In today's episode of AI Daily, we bring you the latest news from the world of AI. First up, we discuss the Google AI Pact with the EU. As Europe pushes for AI regulations, Google is the first corporate partner to join this initiative. We delve into the importance of this partnership and the potential impact it may have on AI regulation worldwide. Next, we explore the exciting Alexandria Initiative, which aims to embed the entire internet. This project, led by microcosm, focuses on making research papers more accessible and improving collaboration and search capabilities. We discuss the benefits and potential applications of this groundbreaking technology. Lastly, we dive into the soaring growth of NVIDIA and the buzzing discussions around its stock. While NVIDIA's market cap continues to rise, we caution against FOMO (Fear of Missing Out) and remind viewers to consider other players in the market. Join us for this informative episode of AI Daily!
Key Points
Google AI Pact with EU:
Google and the EU are in discussions similar to OpenAI and the US government.
Europe is pushing for AI regulation, aiming to implement it by the end of the year.
Google is the first corporate partner to join the AI pact with the EU.
There are concerns about the lack of public involvement and actual actions in AI regulation discussions, with accusations of virtue signaling by politicians and companies.
Alexandria Initiative to Embed the Internet:
Microcosm has developed Alexandria, a project aiming to embed the entire internet, starting with titles and abstracts of research papers from archives.
Alexandria improves search and collaboration by providing easier access to research papers and enhancing semantic web capabilities.
The project is open source, allowing users to download the models and participate in the open collective to decide what content to embed next.
The advancements in embedding technology, such as cheaper and more useful embeddings, along with open source models, have made projects like Alexandria possible now.
NVIDIA’s Growth:
Nvidia's stock market value is skyrocketing, prompting discussions on whether it's still a good time to buy the stock.
While the consensus is that Nvidia will continue to grow, there is a cautionary note about other players entering the market and potential price fluctuations.
The situation is reminiscent of internet-driven hype and market volatility, where stocks can experience significant increases and decreases.
Nvidia is seen as having a monopoly on GPUs and AI infrastructure, but the emergence of custom boards and other companies may challenge their dominance in the long term.
Links Mentioned:
Follow us on Twitter:
Subscribe to our Substack:
Transcript:
Farb: Good morning and welcome to AI Daily. We're here with your hot AI news for today. We got some news from Google, from a new Alexandria, and from our friends over there at Nvidia. Let's get started with the google
AI Pact with the eu. It looks like Google and the EU are. Talking in the same way that, uh, OpenAI and the us, uh, government were talking mm-hmm.
Google not to be left behind, uh, wants to show that they are also, uh, working with governments to regulate AI in ways that nobody is apparently willing to mention yet. Ethan, though, what do you think about this story? Why is it important?
Ethan: Yeah, this is, you know, just as you mentioned, the US government putting stuff out around regulation.
We've talked about before, we've talked about European regulation before, but Europe is, you know, GDPR, they've had a history here. They are really gung-ho about this. Some of their leaders are saying, Hey, we need to get something implemented by the end of the year. So then we have an implementation cycle over year, and this actually starts happening now, so, If they're in a rush to get this done, they're looking for, you know, the big corporate partner to kind of push this forward with them.
Google is the first one who signed up, so they're calling this their AI packed. And like you said, they have a little bit of framework we've talked about before around AI use, you know, limiting open source use and stopping them from reaching European citizens, but, They're pushing hard on this, and I think it's important for people to know that, you know, Europe has implemented gdpr, which affected the pretty much entire internet around the world.
We got beautiful cookie banners, um, out of that. So seeing them push harder and harder day over day to get AI regulation, not five years from now, but in the next year, is what they want. It's important to know and keep an eye on.
Farb: This so far sounds like a lot of people trying to make sure that they're in the news cycle about talking about AI regulation.
I, I couldn't find the single thing that they were discussing. It sounds like a lot of, you know, trying to get ahead of the story, which is smart, which is smart for the EU purposes and for Google's purposes. But it's getting kind of tired in my humble opinion around, you know, how many, how much coplay and virtue signaling we can do about AI regulation without actually doing anything.
And you know, I haven't seen a single example, whether it's the US government or the eu. Establishing some sort of way for the public to actually interact with this, for the public to have it say, it just seems like it's something for politicians to make it seem like they're, you know, doing things with big tech companies and, you know, doing lots of regulation.
Uh, where are the people, where is open source involved in this? Where is the average EU American citizen involved in Conner? What do you, what do you think?
