Leveraging AI

143 | A 5 Pillar plan for US AI dominance, DOJ is breaking Google, as OpenAI is working on their own browser, and many other important AI news from the week ending on November 22nd 2024

• Isar Meitis • Season 1 • Episode 143

Share episode

Is the U.S. Prepared for the AI Race?

The global AI race is heating up, and the U.S. is gearing up for the fight. But is it moving fast enough to stay ahead? What are the real challenges—and opportunities—facing America's infrastructure and AI development?

In this episode of The Leveraging AI Podcast, host Isar Meitis dives into the biggest news of the week, spotlighting the five-pillar plan OpenAI has laid out to ensure the U.S. remains a global AI leader. From data centers to government partnerships and even nuclear power, we break down what it will take to keep America at the forefront of AI innovation.

But there's more—beyond just keeping up, there's staying ahead of competitors like China, who are pushing the limits of AI development with groundbreaking models. What does this mean for U.S. businesses, jobs, and economic strategy in the next decade?

In this session, you'll discover:

  • OpenAI’s bold five-pillar plan to keep the U.S. competitive in AI
  • Why AI-focused economic zones and fiber networks are the new “highways” of innovation
  • How U.S. government initiatives could create jobs—but will it be enough?
  • The surprising rise of China’s AI models and what it means for U.S. dominance
  • Nuclear energy’s critical role in powering future data centers—clean energy or not?
  • OpenAI’s $138 million challenge to U.S. competitors
  • Major developments from Microsoft’s Ignite event—AI is everywhere
  • The DOJ's bombshell move to break up Google’s search monopoly—what’s at stake for the future of tech?
  • New AI models, breakthroughs, and infrastructure updates you need to know to stay ahead

Stay informed. Stay ahead.

If you enjoyed this episode, take a moment to share it with a colleague or leave a review on your favorite podcast platform. Your support helps us continue bringing cutting-edge AI insights to your feed!


About Leveraging AI

If you’ve enjoyed or benefited from some of the insights of this episode, leave us a five-star review on your favorite podcast platform, and let us know what you learned, found helpful, or liked most about this show!

Speaker 2:

