Ep 404: AI News That Matters - November 18th, 2024

This is the Everyday AI show, the everyday podcast where we simplify AI and bring its power to your fingertips. Listen daily for practical advice to boost your career, business and everyday life.

OpenAI is releasing AgentIC AI. Google now has the world's top AI model. But for how long? And Microsoft is hours away from its biggest AI announcements of the year. That's a lot going on and that's just the very tip of the iceberg. You can spend hours every single day trying to keep up with what's happening in the world of AI and how it impacts your company and your career. Or you could just tune into Everyday AI every single Monday as we bring you the AI News that matters. What's going on, y'? All? My name is Jordan Wilson and welcome to Everyday AI. This is for you. It is. Your daily live stream, podcast and free daily newsletter are helping us all not just understand what's happening in the world of AI, but how we can all actually use it to grow our companies and grow your careers. So if that sounds like you, you are a hundred percent in the right place. So if you're brand new here, thank you for tuning in. Make sure if you're tuning in on the podcast, check out the show notes. We always have a lot of great resources in there for you. One of the best resources is our website. If you haven't already, please go to your everydayai.com Sign up for the free daily newsletter. We recap every single day's episode Monday through Friday and you can go in there, click the little Episodes tab and there's literally more than 400 episodes. It is a free generative AI university. Whatever you want to learn, marketing, sales, HR, it's all in there. All right, so thank you for tuning in, y'. All. Let's get straight into the AI news that matters for the week of November 18th. And hey, thank you to our live stream audience tuning in as well. Big Bogey and Michael on YouTube, Brian, Rolando, Marie, Fred, everyone else, thanks for tuning in. All right, let's go get straight into the AI News that matters, y'. All. Well, AI News that really matters is going to be happening at Microsoft Ignite, so I will be there tomorrow in Chicago. So if you are going to be at Microsoft Ignite as well, or maybe your company is, make sure you reach out, say what's up, let's, let's connect and chat AI while we're there. So the keynote kicks off tomorrow and hey, I'll just tell you this, there's big things that are going to be happening at 7:30am Central Standard Time. So make sure you tune in. All right, let's get to one of the biggest pieces of AI news this week. Google's Gemini is now the top AI model in the world. So Google DeepMind's latest AI model, Gemini X exp 1114 so experimental 1114 after November 14th has now surged to the top of the LM arena chatbot arena leaderboard, surpassing OpenAI's GPT4O. So this achievement highlights Google's advancements in AI particularly the model does really well in math and vision tasks, areas where Gemini models traditionally excel. So the LM Arena, Chatbot arena. So formally that was known as the LMSYS Arena. Yeah, a lot of acronyms already three minutes into the show. So the LM Chatbot arena allows AI labs to compare their models in a blind head to head format with users voting on performance without knowing which model model they are evaluating. Y' all remember that like Pepsi versus Coke taste test? That's essentially what the LM arena is for. Chatbots and open AI, I'd say for 95 of the days that this has been up, OpenAI is usually the top model in the world. But Google on November 14, so just a couple of days ago released a new model and it is now the top model. So right now the top five models in the arena are dominated by OpenAI and Google, with Xai's Grok 2 being the first non Google or open AI model on the leaderboard. So despite its success, Gemini 1114 is not yet available in the public app or on the website. So access right now is limited to those with a Google AI Studio account. Also, the Chatbot Arena's evaluation is based on human perceptions of performance and output quality rather than strict data based testing or benchmarks. So how long is this going to last? I'd say not long. Here's why. We just spent a little bit of time over the weekend digging into the Chatbot arena and we also saw that OpenAI is actually testing a new model as well. So it's anonymous. Chatbot model from OpenAI is being blindly tested right now. So I don't know how long this new Google 1114 is going to be the most powerful model in the world. But interestingly enough I also saw Google is testing two follow up models. So they have their mystery Gemini 3 and secret chatbot named models. Yeah, gotta, gotta love the the model naming for the secret models. Like secret Chatbot. Love it. Right? It's like, it's like a bad spy name. But we'll see I would assume, y', all, that OpenAI in the next, I don't know, week or two would probably update its GPT4O model and would presumably be the next top model. So normally, is that the name of a show, right? Normally OpenAI doesn't waste time in being the top model. So it's always slight refreshes, right? It's not like going from, you know, GPT 3.5 to GPT 4 and you know, all of a sudden OpenAI is on the top again. So kind of the trend over the last six months has been these incremental updates, right? So on the OpenAI side, it's actually called GPT4O latest, even though we don't really say that name a lot. Also, what's important to notice here, important to note here about the Google Gemini 1114. If you go on, you know, Gemini, their front end, you're not getting this model, you're only getting this as a backend developer. So this is something that I wish that Google would change a little bit because everyone else, and when I say everyone else, the other kind of two big players