That's really what's so exciting about this launch for a lot of people who have been disillusioned with the updates at Twitter as of late. I mean, they have tried platforms like Masdon, like Blue Sky, like Post, you name it. There's so many Twitter competitors out there that haven't quite gotten the network that they need to truly be a
threat to Twitter. And here is Meta, which has upwards of three billion users on its products around the world, and Instagram, which has a following structure similar to Twitter, where where unlike Facebook, where you're friends with the people who are friends with you, on Instagram you can follow people who are interesting. That is the same as a
Twitter model. So it makes sense for them to say we're having a Twitter copycat app over here, and you can choose to follow the same people you follow on Instagram, connect with those same people, and keep your handle, keep your verification. That it's a very powerful head start for this app Threads if people choose to use it, and I do think it could be a major threat to Twitter.
And that's ask you Asha Water.
I mean, we know Twitter is had to sometimes elicit Eve's point, but since Linda Yakarinos come in and take over the CEOO role. They have been much more communicative. Are you finding that they're worried? People you speak to do they see it as a big threat?
I mean, like Sarah said, I think it is a huge threat.
Right.
You have the alternatives, as she mentioned posts massed on Blue Sky that just did not take off in the way that people thought they would. People went to those platforms, but then we're not exactly enthused by maybe the user experience or not having enough people not being able to take over the same handle, and so Integram can come in with threads and do all of those things and sort of solve a lot of those problems.
I could see people moving over.
Twitter has had a lot of changes with the verification system and how you get blue check marks, having to pay for additional features over the past couple of days.
Yeah, Asia, just dig into the limited usage because there's been more clarification over the holiday day yesterday and over the weekends. But what are you hearing about some of the pushback against the latest moves that Twitter itself says is to ensure authenticity.
Yeah, I mean, people are not happy about the fact that you're gonna be limited in the amount of tweets that you can see. And it is supposed to be a temporary measure, but the reason people go to Twitter is to be able to see tweets and to be able to see activity going on, and so when you start limiting that, that's concerning. Some people have also said maybe advertisers could be concerned about that, right if if you there's less tweets, that means there's less ads that people are seeing.
So it it's it's challenging. Now.
It is temporary. We don't know how long it's gonna last. But there has been some pushback just on Twitter and just amongst users.
Sarah talking of pushback, there's who is pushing back in the EU. We understand that threads isn't gonna be available there as soon as tomorrow. If you're in the EU, is this regulators saying no, you can't do this? Is a this meta saying we are not bringing our innovation to you because of.
Your stricter regulations.
I think that that there's a little bit of both going on. I think that you know, regulators are are certainly concerned and saying, you know, if you're going to add this new product in the EU. There are other considerations around competition, around the fact that Meta already has you know, all of the major social platforms. Do they really need another one? There are the concerns about the
usual privacy details. I think there's just a little bit more to work out before they can launch their I think I think that, you know, given Meta's global focus, they're going to want to launch there eventually, but for now, they're going to want to establish this as a product
that people want to use. And I think it's a really interesting time because on the across their other products, they're really moving away from tech spased communication in towards short form video, and they're trying to take on TikTok on Instagram and and Facebook right now, and so threads may be an option not just for people who use Twitter, but for people who want to just post about their lives in a tech based way and find that those kinds of content don't do as well on Instagram and
Facebook the way that they used to. So it could be an interesting, more mainstream version of Twitter.
What about the advertisers, Aisha, How much are companies looking to go towards a Meta visa via Twitter?
Well, I think advertising is a huge opportunity for Meta with this right, we advertisers moved off of Twitter. Advertising revenue went down significantly since Must took over, he said. Some advertisers have come back. A lot of people left for all the reasons that we just outlined right, some of Must behavior, some of the changes to the platform. Already,
advertisers are going to some Twitter alternatives. So it's a huge risk I think for Twitter if Meta can build a competitor and take away some of the issues that Twitter had with content moderation and so on.
And so forth.
However, Twitter does have Linda Yakarino, and she has a lot of industry experience, she has relationships. Twitter has made some changes recently. They've been talking more about brand safety, so they might be able to keep advertisers.
I don't know.