Hello and welcome to a News Weekend episode of The Leveraging AI Podcast. The podcast that shares practical, ethical ways to leverage AI to improve efficiency, grow your business and advance your career. This is Isar Meitis, your host, and we have a jam packed episode of news for this weekend. We could have probably done two episodes and yet I'm going to try to stick to the format I've been doing in the past few weeks, focusing on a few bigger topics and then doing a rapid fire announcements on some of the other important news from this week. The top focus for this week is going to be the dramatic changes in global and specifically US infrastructure and how it is changing right now and how open AI envisioning to change in the future to keep the US in the front of the AI race. We're also going to talk about some big announcements from Microsoft events this week, as well as dramatic changes in the world of search. So lots to talk about, even in the big topics. So let's get started. This week, OpenAI released a comprehensive five pillar plan for the American AI innovation. And this five pillar plan is supposed to keep the U. S. ahead of its global competition in the AI race. The five pillars are AI economic zones that will be developed to put data centers and everything that's related to them with university linked areas in order to create locations Where industry, research, data, and science can all coexist in multiple places in the U. S., they're suggesting to create a National Highway Act for fiber and other infrastructure like wireless and gas in order to support this new development. They are seeking increased government funding for AI job creation projects, and Pillar number four is a grand alliance with allies from other supporting countries to counter Russian and Chinese development of AI. And the last and fifth pillar is investment in nuclear development in order to enable the power requirements with clean energy. for these new data centers. Now, none of these things are new. If you've been listening to this show, you know that most of these topics have been in discussion through everybody in the industry for a while, but putting it together as a plan and putting it in front of the government is definitely not a bad thing. How will that turn into anything now with the changes in the government? I don't know. I don't think anybody knows We're going to talk a little more about government impact and what that means to things that are happening right now Later on in this episode. But the obvious things are there's a huge required investment In infrastructure across multiple types of infrastructure, Whether it's data centers, power, supporting these data centers and even everything around them as far as keeping those systems running. There are several different assessments over there on how many tens of thousands of jobs every one of those hubs will create. there's also obviously the point that needs to be discussed of how many jobs it's going to take because it's very obvious to anybody who's in the AI world. That AI will displace many of the jobs that we know today, and hopefully these new jobs that are going to get created in order to support these data centers and data hubs and innovation hubs will offset at least some of it. My personal opinion that A, yes, it will create a lot of jobs. B, it will take a much longer time to create these jobs than it will take to displace many of the jobs that we know today. And I think there's going to be an area in between in time that we're going to be in a, I'll say, interesting economic scenario where many, many people with high paying jobs will find themselves without a job. And that will have obviously a very significant impact on the economy, which may have an impact on the ability to invest in those initiatives. But as a direction, I think it's a very good direction for the U. S. to do that. Just to prove the point, a Chinese company just released a new model called DeepSeek R1. DeepSeek R1 is a new type of model that follows the same concept of O1 from OpenAI, which is a model that OpenAI released recently that uses a slightly different approach to advancing AI, which they call test time compute, which we know as a thinking model, meaning a lot of the effort is invested, not necessarily in training bigger models with more data and with more compute, but rather in letting it time to think and evaluate the task ahead and providing a time to think about the problem and come up with better answers. And that has been proven to be significantly better. So that's this kind of model that they released. This model is already matching OpenAI's O1 preview, in several different benchmarks. Now the interesting thing about this is First of all, it takes longer time to think than average models, sometimes tens of seconds before it provides an answer, which sounds counterintuitive in a world where everybody wants instant gratification, but they're going to release it as an open source model. So an open source model that can compete with O1, maybe not on everything, but on several different directions. But the most interesting thing about this is that they trained this model on a 10 bit machine. 