here, Anthropic Claude and OpenAI, when they release a new model, they make that new model publicly available on their kind of chatbot side, right? So if you go to chatgpt.com or Claude AI, you'll get access to OpenAI and Anthropic's most powerful models, respectively. Google, not so much. Google really restricts their most powerful models just inside of their AI studio. So, you know, if you're using Google Gemini on the front end as a chatbot, you're generally working with a model that is anywhere from three to eight months old. So I'm not sure what Google's, you know, go to market strategy is there. Maybe they really want to make sure that a model is stable before they roll it out as the de facto model. Not a big fan of it because I think it creates a lot of confusion. All right, our next piece of AI news. Chatbots are outperforming doctors in medical diagnosis studies. So a recent study published in the Journal of the American Medical Association Network Open reveals that ChatGPT4 outperformed human physicians in diagnosing medical conditions from case histories. So this study involved 50 doctors and it showed showed that ChatGPT scored an average of 90% in diagnosing medical cases, while doctors using the chatbot scored 76% and those without it scored 74%. So slight increase there. So this experiment highlights the potential of AI to serve as a valuable tool in medical diagnostics, yet also underscores the challenges in integrating AI into a medical practice. So Dr. Adam Rodman, who helped design the study, noted that doctors often remained confident in their own diagnosis even when the chatbot suggested potentially better alternatives. Yeah, nothing like, you know, second guessing chatbot, that is, you know, has access to human history. So the study suggests that many doctors are not fully exploiting AI's capabilities, often using it more like a search engine rather than leveraging its comprehensive diagnostic abilities. So Dr. Jonathan Chen from Stanford, who was a study author, emphasized that only a few doctors realized they could input entire case histories in the chatbot for more comprehensive answers. The findings point to a need for better training and understanding of AI tools among medical professionals to fully harness their diagnostic potential. So the study used real patient cases that are part of a set of 105 cases used since the 1990s, ensuring that chat GBT had no prior training on them, making the results particularly noteworthy. So this is not the first study that kind of pits AI chatbots against doctors. There was a pretty famous study about a year and a half ago that actually showed chat chat GBT was more empathetic than doctors and had better bedside manner in kind of a blind test. So this, this, I, I do think this is going to be one of those studies that commonly gets referred to in the coming months. Right. When we talk about chat GPT or AI models capabilities. I don't know, live stream audience, would you want AI as a doctor? Right. I know it's a weird, a weird concept, but I do think that is where we're heading. Just a lot more reliance on large language models in the medical setting. I, I don't care. I would love to hand a chatbot over all of my medical history if it meant that I didn't have to wait like nine months, you know, for a stuffy nose. Right. The medical system here in the US is broken and I am very excited for AI to disrupt that system. All right, our next piece of AI news, Nvidia earnings are coming out on Wednesday and there is going to be a lot of eyeballs on that report. So Nvidia is set to report its quarter three earnings on Wednesday, providing a crucial update on the AI market's momentum. So, as the world's largest publicly traded company by market cap, Nvidia has become a barometer for AI growth due to its significant advances in AI technology and applications. So the company's stock has soared nearly 190% year to date, driven by the surging demand for AI Technologies making it the largest publicly traded company by market cap in the entire world. So, yeah, technically the most valuable company in the world. And y', all, I told you two years ago when Nvidia was kind of like, no one really knew about it. I'm like, Nvidia is going to be the biggest company in the world. No one listen to me. Yeah, you got to start listening, y'. All. So investors are eager to see if Nvidia will exceed expectations and raise its quarter four revenue outlook, projected at $37 billion. So despite past strong performances, Nvidia's shares actually fell 6% after quarter two earnings, possibly due to profit taking or unmet high investor expectations. Yeah, expectations are through the roof, especially when you hear about companies like, you know, as an example, Xai ordering 100,000 GPU chips all at once. You know, expectations are pretty high for Nvidia to deliver. So CEO Jensen Huang may provide updates on the next generation Blackwell AI chips, which are expected to significantly contribute to future revenues. The demand for Blackwell chips is already outpacing supply, indicating robust future sales potential. So the company's performance is seen as a reflection of the expanding influence of AI across various sectors beyond just technology impacting industries like healthcare, automotive and finance. So yeah, I talk about this pretty much all the time, y'. All. The reason why we talk about chips a lot and we cover GPUs and kind of the political aspect of this is because GPUs mainly from Nvidia power. All AI. Right. All the large language models. Right. For the most part, any AI tool you use, it is a high likelihood they are using Nvidia's chips. So, so much of our day to day work and how the future of work is changing is actually impacted by Nvidia chips. All right, so that's why it's pretty important. So Fred's