We'll see how it plays out.
We'll have to wait with that cage match occurs as well, But for now, the figurative cage match is about to ensue as soon as tomorrow. Sarah fryar ash accounts, thank you both so much for breaking that down.
Meanwhile, coming up, Rivian.
Started delivering the electric bands it makes for Amazon to Europe and there's first commercial shipments outside of the US.
More on that next. This is a Bloomberg time now for talking tech.
First up, Meta CEO Mark Zuckerberg is facing strudiny in China.
Social state backed.
Social media account published is scathing editorial on the Facebook co founder. Now first reported by Hong Kong newspaper that the editorial emerged after The Wall Street Journal reported that Meta was in talks with ten Cent Holdings about selling its quest for the Yar headset in China. Ssuckerberg has openly criticized Chinese censorship. Meta and ten Cent didn't respond
to requests for comment. Meanwhile, Spotify's CEO has raised over sixty five million dollars in venture capital to expand his medical startup. It's called Nico Health, and Tomaco Lakestar General Catalyst joining the investment and some of the company's biggest backers. The startup provides full body scans for skin and heart conditions. Over there in Sweden plus Rivian and it started delivering the electric vans it makes the Amazon to Europe's first
commercial shipments outside of the United States. Amazon will roll out more than three hundred delivery vans in the coming weeks to German cities including Munich, Berlin and Dusseldulff joining us to just talk through that a little bit more as Spoomberg News is Craig Trudell.
So how big a deal is this for Revian because the shareholder is.
Like it, Yeah, it's absolutely a big deal for Rivian. I think anytime can be a sort of reminder of the ties to Amazon and the role that Amazon plays as a big customer is a good thing for Rivian.
Not that folks who know this company well and have been following it closely will forget that, but I do think this is a company that, you know, if we take a sort of step back and think about, you know, what's happened since they're Blockbuster IPO in late twenty twenty one, they've had a heck of a time getting production up and rolling, and so this is an indication that, you know, yet another indication that you know, production is is kind
of being smoothed out. We saw that a little bit earlier this week with the production numbers that they reported for the second quarter and the fact that they're able to now you know, not only you know, boost deliveries locally, but start some exporting is a positive sign for production going forward.
I'm positive coming from the analyst as well, need amnalyst Chris Piers lifting the price target after those production numbers DA Davidson the analyst, they're really upgoading the Stockton neutral from an underperform and there in line is the fact that it has been underperforming, right, has been a lot of worries and concerns just remind us of what sort of being the choke hold on the production side of things.
I mean, they were suffering from, you know, the issues that the broader industry had has for a couple of years now, right of just supply chain duress and the chip shortage, and just when we think that we're through that, now we have you know, these concerns about these metals in China and whether or not you know, that will be the latest, you know, thing to slow down supply
of semiconductors. I think one of the things that this story speaks to, though, this idea of starting to ship vans to Europe, is just the need for electrification in this commercial vehicle space and this role that Rivian could play in that. When you think about sort of the biggest bang for your buck in electrifying things that are on the road, if you can electrify these vans that do a ton of miles a day that are doing increasingly these last mile deliveries because of the rise of
online shopping and deliveries and city centers. We know for a fact that you know, cities want more of these on the roads. Incumbent manufacturers are having a hard time meeting demand. Startups are having a hard time meeting demand. So if Rivian can get cranking on these vans, we do know that there's a lot of interest in these if they can, you know, sort of follow through on this big order from Amazon for over the next half decade or.
So, one hundred thousand big order for the vans, a lot of evans. There's a lot of vans most Craig Trudell, thank you so much for joining us. Welcome back to Bloomberg Technology. I'm Caroline Hide in New York, Ed Ludlow. He's currently on assignment. Let's check in on these markets today there, because we're pretty muted after the day of celebrations from July fourth, We're just off about a ten percent on the NASDAC and Ashley Tech is outperforming.
The rest of the market.