10, 000 NVIDIA A100 GPUs, meaning the older version of GPUs and only 10, 000 of them. To put things in perspective, Elon latest AI data center has 100, 000 of the more advanced GPUs or a mix between them, and he's planning to add another 100, 000 to that. another order of magnitude bigger or an order of magnitude times two and it only cost them 138 million dollars to train the model. Now I know the sentence of only 138 million sounds ridiculous but it's not ridiculous in the AI world where companies are talking about a billion or 10 billion or 5 billion to train a model. training a model that can compete with the most advanced model that a U. S. company has so far with only 138 million is a very significant achievement, which shows that, first of all, the test time compute approach is in the right direction as far as providing us more AI capabilities for significantly less money But it also shows that the chinese are definitely in the race Despite all the attempts to limit their access to gpus and advanced compute capabilities. Now the model still has multiple issues and it struggles in specific aspects And it can easily be jailbroken and there's already been multiple examples, which is You A little scary when you think about it when you have a very advanced model that people can jailbreak and do different things with it it is blocking some sensitive queries But not everything and it obviously aligned with anything the chinese government wants but that kind of shows you Why having those kind of investment that I mentioned in the beginning to keep the U. S. ahead are so critical. Staying on the same topic of infrastructure in the U. S. Kairos Power received the NRC approval to build two test reactors in Oak Ridge, Tennessee, making a significant advancement in the next generation of nuclear technology. This comes in together with a deal that they have with Google to build new smaller type of reactors to power their data centers. Now the project is supposed to have the first reactor up and running in 2030 and the full project completion in 2035. This is on one hand very exciting news because it's going to provide clean energy to these data centers. The question is what happens between now and 2030 or 2035 and it's obviously going to be carbon based fuels that are going to be powering these data centers at least in the near future. Now two interesting things as far as tying it back to the U. S. government, U. S. Department of Energy has awarded 303 million dollars for this project, and they have expedited the approval process, and again that will sound funny, but the expedited process for that has been project was 18 months, which apparently is shorter than what it usually takes, but it shows that there is a government commitment in order to push new types of nuclear energy to support the needs of the AI development. Staying on the same topic, Crusoe Energy just raised 686 million for data centers. Now, they were trying to raise more. They were trying to raise 818 million, but 686 million is definitely not bad. Multiple companies have participated in the funding. And the total funding that they've raised is close to 1. 5 billion, including equity and debt. And they are planning to build massive data centers in Texas. They already have a conversation on leasing these data centers to giants like Oracle, Microsoft, and OpenAI. Bye bye. Now, to put things in perspective, This field of building data centers is obviously on fire. So other competitors have raised significant amount of money or weave, which is another company in that field raised 12. 7 billion. Lambda Labs raised 500 million. TogetherAI raised 106 million. And so lots and lots and lots of money is being poured into data center building. And the IDC projects the global data center electricity consumption will double by 2028. And it's also really, really big. And Morgan Stanley estimate that 2. 5 billion metric tons of CO2 emissions will be generated just by the power consumption of data centers by 2030. So that shows you that The fifth pillar that OpenAI refers to as far as accelerating nuclear development is critical just because of the amount of power these new data centers will require. Staying on the same infrastructure topic, Logic Monitor, which is a company that builds software that monitors data centers and how they operate, has raised 800 million. million dollars on a 2. 4 billion valuation. Their total funding has reached 942 million in the past few years. And they are building, as I mentioned, monitoring systems for these data centers that provide a lot of A efficiencies and B safety with many significant benefits to the companies who are implementing those capabilities, I'm not going to dive into all of them, but it's very, very obvious that there's an entire ecosystem with a very significant supply chain that are all benefiting from this new push to data centers and power, and that is not going to stop anytime soon. Now speaking about supply chain, XAI just cancelled a 6 billion AI server order to Supermicro SMCI that comes immediately after they came under investigation for accounting manipulation, export violation to China and Russia. and potential NASDAQ delisting. If they are going to be delisted, they will have to pay 1. 7 billion in convertible notes. And their stock took a 35 percent plummet the day those allegations were announced. And so they're not in a good situation. And as I mentioned, to make it worse, XAI just pulled those 6 billion of server order from them and pushed it over to Dell, which they had an arrangement with before. But before, the order was broken into Supermicro on one side and Dell on the other. And now the entire order from X. AI is going to go to Dell. Dell. com. And other beneficiaries are companies like Wistron and Inventech. Both are Dell suppliers. So again, there's a very long supply chain that is benefiting with very significant amounts from everything that's going on right now in the AI space. Now, speaking about XAI and their ability to spend this kind of money, they have just raised. Additional$5 billion in addition to the$6 billion they raised just earlier this year, valuing extra AI at$50 billion. the interesting thing about this number beyond the fact it's an insane number for a company that is about a year old, is that the fact that