asking here, how will increased tariffs impact Nvidia? Yeah, we've seen some reports from the Trump administration, the incoming Trump administration here in the US of tariffs anywhere between 10 to 20% and I believe up to 60% on Chinese imports. So yeah, we are probably going to be seeing, if these tariffs do come to fruition, we're going to be seeing probably higher costs on the hardware and software side. That's not my personal opinion, that's just, you know, third party reports that are looking at this. So we will see. All right, our next piece of AI news, Chat GBT has rolled out some noteworthy desktop updates on both Mac and windows. So OpenAI has announced a new feature for the chat GPT desktop app on Mac os, allowing it to read code directly from several developer focused coding apps like VS Code, Xcode Text, Edit Terminal and iTerm2. So this development eliminates the need for developers to copy and paste code into ChatGPT, streamlining the coding process by sending code sections directly to the Chatbot as context. So unlike other AI coding tools such as GitHub Copilot, ChatGPT cannot yet write code directly into these apps, but it can assist by providing code suggestions. So I did test this out last week as it came out. So there's also some confusion out there. People are thinking that this is the new operator kind of agentic AI from OpenAI. It is not. All this does so on the Chat GPT Mac app. So you need to update it to the newest app. There's essentially kind of like how you would mention a GPT. You can essentially read different files from those programs, right? So as an example, let's say you are using VS code or something like that to code. ChatGPT can simply read the file that you are working on, but it cannot write. So essentially what this does, it's a great user kind of UI ux. So user interface, user experience update, you know, it will save a little bit of time. But right now I think this is just more of an incremental feature, right? And hopefully creating a stronger bridge between the ChatGPT desktop app and how you work with native Mac apps. So we do not yet have that same capability on the Windows side. However, Windows users were not left out of this because also last week Chat GPT announced that free users can now download and use the ChatGPT app for Windows. So before this was limited to paid users only. So if you are a free Chat GPT user, you can now download the desktop app for Windows. All right, more Open AI news. Let's keep it going. So according to reports from Bloomberg, OpenAI has proposed the North American AI alliance to compete with China. So OpenAI is urging US legislators and regulators and its allies to collaborate. To collaborate on AI infrastructure to maintain a competitive edge over China, according to Bloomberg reports. So the company suggests forming a North American compact for AI to streamline access to talent, financing and supply chains. So OpenAI envisions this compact expanding globally to include US allies and partners such as countries in the Middle East. The proposal was unveiled at an event by the center for Strategic and International Studies, highlighting its significance. So OpenAI's policy blueprint emphasizes the need for the US to support energy infrastructure projects by committing to purchase power so the company recommends establishing AI economic zones to expedite permitting processes and revitalize nuclear reactors. Yeah, one, one big thing. We saw another report that we covered in the newsletter last week that said up to 40% of AI data centers are going to run into power problems in 2025. Not good. So expanding nuclear energy capacity by utilizing compact reactors from the US Navy was also proposed. So OpenAI views AI as an opportunity to re industrialize the US and foster broad based economic growth. The initiative is seen as crucial for national security, promoting democratic values in the AI development against China's influence. So OpenAI has sought funding from investors in the Middle east to support China chip energy and data center expansion. So OpenAI's CEO Sam Altman has reportedly engaged with US officials to garner support for the infrastructure plan. So the proposal coincides with an upcoming change in the US administration with President elect Donald Trump acknowledging the need for expanded energy capacity. So this one is pretty interesting here, right? It almost seems like on. It almost seems like on the AI side at least according to reports that OpenAI is almost trying to establish a NATO esque partnership among ally nations to the United States just for AI infrastructure and development specifically to compete with China. So I do think that China was maybe a little late to the game, but I think if you look at just over the last six to nine months, China, aside from the U.S. i think has been the leader in AI development. And it's not even close in terms of first it's like 1A, 1B and there's really no one else in second or third right now. It is at least as of recently it really is just the US and China. So a pretty strategic move here. According to reports from Bloomberg from Open AI trying to take a driver's seat in terms of kind of global dominance on the AI side scene. So it is a very important, I'd say piece of the puzzle to keep a look at, right? I think people sometimes devalue AI's role in the world economy. It is everything right now, right? AI is being used in the forefront of military and defense, it's being used in transportation supply chains. Literally every aspect of the world is becoming maybe overly reliant on artificial intelligence, on GPUs, on the ability to generate power, right? So much of world economies right now is heavily reliant and quickly becoming potentially over reliant on AI and kind of related industries. So I do think it is important