The rest of the market nervous about Chinese economy, the growth story lackluster there when you look at the services number also. Then then as that one hundred does managed to perform in terms of the big names, we'll get into the individual movers is to support there. But look at the tenure yield once again moving higher by six basis points. Remember we get those FOMC minutes from the previous rate decision where they stood pat They did nothing for the
first time in ten rate hikes. So we're interesting to see if there's any steer, any color as to just how divided the FOMC is in terms of tackling inflation here in the United States. Let's move on and have a look at some of the individual movers, because we do have companies on the move metal platforms. As I say, really the Star performer today up more than three percent as we anticipate that unveiling of threads as soon as tomorrow.
Of course, their Twitter take on Intel off by two percent, and in fact most of the chip stocks are down today as we worry about the geopolitical risk coming from China, the fact that they are limiting the materials the rarer in particular that go into semiconductors' evs, you name it. Those two key materials that are going to be even harder to get a hold of. Intel on the downside, and Pegger Systems just want to show off this particular
software company and interestingly up fourteen percent. Of course, a small company that really getting a rating upgrade is some analysts out there and really feel that some of the legal headwinds are not something to be as agitated about as was previous at fourteen percent.
Meanwhile, let's talk a little bit about.
Software and hiring in particular here in New York City taking steps in fact, the city is to regulate artificial intelligence when it comes to hiring. Today, a new law comes into effect that requires companies using AI software in recruiting to notify candidates that an automated system is being used.
It's also requiring.
Companies to have independent auditors check their technology annually for bias. Let's talk about all of that with George Wycount of Bloomberg Law and George, why is there potentially the threat of bias within these AI systems?
So yeah, AI is used in all kinds of hiring mechanisms around the country, ranging from resume reviews, for evaluating video interviews.
It's all over the place.
A twenty twenty two survey from the Society of Human Resource Management show that nearly one in four organizations use AI in some HR capacity.
So it's really all over there's a lot.
Of concern and with that, there's a lot of concerns about lawmakers and advocates. A like, these technologies, which are relatively novel, they haven't been in effect for that long, rely on data that bakes and prejudice. Because a lot of these systems, like a lot of other things, are being trained on historical data. This data is imperfect and could have a discriminatory event.
Okay, so talk us through the new requirements. First of their coming from New York, what would a company have to do?
How would they go about disclosing?
Yeah, so the law on its face requires all New York City employees to independently audit their systems. That means they have to get an independent firm to audit, it has to be done once a year, and they need to publicly disclose the results on their employment website. And the law also requires that these employers inform job candidates that they are using AI prior to their use in
any kind of hiring mechanism. The law does not currently have any private right of action, meaning an individual you cannot file an individual suit.
Against organization if they don't comply.
But if an employer is to be found in violation of the law, they will encoura a five hundred dollars fine on the first and a fifteen hundred dollars fine on each subsequent offense.
And some companies had actually fought back against it, or at least lobbied against this new law, and it had been tweaked around the edges, George, But I'm interested about now.
It's officially enforceable.
Does this become the regerring across the rest of the states?
What do we see in other cities? What do you think comes next?
Yeah, so this is the first of the nation law and it's likely to be followed. There's been growing pressure in recent years for jurisdictions to regulate AI. It's been a high priority for the Equal for the Equal Employment Opportunity Commission, which is the agency that's in charge of enforcing Title seven of the Civil It's Act of nineteen sixty four.
Excuse me.
Some other states have laws on the books Illinois and Maryland in particular, they have laws and the books that regulate AI in some extent for video interviews, but this is likely to be followed in while right now this just applies to employers within New York City. Is the thing all employers need to be thinking about right now as they use artificial intelligence in the hiring capacity.
Great round up, George Wykamp, Thank you so much of Berg Law of a feeling we might be coming back to you at some stage soon. Meanwhile, let's take a deep dive into a so called artificial intelligence whisperers. Look, you can earn six figure salaries and you don't even need any programming experience. Bloomberg's Conrad Quelty Harper one of these prompt engineers, as are known to find just how to coax the best out of those large language models.
Albert Phelps is a full time prompt engineer, or, as he describes it, an AI whisperer. I sat down with him to understand what's behind this increasingly coveted role that can pay up to six figures.
So we're saying, you know, you're a highly creative and talented journalist working for Bloomberg News.
Thank you and you create.
An introduction to an article on prompt engineering in the House of Bloomberg style.