this 50 billion in valuation is more than the$44 billion that Elon Musk paid for Twitter now X. So he made a very big investment in a company. More or less crashed it to the ground. So the that people are giving Twitter right now, or X right now, are in single digits. So 7 to 8 billion instead of the 44 billion it was worth when you bought it. But XAI, which was spun out of that, is now worth 50 billion. So In any situation, I wouldn't bet against Elon in anything that he's doing, but definitely they're on a very high growth. They're building gigantic data centers. They built the recent one that is now the biggest training data center in the world in just 122 days, which is about a quarter of what it took anybody else to build those data centers. And as I mentioned earlier, he's planning to double that data center. And staying on the topic of raising an infrastructure in this field, there's a report about the current cash position of NVIDIA and its future cash positions. So right now, NVIDIA is sitting on 35 billion in cash. They are Projected to have 175 billion in cash by 2027 because they're generating about 15 billion in free cash flow every single quarter. right now, and that is expected to grow. Mean? It means that they are most likely looking for significant M& A activity, which so far has been somewhat cooled down by different regulations by the current administration, and these are expected to either completely go away or be significantly more loose once the new administration comes in. So they've already made significant investments in Databricks and serve robotics and figure and charm therapeutics and other companies that are building stuff around the AI space, but they could go after other kinds of companies like optical networks and robotic companies and drug discovery firms and semiconductors and You name it, they have enough money to buy whatever they want right now. And again, right now, the limitations are a lot, many of the limitations are regulatory, but these may not be there once the new government takes place. So from the topic of infrastructure and the impact of AI on our world, definitely in the U S let's switch gears to the next big topic that I mentioned, which is Microsoft. Microsoft had their Ignite event this week and they made a huge amount of announcements. Many of them are AI related. It's very, very obvious that there's a serious AI focus in Microsoft. The first clear sign for that is that the Microsoft 365 app icon is going to change from its current icon to the co pilot icon, just saying 365 on the bottom of it. So that basically means That the two becomes one. Microsoft 365 becomes Microsoft CoPilot, at least from branding perspective. They also announced Azure AI Foundry, which is a new unified AI application platform for everything Azure. So there's obviously a very between Microsoft, Google, and AWS. On providing AI infrastructure for anybody who runs over their platform, and that's another step in Microsoft side to provide a unified technological front that companies can easily use in order to deploy AI solutions on the Azure cloud environment. And what Microsoft himself said in the blog post is Azure AI Foundry helps bridge the gap between cutting edge AI technologies and practical business applications, empowering organizations to harness the full potential of AI efficiently and effectively. Very clear statement. I wouldn't expect anything else from any of these companies in this current race, but it's very clear that Microsoft is pushing very aggressively in the AI direction after their huge investments in open AI, as well as their internal AI development capabilities. it's very obvious where they are going. And the last big topic for this week is the Department of Justice has just released a huge landmark plan to break Google's search monopoly. So this includes several really important. gigantic tectonic shift level demands. One of them is to sell Chrome. So the Chrome browser is obviously one of the most successful things Google has done in the past two decades, which provides them access and complete monopoly or not complete monopoly, but they control, I don't know, 80 percent of the browser market through that. They control search and many other things, people access and driving them to Google related capabilities. And so they will be forced by the DOJ to sell Chrome. They might need to also divest Android, which is their phone operating system. They'll be banned from signing any exclusive deals on search with any company. They will be restrictions on using data advantages that they have through their ecosystem to develop new AI products. and they'll be forced to fund education campaigns for search engine switching. So they will need to educate people on why they shouldn't use Google and actually use other stuff. All these are obviously extreme measures, but this is what the DOJ is trying to force on Google to do. The final revisions date are due on March 7 on 2025. The remedies are supposed to last 10 years. Unless there's an early termination in five years, if competition substantially increases, or Google's market share drops below 50%. And so this doesn't look good for Google at all. Now, another impact of that in the AI field, and there's probably a lot more, but the first thing that comes to my mind is Google invested 500 million in Anthropic initially. committing to invest up to 1. 