to see what happens in terms of partnerships on a global scale. I don't think people are paying to attention, too much attention to it, but I think that we should so pretty big news there coming from Bloomberg. And hey, as a reminder y', all, this is always in our newsletter. So if you're listening to this on, maybe you're on the treadmill or, or walking your dog, something like that, and you're like, what happened there? It's always going to be in the newsletter, so don't worry about that. Are you still running in circles trying to figure out how to actually grow your business with AI? Maybe your company has been tinkering with large language models for a year or more, but can't really get traction to find ROI on Genai. Hey, this is Jordan Wilson, host of this very podcast. Companies like Adobe, Microsoft and Nvidia have partnered with us because they trust our expertise in educating the masses around generative AI to get ahead. And some of the most innovative companies in the country hire us to help with their AI strategy and to train hundreds of their employees on how to use Gen AI. So whether you're looking for ChatGPT training for thousands or just need help building your front end AI strategy, you can partner with us too. Just like some of the biggest companies in the world do. Go to your everydayai.com partner to get in contact with our team. Or you can just click on the partner section of our website will help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on Gen AI. All right, our next piece of AI news Some big news from Google. So Google's Gemini AI is reportedly set up for a major upgrade with some conversational capabilities. This sounds like a small feature, but stick with stick with me. I'm going to tell you why this is actually pretty big. So Google's Gemini AI is poised for a significant upgrade, potentially enhancing its ability to interact with various media types like files, photos and recordings. So this is with the Gemini Live kind of Live interactive voice agent. So this development mirrors the capabilities of OpenAI's ChatGPT, particularly a feature missing in its advanced voice mode. So according to Android Authority, a recent beta version of the Google Gemini app revealed code suggesting that Gemini Live might soon be able to enable discussions about user uploaded attachments. So although the feature is not yet active, the code hints at a user experience where questions can be asked and insights can be gained directly from files. Here's why this is important, because also in the last couple of days, Google has released Gemini as a dedicated app on the iPhone, which includes access to its Gemini Live voice assistant. So this is important because the combination of having both Gemini on the iPhone, which has access to the neural Gemini Live Live agent, and the ability for Gemini Live to potentially read your files would put it ahead of all other voice assistants. So I'd say right now there's really only at least here in the US I'd say that by far OpenAI's advanced voice mode is the best live AI voice assistant. Right. And presumably we'll hopefully see some of these roll out to companies or to services like, like Amazon's Alexa, like Apple's Siri, et cetera. But right now it is really just OpenAI's advanced voice mode and Google's Gemini Live that I think are pretty far above everyone else. But the downside, at least right now is these very, you know, neural, low latency, very human like sounding AI voice assistants powered by large language models do not and cannot access your actual data, which makes them a little less useful. Right. So we've done a couple of shows here as an example on Advice advanced mode from OpenAI and ChatGPT. The downside though is you can't use all of the other tools so it cannot access the web. You cannot even for Advanced Voice Mode on Chat GPT's app, which is right now only available on the iPhone. Actually, no, I do believe it's, it was rolled out. Yeah. During my, my little vacation that I had celebrating my anniversary, it was rolled out on desktop as well. But the downside is you cannot upload files to OpenAI's advanced voice mode and you actually can't even type. As soon as you type, you can no longer use Advanced Voice Mode. So this is some. Actually, like I said, it sounds like a small little feature, potential feature from Google Gemini and Gemini Live, but it's actually pretty big because this would, if it does actually roll out, this would mark the first very smart large language model powered AI voice assistant that can access your data, which I think would really change how we work. Right. Think of as an example, if you have a long commute or if you're taking a long walk, talk, things like Advanced Voice Mode and Google Gemini Live are great to talk to. However, it's not super useful, right? Because if you do want to really gain some great insights when it comes to your business, when it comes to something that you're personally working on, you have to spend a lot of time verbalizing and talking all of that information into the context window, which is, I mean redundant number one. Number two, it's extremely time consuming. So the ability to upload Files and use Google Gemini Live and their new iPhone app that brings Gemini and Gemini Live to iPhone users. Could be a huge announcement. All right, last but def. So, so not necessarily Michael, great question from YouTube asking is this kind of like Notebook LM? Not necessarily. So the big thing right now with Gemini Live in advanced voice mode is you can can talk to these agents, right? Notebook lm I think it is the product of the year for 2024. Unless OpenAI drops Sora or drops, you know, GPT5 Orion it I, I, I still think Notebook LM right now is the product of the year for 2024. So Notebook LM is by Google. It is a grounded