It's not got the kind of clean pros of Bloomberg, but it's a bit fowery, I'd say. With the rise of large language models like GPT three and GPT four, demand for a new kind of AI job is also growing. Prompt engineers spend their day coaxing the AI to produce better results and help companies train their workforce to harness the tools. The role can pay as much as three hundred and thirty five thousand dollars a year, and for many it doesn't even require a degree in computer engineering
or advanced coding skills. I'd love to see how you spend your day. Could you maybe build me a demo of a Bloomberg editor. Let's say absolutely, Let's go and do that.
Okay.
Open AI's consumer facing tool chat GPT is pre prompted. It's set up to be a helpful assistant. To be a professional prompt engineer, you have to get into the guts of the large language model. That means changing the default prompt. This is the chat the open a AI playground. So this is a more advanced, more neutral more powerful, customizable version of what you see when you load uptrat GBT.
Yeah, exactly what we're going to look to do is help you write some interesting content around different prompt engineering techniques.
So this is you're making a sort of helpful assistant for me to write a story about prompt engineering as a Boomberg journalist.
Yes, exactly.
Yeah, a seemingly semi mystifying technique. Prompt engineering is an action packed interface between human creativity and AI's data crunching power. Yeah, it's making me a bit sick.
Yeah, I think so, I think so. I think there's more work that we need to do.
There are a few tips and tricks that Albert uses to improve as results change the temperature or the randomness of the output. Tell it to take things step by step and feed it a previously written article as an example. After thirty minutes of work, we didn't quite manage to replace my editor with AI, but the responses did get more useful as Albert added more context and instructions.
The one thing that probably haven't done a lot of here is actually treating this as a dialogue is actually a really useful way often of getting better rest.
Okay, we're getting somewhere. Now there's a germ of an idea that's interesting. Albert says PROMPT engineering has turned the English language into a programming language. He's had no technical education or programming experience. In fact, he studied history a university before starting his career as a consultant. He says his non scientific background mix him more not less qualified to wrangle AI tools. When it comes to fearing this
technology replacing our jobs, he's not as concerned. And when will my editor just be a Tracy BT type model that we've developed together.
I would probably think of it the other way around. I think we as humans actually might become more like than editor. Will outline what we want the model to do, we'll give it some test cases, we'll give it some examples, and then we'll be the ones essentially doing that that.
Editing liberal arts.
Back on the RST, Prompt Engineer Albert Feltz. They're speaking with Bloomberg's Conrad Cuilty Harper coming up runway securing an addition one hundred and forty million dollars in funding from some big names in the world of tech. You're going to sit down for an exclusive conversation with the CEO next, there's a Bloomberg.
AI startup, Runway, whose.
Software can create a short video from just a few type words, has raised one hundred and forty one million dollars in funding from the likes of Google and Video Salesforce and other investors joining us. Now, I'm pleased to say it's the CEO of Runway, Chris Valenzuela.
Chris, congratulations. One hundred and forty.
One million is actually an extension of your series C was this inbound interest coming from these big players?
Yeah, thank you for having me.
We've been working on Runway for now five years or so and so very thrill and excited to be partnering with with the likes of Google and Nvidia Saltfares on this new round that extends our kind of like bandwidth and capacity to go continue innovating on core research that we're working on since the release of Gen one and Gen two, which are video generation models. There's still a long break to go, but we're excited about taking this new funding to improve and scale our research efforts.
And it's interesting, isn't it that previously been raising money from the big VC names, but now it seems like it's strategic investors coming in, How do you think their interest was peaked and wanting to come in and do this extension?
Sure, I mean there are different parts of if I would say, the development of these models, and you can think about it as two phases. The first phase is building the baseline foundational models, and that takes a lot of time and kind of like work and engineering counter efforts.
The second part is kind of I would say, the phase that we're in right now, which is getting these models out to more people, to more customers, to most enterprises, and that being able to work with companies like Google and Biding selfware so really like speeds up that and allows us to innovate at the speed that we need
to innovate. We're entering that phase of development that requires a different set of like attitudes and I would stay research kind of like mentalities and so STUDI early, but excited to be working with some of the best companies in the world.