Conner: I think many people, many companies are worried this is gonna be as big of an impact that Ethan said as GDPR was. So I think, like Ethan said, they're really trying to get ahead of the curb.
Um, apparently EU is going to or is already working with their companies. Google's just the first one they felt we talked about, but many companies are worried about this. We saw even open ai. We saw Sam Alman say that if open AI can't figure out EU regulations, they would pull outta the EU.
Farb: You know, but Sam's going out there and, you know, talking in public, he's talking on Twitter, he's doing lots and lots of talks at different conferences.
You know, he asked the group of folks at, I, I think in, in Germany, whether they think open source, you know, uh, uh, G P T five should be open sourced right away. Uh, most of the people raised their hands. He said, well, we're not gonna do that. But it's interesting to see how many people voted for that. So, You know, I'll say I think there's a big difference between what Sam is doing and what some of these other folks are doing.
They seem to be doing a lot of virtue signaling. Uh, where he seems to be actually trying to interact with real people out in the world to discuss, you know, what to do about this. I think Mike, I think he and
Conner: open AI know more of what they want. I think the problem with Google, the problem with eu, the problem with Microsoft, they only know what they want yet, so that's why they'll probably very hush hush is, would be my guess.
Farb: I mean, yeah, I think somebody at Google said, Hey, Sundar, you need to, you need to, we need to get some news stories about you talking to governments as well. Uh, and he's like, okay, let's, let's do some, let's do a AI pact, some, you know, just so that there's enough of a word around it that some, you know, news stories will cover it.
What were you saying, Ethan?
Ethan: Well, we saw them put, you know, we talked about this on a previous show. The EU has put out the framework that they're kind of pushing towards. They wanted to limit some of the open source access. They wanted to limit some of the more data privacy concerns and training of these models.
And at least Commissioner Breton, you know, he's been active on Twitter. He is trying to push these regulations forward. I think, you know, as much as you both have valid concerns, um, about, you know, the lack of public intervention in this and how much virtue signaling goes along, which I completely agree with, they're worried about it and they're pushing towards it.
And just like they did, what are they worried about? They're worried about AI risk. They're worried about what's the risk? I don't know.
Farb: Yeah, okay. That's what I'm saying. Like what? What's the risk here? Understand if Sam is saying there's risks involved. Uh, I can understand that because the guy's been building this for years.
He knows what's coming. Uh, for some, you know, EU had to say there's risks involved and not be able to mention a single one of them, uh, seems most spurious.
Conner: They probably think it's the same risk that Sam Self has said about jobs, about privacy. I think the EU is just more preventive.
Farb: I don't think those are the risks that Sam's really talking about.
Conner: I mean, that's what he mentioned. At least jobs in front of Congress.
Farb: I think his concerns are more existential and it's easy to say, well, my risks, my concerns are existential too. Uh, but I don't think any of these people could be pushed to actually explain what any, any of those existential risks are. I bet you Sam could
Ethan: possibly, I, I don't think it matters at the end of the day, I think they're trying to get ahead of the curve.
We saw what happened with the internet with Cambridge Analytica, and how much happens when you don't talk about this stuff at all until you know exactly what the risks are, know exactly what to say and know exactly what to do. So you can't just wait till that stuff happens. You gotta get ahead of the curve, and that's what they're trying to do.
Farb: Yeah, I don't, I'm not buying it from the EU. Good luck guys over there. Real American. Yeah, that's right. Go America. If you don't like it, go to the eu. Uh, let's go on to our next story. The entirety of Alexandria is being rebirthed. Uh, this project is trying to embed the entire internet. It's super cool.
Ethan, can you tell us a little bit about what it is and, and why you think it's important?
Ethan: Yeah, it's super cool. So this comes out of microcosm, which is, you know, they're kind of a new collective. Some amazing engineers are a part of it, and they've put together Alexandria, which what they've done is take all of archives, research papers.
What they're only doing is the titles and the abstracts right now. But they're embedding all of these. So improving, you know, if you've ever used archive and you're trying to search up, you know, Hey, how do I do distributed computing? You're gonna get not the results you're looking for. So, This move towards embedding, you know, some people call it the semantic web.
How can we improve collaboration? How can we improve search? And seeing this like open collective, it's open source code, you can download the models right now. So they're putting together an open collective to take all of archives, research papers, take the abstracts and make them easier to work with. And I think, does it stay updated though?
I'm not sure how they're continuing to update them. Um, I know embeddings, they talked about how embeddings are really cheap. Um, they've done I think a couple million of the research papers now, so I'm not sure how often they're gonna update them, but it is an up easy update process, so hopefully they keep up to date with
Farb: it because currently, you know, if you put a, you know, uh, chat, GPTs browsing doesn't read PDFs right now.
