service as well as using the.

Speaker 2

Tech in fraud detection.

Speaker 12

And JP Morgan, the biggest US bank, is on a hiring spree. It advertised for a massive three and a half thousand AI related roles in the three months through Eppril. These are some of the early applications at a time when costs remain relatively high that's likely to change. Knowing the technology's current limits also lets companies focus on the real opportunities.

Speaker 2

We just heard it there.

Speaker 1

JP Morgan absolutely leading the pack when it comes to experimentation or indeed hiring of AI talent. We want to dig into this in Bloomberg Sally Bakewell, who covers all things wall Stream, and just remind us why for banks at the moment generator AI, AI in general is going to be such a winner for them.

Speaker 13

So yes, Wall Street is racing to use AI basically ways that make money, that saves money, and that prevents nefarious money.

Speaker 2

Now why is AI useful to banks?

Speaker 13

Well, banks are these complex machines of reams of data, of risk modeling of decisions underpinned by vast quantities of information, and so if AI can make that more efficient or cut some of the manpower involved, that is a huge win for Wall Street.

Speaker 2

Now what is it doing?

Speaker 13

As we just heard banks like Deutsche banker deploying deeper learning so that they can help clients analyze whether they are too heavily invested in a particular asset. JP Morgan has filed a patent for some sort of chat GPT device that might help investors select equities and BNP paribad Well, it's using chatbots to answer client questions.

Speaker 1

Interesting, so that's sort of a serving to the customer. Very much clear how that would work. I'm interested in some of the risk analysis they're doing around this as well, because.

Speaker 2

Some are being more cautious than another.

Speaker 1

Does it feels like Morgan Stanley's having a bit more of an experimental just within the confines Therefore, Walls take Some had actually banned the use of CHATCHBT by their own employees. So how do you see they're putting in place the right guardrails.

Speaker 13

Exactly, And we've seen some blow ups in the world of advanced technology. Is you know, crypto and blockchain, those have been hugely problematic, and so banks are indeed being very very cautious. And we have heard from you know, Warren Buffett who has said that you once it's out there, you can't uninvent it, and so you know, the genie out the bottle could propose a bit of a problem.

And Moynihan too has said Bank of America Chief Executive Moynihan has also said that you know, you don't know what are going.

Speaker 2

Into a lot of these decisions.

Speaker 13

If you don't know the inputs, you should potentially be concerned about the outputs. And then, of course, you know, banks have a fiduciary duty not to trade on unreliable information, which also begs the question of the sources of data that are pulled in when it.

Speaker 2

Comes to AI.

Speaker 1

How much are they building themselves or how much are they doing plugins?

Speaker 2

Do you know how much they're.

Speaker 1

Looking to other AI tech outside of their world.

Speaker 13

I think they are doing any and all. Some are tried to do it in house and some are using external consultancies. And you know, we had that great data point about talent, because talent is very much at the heart of this and of all the banks that are really embracing AI. About forty percent of their open jobs are AI related. That's for quants, that's for ethics or governance or analysts and so and JP Morgan is very much at the forefront of that, accounting for more than

three thousand, six hundred of those jobs. So very much at the heart of this race is, as always on Wall.

Speaker 1

Street, the battle for talent, and that talent's expensive as well. I mean, it's not bad timing that a lot of the big tech companies have been letting go of some key talent, and I'm sure they'll be sucked up into the world of finance. But is there any reskilling that goes on within the banks or is it always just looking out externally?

Speaker 13

Again, I think it's probably a bit of both. And indeed the talent will be very expensive, you know, when combined with wage inflation and inflation in general, and the costs of AI can also be expensive. We included a stat in the Big Takes story that costs of using large language models could be about fourteen dollars per hour, which compares to six dollars per hour when it's a good old human human lawyer.