And of course, Chris, I mean, you haven't disclosed your valuation, but we understand that's there about one and a half billion dollars. Just talk to us about the costs from your side, you're going to be using this to of course beef up team to keep on creating innovating, but the compute power must be massive. Is that something where Google comes in.
I think there are two parts again of the stage to where we are with these generaling models and video models and image models tend to work very different with the language models that perhaps are the ones that are more discussed and discuss nowadays. For video models and for image models, there's a scale component on the infrastructure side of things, and making sure that models are fast and reliable and you can train large video models is a
complex ask. So we're putting a lot of resources on making that happen with state of the art technology and research and compute. But on the other side, there's so much on the infrastructure that needs to get built that hasn't yet been built. That also requires a lot of effort and time and great people to build. We are an incredible small team that's been really pushing the boundaries of.
The field for the last four or five.
Years since we since we started on building these models and these technologies, and so really making sure that you can work with the best on both the research aspects and the engineering aspects. It's really important and the resources will go pretty much to both ends on the research and engineering side as well.
And you have been working on the infrastructure side a lot already Chris Layman's terms, can you tell us like some of the issues there, what actually what architecture you need to build to scale at the rate.
You want to?
For sure, that's a great question.
So if you think about it really on Jenney one and Gene two, which are two of the video generation models we just release, and they say just because it's only been like four or five weeks since we did it, there's so much work that goes into making these models do what they're able to do, which is generate video. Right you're generating video. You stop for a minute and
think about it. You're generating video with words, which five years ago, ten years ago we have thought to be like impossible or very hard, Allen most feels like magic these days, and so making sure that you can do that, it's in itself a very technical, difficult fit. And so research have gotten into a phone we're able to kind of like generate the videos you're seeing right now, and
those are just being generated with words. But then on the other side of things, this is a very complex infrastructure that allows anyone really with a mobile phone or a desktop application or an Internet connection to be able to do this. And in order to do that, you need to have an inference and a streaming service that's able to generate those frames almost on a real time basis.
Our videos take like forty seconds to generate, which is fast, and we're making sure we can reduce that to hopefully at some point or at some time do it in a real time basis. So everything that we need to build make that happen hasn't been built because these models and these technology didn't exist literally a couple of months ago.
And so really from the growna with think ourselves as a full start company, which which we ask or we ask ourselves to be able to build everything that allows you to deploy these models to our filmmakers and studios and creative companies around the world that are already using and leveraging the technology.
I've been using and not really leveraging, but interested in it and playing around with Gen two today. And it's beautiful some creations that you make. But it's interesting that sort of faces are the area that perhaps aren't that easy to provide, and also the way in which you have to have an abundance of caution. How are you thinking about some of the well still evolving regulation and the way in which that you can ensure content remains well moderated in some way.
Really first, and we take safety extremely seriously in Run. We have a team dedicated just to make sure we have the right filters and safety policies in place, and so we're taking an our really.
Costious approach here.
The analogy I've been trying to I guess, really think about and best communicate the stages of development that we're in with these models is to think about this as a new kind of camera, Right, so we're putting a new kind of camera in the hands of filmmakers and artists, and there's new norms and new policies and new kind of like ways of using descinality that will have to be evolved and created as we start improving their quality
of these models. Again, these are very new models, and so there's there's a long way to go by taking a very i would say opinionated matter or making the similarity safe and really important aspecture is controllable, like these tools are tools for artists that are meant to augment a creative process, and so by making sure that you can use them in a safely environment in our creative environment.
For us, it's always been and would always be our priority.
Our actors right to be worried in the way that we're currently saying.
No.
I think you know, most of the of the usage of the overall thoughts around video and AI video generation comes from the sea that you're going to be able to like just type in a few words and generate a Hollywood movie.
The truth is that steps far from being actually the case.
These are tools that would allow you to itterate faster on your processes and your ideas and your projects, but in no mean they're trying to replace anything, They're just augmenting those things. And so perhaps a question is like how fast and how easier are we're going to be able to shoot and do what was seen with an actor in the way that perhaps you couldn't today because it's either too expensive or too time consuming these day to do it.