It's kind of, it's kind of interesting and so it's tough for it to dig into anything on archive. And this project seems. Amazing and useful. I guess my question is, there's so much stuff hitting archive on a daily basis. If it's not being embedded pretty regularly, you end up back in this situation where you're searching for something and it's like, well, sorry, we, we haven't read that one yet.
Connor, what do you think?
Conner: Yeah, I mean the big issue with get, again, like you said, like search or getting high quality data is open data, like a, a lot of moats of these big companies is this data that can't be synthetically generated and that can on, you only want the real data and having the actual archive papers entirely embedded and whatever they're going to, and bed next is very nice and safe, so.
Farb: Very cool. What, what else could you know you, what else do you imagine they could use this technology for, other than the ones that they've kind of mentioned? I think they mentioned patents, uh, some financial stuff. Anything else come to mind that. I think it would be a cool application of this news they're doing as well, I think.
Ethan: Um, so if you can just better embed news, you can get more up to date and kind of search for what you want, what are you're trying to find out versus just, you know, Hey, what's going on in X today? You can actually answer your questions about current news. So a lot of things, probably just the fact that they can create an open collective and get a lot of people working on it together, voting on what they want embedded next, you got this like decentralized movement to, you know, semanticized the web.
Farb: Why do you think this is happening now and you know, didn't happen already?
Ethan: Why is, why is this embeddings are so much cheaper now? Number one, they're a lot easier to use. You also have, they're using, I believe, instructor xl, so an open source model, so cheaper embeddings, open source, LLMs, both affecting this, you know, decentralized movement to semanticized the web, like I said.
So two fundamental technologies that just weren't here last year.
Conner: Embeddings are far more useful now too, especially having these better models, especially having open source models like LAMA Embeddings are now more useful. You can search and you can embed, and then you can use those. Now we have open data sets for search inside of open models.
Farb: Why are, why are embeddings better now? Why are they more useful now?
Conner: Because, I mean, before, before Llama, before these models, you could search and you could find these embeddings, but you would just get back the entire abstract. Now, LLA has something to explain. Or, and embeddings have better
Ethan: Vector databases too. Better infrastructure to search these things. All these pieces have been kind of coming together over the past few years. Mm-hmm.
Farb: Very cool. Very cool. Well, all right, let's move on to our last story. NVIDIA blowing up the market cap, going through the roof. Everybody on the inter, everybody on Twitter asking if it's not too late for them to buy Nvidia stock.
I'm gonna go ahead and say Nvidia stock is gonna continue to continue to grow. I don't think that's a controversial statement. I will say, however, that there are a lot of folks that are gonna be getting into this game, and I don't think this is a, you know, one horse wins the race type of situation. So if you're gonna start, uh, spraying and praying your, your money around, you may wanna, you know, save a little bit for some of the other folks.
This is, this to me, smells of classic internet drives everybody crazy around some topic. Um, prices go through the roof of something, prices fall through the floor of something. You know, this is the, uh, this is the positive version of being canceled on the internet when your stock gets driven through the roof.
Uh, because some people, you know, there was some news story, uh, and then some dudes on Reddit started pumping the price. Uh, I say main, you know, remain calm. You're not gonna miss out on a 10 x on the Nvidia stock. Probably that that happens over a week. Uh, NVIDIA's, NVIDIA's probably gonna continue to grow, but there's gonna be a lot of other players here.
Ethan, what do you have to say?
Ethan: Yeah, I, I'll definitely say first, you know, not financial advice from any of us, but w we were talking about this on that air chat channel of Nvidia, and you know, someone made a great comment how early days of the internet, you know, everyone was pouring into Cisco. They had this monopoly on internet infrastructure and everyone thought, you know, they were the long term winners.
So you had a lot of growth priced in with something like Cisco in the early days. And I think we're seeing a repeat of that here with Nvidia. You know, they have a monopoly on GPUs. All AI infrastructure is pretty much either trained or ran on Nvidia hardware, and I think people see this, you know what is a clear monopoly and a clear winner in this space and are pouring in, so, you know, how will it go long term?
I think likely like Cisco, but we'll see.
Farb: Little reflexivity may occur here. We'll see. What do you think Connor?
Conner: Yeah, I think of course NVIDIA does definitely have a bit of a monopoly on GPUs here. But as we saw with Meta, as we saw with Edge J ai, we're definitely gonna see more custom ASIC boards, more custom boards that work for these models, that we don't need these general purpose GPUs anymore and even GPUs themselves.