5 billion in them over time. And this may take that away because this is another monopolistic approach in that direction. And so this is obviously This whole thing has a huge impact on the digital world as we know it. The only thing that may stop this, the way I see this right now is the new administration. It is very clear that Trump's approach is to reduce regulation and impose as little limitations as possible on American companies to enable them to compete in the world, regardless of that impact in the local market. So there's pros and cons, obviously on both sides of that argument. On one hand, this monopoly is preventing competition in the US. On the other hand, it's keeping a US based company ahead of its competitors around the world. Which one is more important? I'm not a hundred percent sure. And like I said, there's argument on both aspects of this, but it will be very interesting to see if this plan will move forward. Google will be a very different company than it is right now. Now, obviously, Google's response was, as you can expect, very extreme. They call it a radical interventionist agenda. They claim that Chrome sale will endanger user security, which I'm not surprised that they're saying that, whether it's true or not. They're claiming that these moves will reduce the ability of the U. S. to be competitive in the AI development market. And they're arguing that the DOJ completely exceeded the scope of search distribution case, Which is what this is all based on so it will be interesting to see how this plays out but as I mentioned There's still time between now and March for the new administration to throw this into the trash or at least change it one way or Another. Now to put more pressure on Google open AI is planning to release its own browser, so According to an article on the information, which has been a very reliable source to anything of what's happening in this entire field, but specifically in OpenAI, they're claiming that OpenAI are planning to release a new browser that will be integrated with ChatGPT. They're also developing an NL web search product for websites that will allow to have a natural language conversation, either through text or through voice, in order to understand any information that is in any website that will participate. And in order to do that, they hired the founders of Chrome, Ben Goodger and Darren Fisher. So they're very serious in that direction. If OpenAI manages to deploy that successfully, it will be immediately accessed by their 300 million active users that they have right now. And that number is growing every single day. So that puts another pressure on Google. And if that's not enough, Perplexity just launched AI Power Shopping platform with a one click checkout. And before I dive into this topic, I want to connect it to something I've done this week. So in this past week, I've been at a Large U. S known brand in the apparel space doing two days of a training for about 60 people from different levels of leadership from their entire company. It was fascinating two days and through these two days. First of all, we're doing training and we need a hackathon that was extremely successful and at the end of the two days, many of the departments were Actually, all of them were able to create actual tools that they can start using right now that allows them to drive significant efficiencies in many of the processes they're doing today. The use cases included things like data integration for multiple sources, data analysis, internal training, warehouse process mapping. Multiple aspects of marketing and brand reviews and many, many more time savings that these new initiatives are going to create varies between a few hours per one of these tools to a few days for some of these tools. So things that were done in days until two days ago in that organization will now be able to be done in 20 to 30 seconds. And so a huge and amazing success. training. But why is that related to perplexity? It's related to perplexity because one of the things I did is I had a conversation with the leadership team about strategic view of worth of how their world of selling retail apparel will change. And one of the things that I shared with them is that the way we shop is going to change dramatically. Literally everything we know from online shopping and potentially obviously in store shopping will, is going to change. And what Proplexity just released just strengthened my view that a significant part of the shopping that we're going to do is going to be done by agents and not by us. So what is the feature that Proplexity released? Proplexity. com There is the feature that is called buy with pro, which allows you to basically shop around across multiple stores online. It also has a feature that allows you to take a picture of an actual product or an online product, and it will go and search for it across multiple online stores. And it will give you the best options based on your criteria. That could be price, that could be shipping, that could be Many other aspects that you're looking for. And once you will have all these options, it will make it very easy for you to make a decision. What do you want to buy? You can then check out on the perplexity platform with one click and enjoy shipping through them. Now, in order to accelerate the adoption of that, they are adding free enrollment to large retailers onto this platform. And that obviously in return to giving them Increased visibility through that solution. There's a payment integration that again is built into the product itself and complimentary API access to allow search of the products through their platform. Now, this is already built on an integration that they have with Shopify, but I assume this will go Shopify very, very quickly, allowing you to search a lot of other online retailers. What does that mean? It means exactly what I said before. It means that As humans, we will stop going to retail websites and these agents will go and search the web for us and will give us the best options per our needs. Now, if you think about historically, what happened, Amazon