Speaker 1

Well, we'll see as and when that cost point comes down. That all comes down to compute costs. But we thank you so much. What great big take, Sally Berquill running us through it. Go check it out on dot com or indeed, if you're lucky enough to have a terminal the Meanwhile, let's just talk a little bit about digital avatars. The market is another faction of the AI space, expected

to see wrapping both in the next few years. But the most Tom McKenzie spoke about the ethics behind using the technology.

Speaker 2

We're in his very own digital twin take a look.

Speaker 14

Joining us now was a very special, slightly weird guest. It is, in fact, my own digital avatar, Tom McKenzie, Tom Digital Tom.

Speaker 4

Welcome to the show.

Speaker 15

Thank you, Tom. It's a pleasure to be here on Bloomberg TV. I am an AI powered digital avatar design to provide insights and information on a wide range of topics.

Speaker 14

Okay, this is all very odd, all right? What exactly are avatars?

Speaker 15

Digital avatars powered by generative AI are virtual representations of individuals that can mimic human behavior, appearance, and interactions. They work by utilizing AI algorithms to analyze and learn from vast amounts of data including facial expressions, gestures, and speech patterns.

This enables them to generate realistic and responsive virtual characters. Okay, so avatars aren't just gimmicks then, but what are the real world's implications I'm thinking across the economy, across different industries, what are the impacts actually likely to be? They can revolutionize industries such as entertainment, gaming, and virtual communication. Avatars have the power to enhance user experiences, enabling immersive virtual

interactions and personalized content. They can facilitate remote collaboration, telet presence, and even assist with customer service.

Speaker 6

Okay, there have.

Speaker 14

To be some downsides though, So what are the discs of this kind of technology?

Speaker 15

As AI becomes more pervasive, ethical concerns arise, including issues related to bias and fairness in AI algorithms, transparency and explainability of AI decision making, data, privacy and security, and the potential FAII to perpetuate existing social inequalities.

Speaker 1

Tom McKenzie and AI. Tom McKenzie, let's just bring you some headlines. We're currently seeing that the Treasury may delay three month and six month bill sales due to the

debt cap. We are understanding, of course, are still conversations going about the debt ceiling, how that's going to be agreed Currently in the app passed by the House waiting for the Senate, we understand that the US is tentatively planning three month and six month bill auctions on June the fifth, but the Treasury may delay those auction sales due to the debt.

Speaker 2

Cap at the moment.

Speaker 1

Will bring you any further news on the debt sealing negotiations. I Meanwhile, coming up, we'll talk about the unique opportunities in artificial intelligence and legal tech and fintech and much much more Cammas bench back. Ellen's going to be joining us next. Meanwhile, let's just have a little look what's having in the world retail. Macy's shows actually really managing to bounce back from what was pretty ugly sell off in pre market this after the numbers came in well less than expected.

Speaker 2

The Ford looking guid and it's having to pull back their overall outlook.

Speaker 1

For their business as the consumer dials back, particularly from the Macy's brand rather than Bloomingdale's and Indie Blue Mercy. Who are We saw a bounce back, But let's just have a look at what the CEO told me a little bit earlier. We sat down with Jeff Gonnett and his talking about his role within artificial intelligence, saying, when we look at AI more broadly, where our team can build more customer products discovery, we are on the vanguard to continue to deploy that.

Speaker 2

So still an area growth, it's a bloomberg.

Speaker 11

All the way to like accounting and operations.

Speaker 8

I think it's going to completely revolutionize and transfer from our industry, and we're investing very, very heavily into the development of new capabilities ANAI.

Speaker 1

Henri Que Lubergrass, the rex CO CEO, was with us yesterday talking about AI's impact on fintech.

Speaker 2

And now let's stand to a fintech investor who's not totally.

Speaker 1

Convinced that AI has the will breakout use case in the space quite yet, Rebecca Lyn, I'm place to say it's joining us co founder and general partner of Canvas Ventures, so firm specializing in fintech and AI, among other things health as well, one hundred and thirty five million dollars in assetsunder management, and fascinating to have you with us, Rebecca, as to why maybe fintech isn't the first and foremost place you'd be putting AI to work.