Definitely gonna see more of other out of o other companies as they become far more useful for AI and AI inference. Apple, Google, they're running these models on devices. So I agree. I think a video will definitely do very well in the future and will continue to, but maybe not as well as many people think.
So processes are eating the world.
Farb: I don't think that's stopping any time in our lifetimes. May, maybe not for the, uh, rest of the entirety of humanity. Uh, so. Well, we'll see how this all, how this all plays out. Well, fantastic. Three great stories today. Let's talk a little bit about, uh, what we're seeing out in the world.
I, I saw this cool example of a tool that you can plug into your CI if you're a dev shop, if you're an engineer. Uh, it'll use, you can also give it some rules around how to evaluate your code and it'll use an LLM to help evaluate your code, run some tests for you. And, uh, what was really cool about it, it seems to just plug, plug right into your ci, plug right into your continuous integration so that it's, it's nice and easy and not a lot of extra work, uh, for folks.
I don't know if either of you had a chance to take a look at that or what you guys thought.
Ethan: No, I, I hadn't looked. I saw you had sent it over, checked it out a bit. You know, these tools that embed in like code workflows, I've always liked. They're, instead of this co-pilot that devs used, they're kind of this agent that exists within your code base, within your company, like automatically doing ci, automatically doing documentation.
So I love these types of angles. I didn't get to look at this tool too much, but yeah. Very cool. Hmm.
Conner: Yeah, it's, it's nice to have something that can directly inference on your changes directly. Look at your commits and mm-hmm. Be a bit of a fail proof for when your tabs mess up.
Farb: So they had some concerns about the pricing though.
They were like, be careful with how you use this thing. The, the, the cost could get out of hand pretty, pretty quickly, so yeah, it'll be cooler to see once the cost of these things come, come down as well. They'll just be everywhere. What are you guys seeing?
Conner: Got it. Uh, I saw Diagram. Diagram, launched their new site and had a big emphasis on their UI AI where they're essentially making AI models that are complete centered around UI.
Figma, of course, is an amazing place to build and prototype ui, but diagram tools are making it very, making it far quicker to make these things. And a couple of, one of the models they have is making glyphs. So you say, Hey, I want. An image of a guy drinking a water and it makes a perfect little icon for you.
Or they're having AI to make the layers of your Figma prototypes or the text or anything really. And it's, it's nice to see that prototypings just gonna get easier, um, especially with Figma.
Farb: So I haven't had a chance to try. But it looks like it connects into your, but plugs into Figma directly. Yep.
Farb: Direct plug. Big win. I, I was playing with Photoshop's, uh, beta generative Phil yesterday, and it's, it's really cool. I tweet, I tweeted about it like the. It, I gave it a little picture and I asked it to sort of expand the picture out and the degree to which it was exactly what you would expect was freakish.
I, I, I, I was almost. I, my head was spinning a bit around it being like, how did it generate that part? It's such exactly what you would want to see. Uh, totally mind blowing over there. Ethan, what are you seeing?
Ethan: Um, yeah, super cool on that. I love, you know, a more aesthetic world people can use easier, but, um, I, I saw spell book.
They're another kind of AI for lawyers. They raised a healthy seed round a bit ago, so excited for them. I love the plays on these kind of vertical lms. How do you integrate in a doctor's workflow? A lawyer's workflow? So the more people working on that, I think. The better. And it was cool to see is it available, can you use it or you know, I don't think so yet.
I'm, I'm not sure. I'm not a lawyer so I haven't dove into it, but there seem like there are a lot
Farb: of these lawyer ones. It'll be interesting to see which ones, which ones win out. Again, it may not be a winner take all type thing.
Ethan: Yeah, it's a B2B SaaS world, so.
Farb: What'll also be interesting to see is, you know, we're gonna have the, the, the, the same problems of legacy software and legacy solutions that, that haunt the world today will probably haunt the world with AI as well.
So, yes, you know, five years down the road, uh, there'll be a whole new wave of AI companies because of the, the ones that have been around for five years, were built on previous ai, and their infrastructure is not made to easily use the new super intelligent ais. And we'll see another wave of, you know, AI lawyers until we've all been sued to death.
Absolutely. Welcome to the future. Thank you all for joining us. Have a great rest of your day, and we'll see you on the next episode of AI Daily.
Conner: Thank you guys.
Google-EU Pact on AI, Alexandria Internet Embed, & NVIDIA's Growth