hijacked lots and lots of traffic from Google. Many people today, myself included, while we're trying to buy stuff, we don't go to Google and search for things. We go to Amazon directly where we have a lot of stuff and we know it's going to be at our house within, you know, Either a few hours to a couple of days. I have a feeling the same thing is going to happen now, just shifting traffic from Amazon and other large online retailers like Target and Walmart into these AI agent tools that we'll be able to shop across the world. All of these platforms. So even if Amazon and Walmart and target, et cetera, will come up with their own agents that will allow us to do the shopping easier on their platforms. Why would I go to them? If I can just talk to an agent that can shop all of them and find the best options for me. So I think this is the direction this is going. This is obviously a very aggressive and interesting move from perplexity, but that puts even more tension and more pressure on the agent. On Google, because right now Google search has been the source of traffic to many of these websites, driving a lot of revenue to Google itself. And now let's switch to some rapid fire items. There's still a lot of stuff to talk about. I would have loved to dive into more of them, but for the sake of your time, we will keep it short and to the point. In an interview with Demis Assabis, the CEO of Google DeepMind, he shared that he thinks that multiple big breakthroughs are still needed before achieving AGI. That contradicts the thoughts that we heard in the past few weeks from both Sam Altman from OpenAI and Dario Amadei from Anthropic, who claims they can achieve that within the next two years. So Demis disagrees with them. He agrees that multi modal AI models are critical On the path, but he's saying that it's not enough in order to achieve a G I and he still thinks like he has in the past 20 years, and that's what he has been pursuing that the greatest potential or a I lies within accelerating scientific discovery and solving many of the world's problems with a I. It's been very obvious through the years and even in the last few years. how much deep mind is pushing the boundaries in that direction? great piece of news from open AI. They've just released their advanced voice mode into web browser. So previously it was available on mobile devices and on the desktop app. I've been using it a lot, and the only disadvantage was it wasn't available on my browser, which is where I use voice mode. ChatGPT most of the time. So it's now going to be available to pro users and they're claiming later on for free users as well on the web browser. This model is powered by GPT 4. 0 and it's really powerful in allowing you to have a conversation, an actual conversation, while stopping it while it's speaking and going back and forth on any topic that you want. So a great addition from OpenAI to ChatGPT on the browser. And speaking of voice, Microsoft is about to launch a really cool and amazingly powerful feature where you'll be able to train your voice into the team's communication platform, and then it will be able to translate your voice in real time to multiple languages while you're having conversations with international clients or people from your company. This got a lot of excitement. In the room in my conversation that I had this week when I was doing my training because that company has offices in Spanish speaking countries and many Spanish speaking people in the U. S. And so being able to have a conversation in Spanish and English. All at the same time, while every person is speaking their own native language, while everybody else is hearing it in their native language, and vice versa, is extremely powerful. This ability will speak at your voice, just in a different language. in near real time. I find this to be an amazing functionality and if they're claiming it will even capture the sentiment of your voice and translate that to the other language. Now, does that have scary implication? Yes, absolutely. We've already had several huge fraud scams based on these kind of fake calls. So that just makes it even more dangerous because now you'll be able to do it in any language as well. But that being said, from a efficiency and usability, it's a very powerful capability that many companies around the world would be very excited to. And from Microsoft to Google. So Gemini now has the ability to save memories and customize and personalize the data. Your conversations with Gemini, similar to a feature that has been released on OpenAI's chat GPT in April. So it will over time learn more and more about you and we'll be able to provide you more and more relevant information to your needs, similar to OpenAI. And even with more openness, you can look into the info that is saving about you and into these memories and It will allow you to view, edit, and delete any of that information so you can control what's actually knowing about you, whether for personal reasons or just because some of the things it's going to save are not really relevant for your future searches. By the way, I have a mixed feelings about that feature in OpenAI because I do a lot of work for many different clients on my Chachapiti platform. And so in many cases, it remembers stuff that is relevant to a client instead of to me, or it's not relevant to other clients. So I have to go in and delete a lot of stuff. all the time because otherwise you will get some weird answers and you won't understand why. The reason might be is that its memory has saved a piece of knowledge about you that may not be accurate or may not be relevant for some other things. So if you're seeing stuff like that in the behavior of Chachupiti, go back and check its