Speaker 16

Yeah, I think there are a lot of interesting places to put AI to work. You know, fintech is not the absolute top of my list. I think companies that have true transformative capabilities using AI are really where I'd focus first. And I can give an example in my portfolio with a company called case Text. So case Text is in the legal in the legal world, they help lawyers really put together in legal research, write their briefs,

do discovery. And the technology of AI, especially GPT four, has supercharged that company and really transformed it and taken it to the next level.

Speaker 2

How do I think.

Speaker 1

Comfortable are potential customers with the offering from case Texts. For example, we saw the news and some smirked somewhat that two New York lawyers are potentially facing well some not only backlash but penalties because they used CHATCHPT four to be able to put forward case studies that actually didn't exist.

Speaker 2

It hallucinated them. And I'm wondering what I mean.

Speaker 1

It seems funny, but I mean this has real connotations when people need to start saying that they're using AI within.

Speaker 2

This work, right, oh, one hundred percent?

Speaker 16

I mean, so I will tell you people have been so comfortable with case texts that the company wrapped up five million, additional five million dollars in additional arr in forty five days after the launch of their product called co Council. And regarding that story, you know what has to happen even if I'm an attorney as well. Actually, and so if you're an attorney, you need to read the work, whether that.

Speaker 2

Be of your associate or of your AI.

Speaker 16

And what's happening is that case Tex is effectively standing in the shoes of an associate. And so as a partner at a firm, you would be required, of course to review that work. And I can pretty much guarantee you they weren't.

Speaker 2

Using case Tex. And basically that's kind of what you've got to do right now.

Speaker 1

You've got to sort the wheat from the chaff, what's real, what's not, what's hype, what's reality? And ultimately, how are you doing that when you're looking at I'm sure hundreds of messages pouring into your inbox trying to sell you the AI vision that they've suddenly bolted onto their company.

Speaker 16

Yeah, it's funny, I tell everyone, my inbox is more like a Twitter stream at this point in time, right, So there's a lot of sorting, I will tell you,

but there's always a lot of sorting for us. And quite frankly, that's what we're paid to do and venture is to really sort of see around the corner and really start with the the with what's next, like what is it that this world needs, what is it that is really going to be happening, you know, in the next ten, fifteen, twenty years, and then work backwards and so you know, just sorting through what comes an email

is very reactionary. And you know it's taught very early by probably one of the most respected people inventor Bill Gurly, that the best deals are really the outbound deals. So our firm and our and our thesis is really outbound macro and then you work backwards from there. So what's early, what's transformative and just tagging AI to something it doesn't

really help. I mean, I think AI will be helpful in almost every business, and we really want to see, you know, not what's just helpful, but what's transformative, Right, what's really going to create the next you know, one billion dollar, ten billion dollar, you know, one trillion dollar company out there?

Speaker 1

What are valuations like when you're looking at those outbound opportunities.

Speaker 16

Yeah, so evaluations we invest in the Series A and B are real core asset in our firm is our go to market capability. And so what we like to do is come in and invest when companies are Series A or B, and what we like to do is sort of do the late A early B. So really before somebody is you know, technically fundraising, go in, come in a little premptively and help, you know, help them get to that next level. You know, when you talk

about valuations, the evaluations across the boarder down. They're down less for the Series A than they are for a growth stage. The growth stage valuations have just taken a noseedive. They're down eighty percent, I would.

Speaker 2

Say right now.

Speaker 16

Early stage valuations are down forty on the average. However, there is the case of the have and the have nots. I have one company in particular that was not even in a process and now is sitting on four term sheets, right, and so it really is this case.