Chris, great talking to you. Thank you on the planning round. Chris Valezuela, he's a Runway CEO. Let's talk streaming wars and the costs involved. Amazon CEO Any Jesse is taking a hard look at how the company's Hollywood studio spends on original TV programming. I'm kind of wondering why it shows seem to cost more than netflixesus.
Chris Palmry is here to break it all down. And we know they're on a cost cutting focus. We know the jobs that are being shedded. It's an obvious place to look at the content. Why is it ending up more expensive?
Well, they did this massive pivot. If you remember when Amazon started in video, they were going to use their sort of experience in books and find out these sort of arty shows that people they thought their book customers might want to watch.
If remember programming.
Like Transparent, and then they Jeff Bezos said I want a Game of Thrones and they started this massive spending push. We saw the Ring of Power of the Lord of the Rings spin off and massive shows like Citadel, which has kind of been a disappointment from the.
Russo brothers this year.
And so now, as you mentioned, Andy Jase, the CEO, is taking a hard look, calling the entertainment executives to headquarters. We're asking for detailed accounts of their spending and cleaning the big shows.
It's part of an overall.
Sweep of what Hollywood's doing, looking at spending in the streaming wars, but you know, exacerbated by Amazon situation, and.
Is there any way that they can reduce the costs without crumping on some of the wins here or indeed the fact that they're investing because they want to keep us addicted to the broader Amazon offering, not just of course Amazon Prime.
Well that's always been the argument and why sort of Amazon's video business has been so much under the radar because they have this ulterior motive. We just get people subscribing to the Prime service and lock them in for shipping and all kinds of purchases. But you know, they can't do that forever and not you know, face the reality, which is everybody's just been spending too much money on TV shows. It look like the era PEAKTV has finally arrived.
That's for sure, just as the economy takes a turn. Of course, Chris palmery, thank you so much for joining us. You feel a little better. Meanwhile, let's have a look at what's happening in terms of Rivian's first commercial shipment outside the US. Of course, Rivian based upon Amazon in many ways, providing they're vans electric vans?
Who are them?
Then?
And Lulo is the expert for.
This, and he's getting ready for a pretty key interview on the back of it, ed where exactly are you?
Surprise just touched down in southern California and we are heading over to Ribban's HQ in Irvine through a conversation with R. J. Scanner and he's the CEO. You're absolutely right. It's been a strange week. I think Ribban's up for a six consecutive session trading at February high. It's been a real turning point for the company and they posted production cousy, production and delivery figures above Wall Street expectations this Monday.
The stop has a lot of momentum.
But remember that this is a company that went from the euphoria of the biggest IPO since Facebook's now known as Meta to a horrible twenty twenty two where it was the second worst performer on the NATS that one hundred and then got taken off then as that one hundred.
So there's two key pieces of news on what we're really talking about.
Characters and we've started to see analysts change their tune, haven't we had.
We've started to see Da Davison.
Some of the other analysts really deciding that now might be an opportunity time to buy. The headlines are looking more positive. What's the headline you want to create today? What do you really want to understand from RJ?
Yeah, I want to understand why they're not doing more. You know that the factory in Normal, Illinois is capable of building many more than the fifty thousand evs that they've guided investors on for this year.
Also just really interested in RJ Scarring. She's forty years old.
He founded this company in twenty eleven, had the big IPO in twenty twenty one.
I just want to ask him is he having fun?
You know, being a founder CEO is a rare thing in this day and age. But he is really at the cutting edge of their engineering. Running the business has brought people in. The question is why is Rivian not ahead at this point They're still kind of lunked in with that large pack of ELI startups to the suffering, but Rivian's actually producing the skin.
Well so and Ludlow battling some of the tannoys over there in the airport. Thank you so much for bringing us what we're really excited about. A live interview a little bit later with Ed Ludlow and of course a Rivian CEO RJ. Scaring, and a little bit of what he wants to get out of that particular conversation.
We can't wait for it. Meanwhile. That does it for this edition of Bloomberg Technology.
Do not forget to check out the podcast find out on the terminal as well as online on Apples, Spotify, and iHeart from New York from Southern California, This is Bloomberg