knowledge. And now the same thing will be true for Google Gemini as well. And from Google to the next giant in line, Meta just made a very interesting move when they poached Salesforce AI chief to lead a new business AI decision. Clara Shea, which was Salesforce CEO of AI until this past week, just joined Meta. To lead that group and Meta is aiming to develop AI tools for businesses using Meta's app. So Instagram, Facebook, and WhatsApp. And the goal is to help these companies reach consumers in better ways. They're already showing a lot of statistics that is showing that companies that are using their existing AI capabilities are reaching more relevant clients and achieving higher conversion. And this is obviously going to enhance. Their capabilities in that direction even further. And since we mentioned Salesforce, Salesforce just launched their AI agent testing platform to ensure that the quality of the agent that you are developing is okay before you release them to the wild. It's supposed to be available on December 2024, which is just a week from today. The goal is to provide a safe tasting environment for the new agent development platform that they recently released. So it's an isolated sandbox of environment mirroring the company's real data, allowing you to test the agents. And it's a part of Salesforce overall new agent lifecycle management category that they're working very hard to create. If you've been following the show, you know, I've been talking a lot about the two things that will limit the success of agent one is consistency and the other is trust right now. Many of these agents are not consistent. They will do amazing work once then once every several attempts, it will do something really weird and inconsistent. That doesn't make any sense. And the other problem is trust. Can you give them access to your data? Can you give them access to your system? Will it allow them to act on your behalf on specific different aspects? And these test platforms are literally generating fake use cases with synthetic data by the millions in order to run many tests on the agents very quickly within a safe and secure environment. This obviously makes a lot of sense and it will dramatically increase the implementation of these agents. Through Enterprise in the near future. Now they're not the only company that's obviously developing and providing these kind of capabilities. Sierra has done the same, UiPath has done the same, and obviously there's similar platform from AWS Bedrock and Microsoft Azure that allows you to test in a test environment multiple agents quickly. And from all the US. Giant enterprises to Mistral. We haven't talked about Mistral for a while, but we have talked about them many times previously on the show. Mistral are a French company that has been focusing on developing very powerful open source AI models, and they just released two new models. One is called Pixtral Large with 124 billion parameters. The other one is called Mistral Large which is a little smaller. This is immediately available on their LeChat platform, which is a chat AI platform that is just built on the platform and you can access it right now. And the key updates to their solution is now it has web search with inline citations, similar to what Chachapiti released recently and Perplexity has been doing for a while. A canvas tool for content editing and visualization, similar to what Chachapiti has with canvas and Claude has with artifacts. PDF and image analysis with equation recognition, which is also something that is available on the other platform. AI agent workflows for automations, similar to GPTs. And image generation via Black Forest Labs Flux Pro, so it really aligned many of the top more advanced features of the top platforms is now available through their open source platform. This is going to be available, as I mentioned, on their website for free for now, at least in the initial release and through their API that is available both on Google Cloud and Microsoft Azure. So the open source world is alive and well, both from Meta, as we discussed earlier, as well as from Mistral and a lot of other players. But they are really providing more and more of the top functionality of the closed source platforms, which makes them a very attractive solution that is much cheaper, and that you can run safely on your data if you are running, in this particular case, on Google Cloud or Microsoft Azure. That's it for this week's news. We'll be back on Tuesday with another how to deep dive on a specific AI business use case that will allow you to learn exactly how to deploy that use case. And if you haven't subscribed to this podcast yet, please do If you are subscribed and you're enjoying the content, I would really appreciate it if you would share it with other people that you know. Also pull up your phone right now. And click on the share button and share it with other people that can benefit from this knowledge, which is anybody in business right now. And in addition, if you haven't provided us a rating in a review on either Apple podcast or Spotify, I would really appreciate it if you do it right now. And again, I know later on, you'll forget, or you may think that it's a good idea, but you don't have time to do this right now. Spend 10 seconds, pull up the phone, click on the review button. Give us a five star review or whatever you think we deserve and write something that other people can find this podcast as well. It helps us get the message out. It helps you drive, provide AI literacy for more people and would, and we would really appreciate it if you do that. and until next time, have an awesome weekend and I will see you again on Tuesday.

People on this episode

Head Shot

Isar Meitis

Host