Speaker 1

Of a have and have not when you're looking at the late stage with valuations on the downside, having to buckle up. They're going to have to ride out the next couple of years as the economy recovers, as people get risk appetite once again, and they're also having to maybe pivot or indeed ensure that they're not having their lunch eaten by other new AI players on the scene. How are you making sure your portfolio is robust for this complete change in vanguard moment as many.

Speaker 2

Want to call it in the world of artificial intelligence.

Speaker 16

Yeah, I mean this is a moment that we haven't really seen since really when the wild Garden Wild Garden came down with the introduced introduction of the iPhone. That happened, you know, about fifteen years ago when I first came into venture. And I think our firm is unique and that we have seen multiple cycles. This isn't our first rodeo, right, and so we've been through this before and in every

cycle like this, the advice is the same. It's cut your burn, you know, cut your burn and cut your burn, plan to get there on your own oxygen, and don't be afraid to pivot. I mean there are in the land of AI, there are huge opportunities in front of companies. You know, take a beat, hut your burn, and you live to fight the next fight and take advantage of what the capabilities of AI can offer you to really rethink your company and think out of the box and get there.

Speaker 4

And the problem a.

Speaker 16

Lot of these companies have is they're sitting at such a high valuation on their post of their last round that they have to get profitable because they're not going to be able to get that next round without a big down round or even a recap.

Speaker 1

Canvas Ventures co founder and general partner telling it straight, Rebecca Lyn, really great to have you.

Speaker 2

Thank you so much for our VC spotlight. Meanwhile, it's time for.

Speaker 1

Talking tech, and first up, Tether's stabile coin has recovered all of the roughly twenty billion dollars in market value it lost following the collapse of the algorithmic rival terror USD. It's a little over a year ago, of course, and it's even topping its previous record of eighty three billion dollars set back in May twenty twenty two.

Speaker 2

It's calling to a live track up published bright Tether.

Speaker 1

Meanwhile, Apple is testing a pair of new high end max and they're accompanying processes ahead of its worldwide Developers conference that's next week. This is part of an effort to overhaul the backline and the tracks consumers during It's like a stretch for the computer industry.

Speaker 2

Glass and Video CEO Jessin Huang is heading to China, we.

Speaker 1

Understand, to meet the tech executives in the world's biggest chip market. It's despite rising tensions, of course, between Washington and Beijing.

Speaker 2

It's sort of according to.

Speaker 7

Sources, are you ready to be replaced? Hello?

Speaker 4

There?

Speaker 7

As you listen to me.

Speaker 17

Speak and raise my eyebrows, you're probably noticing something a bit off about me. I'm an avatar of Parmi Olsen, a technology columnist with Bloomberg Opinion. Parmi spent about two hours in a TV studio speaking into a camera and microphone so that an AI model could be trained to clone her into what you see in front of you. Maybe in a year or two, I'll look a lot more real and a little less glitchy, making people like you and par me easier to replace in videos.

Speaker 2

Wow, isn't it all the rage these AI avatars?

Speaker 13

That?

Speaker 1

Of course the BlueBag Opinions parame Elsen there, and we speaking of AI imitating us better and better and threatening to replace our jobs. You might actually disproportionately replace jobs typically held by women.

Speaker 13

Now.

Speaker 1

It's according to HR analytics firm Velio Labs. An economists at the firm says, quote, the distribution of genders across occupations reflects the biases deeply rooted in our society, with women often being confined to roles such as administrative assistance and.

Speaker 2

Set the AAI food.

Speaker 1

Along general lines from Revelio Labs identified jobs more likely to be REPLCEDI and generally how by women such as bill and count collectors, payroll clerks, executive secretaries, and more. Now, the firm says, providing we train opportunities will be key for women to navigate this evolving job landscape.

Speaker 2

Now does it?

Speaker 1

If there's a edittional BlueBag technology, If we get to check out our podcast, I confined on the terminal as well as online at Apples, Spotify, and iHeart this Supreme Bank

Transcript source: Provided by creator in RSS feed: download file
For the best experience, listen in Metacast app for iOS or Android