state of the art model that has built in retrieval, augmented generation. It is wild that I'm saying those words and that is available. However, with Notebook LM you can't talk to it, right? Right. You can type to it, you can chat with all of your documents and it can also you can generate podcasts. Right? But you can't talk to it. So I think this is where once you can upload your files to a neural AI live agent and talk to it, that's where it's like y' all can the consulting game is going to completely change. So not necessarily like Notebook LM slightly different. All right, our last piece of AI news for the week and I'd say this is probably the biggest news. So According to reports, OpenAI is working on an agentic AI system called Operator. So OpenAI is preparing to release a new AI tool codenamed Operator in January 2025, designed to automate tasks with minimal human intervention, according to Bloomberg reports. So Operator, like I said, is an AI AI agent that acts like a personal assistant, capable of performing multi step processes and executing actions on behalf of users, distinguishing it from traditional AI chatbots. So the introduction of Operator signifies a major shift in how businesses could operate, potentially transforming daily workflows by automating complex and labor intensive tasks. So the tool is expected to be released as a research preview first and also through OpenAI's API for developers. So it will, according to reports, be released in kind of those two tiers first, but at least right not at least not now. Kind of like on the front end of if you go to chatgpt.com or using the ChatGPT app. According to reports, it will not be released like that, but first as a research preview and through the OpenAI's API. So this development aligns with OpenAI's long term focus on agentic AI, a priority highlighted by CEO Sam Altman and reflects the company's ongoing commitment to advancing AI capabilities. So competitors like Anthropic, Google and Microsoft are also making strides in agentic AI with products that we've talked about here on the show in the last couple of weeks a lot. Anthropic's computer use, which is live now Google's Jarvis, which was a Chrome extension that Google launched and then they kind of took away but will reportedly be released in December and Microsoft's autonomous AI agents in Copilot Studio, which according to reports will be publicly released in a beta tomorrow, November 19th. So essentially all the big, I mean those are the big four right there, right? Anthropic, OpenAI, Google and Microsoft, all in the fourth quarter of 2024 are either going to be releasing Agentic AI or reports are coming out that they will be releasing it. So essentially we're seeing from October to now reported reportedly January. So in a four month span we will likely see agentic AI from the four, their biggest players in AI. So the introduction of AI agents like Operator could accelerate the shift towards more automated work environments, emphasizing the need for businesses to invest in AI readiness and redefine job roles. I cannot underestimate or sorry, I cannot over state how important and impactful this is. I think we been kind of, you know, slowly walking our way through large language model implementation over the last two years. This changes everything. This changes how we work, it changes what jobs look like, it changes how humans in the human role in traditional knowledge work. So as AI becomes more integrated into business processes, issues of safety, governance and data compliance are expected to become even more more critical, requiring organizations to establish robust guardrails and human in the loop systems. All right y', all, that was a lot. And just one more reminder, Microsoft ignite tomorrow. Let's just say there's going to be some big announcements and you are going to want to make sure you join the show tomorrow. And you should probably join us live at 7:30am Central Standard Time. I'll just say that I can't say a whole lot more. But as a recap, here is the AI news that matters for the week of November 18th. So first we have Google's new 1114 Gemini 1114 topping the AI arena leaderboards as the world's most powerful AI model. For now, I do expect OpenAI to release an updated version of the GPT4O that would probably take back that top spot. Next, according to a new study, Chat GPT is outperforming outperforming doctors in medical diagnosis. Next Nvidia is poised for strong quarter three earnings that are expected this Wednesday. Next Chat GBT has rolled out some pretty noteworthy desktop updates for both Mac users and for Window users. Next According to reports, OpenAI has proposed a North American AI alliance to compete with China. Next Google's Gemini AI is set for a potential major upgrade with conversational capabilities and giving users the available the option to upload files while using Gemini Live and the new iPhone release. And then, last but not least, reports are saying that OpenAI's new AgentIC AI system called Operator, will be released in January. All right, y', all, that was a lot. I hope that was helpful. If this was helpful, please make sure if you haven't already, go to your everydayai.com Sign up for the free daily newsletter. Hey, and while while you're here, if you're listening, whether it's on the podcast or on the live stream, please click that repost button. Go ahead tag someone in the comments that needs to hear this. If you're on the podcast, please follow or subscribe to the Everyday AI show. Whether you're listening on Apple or Spotify or somewhere else. And if you could leave us a rating, we would super appreciate it. We spend so much time every single day. Yeah, the show is Monday to Friday. Maybe you only listen on Monday for the AI news that matters, but we do this every single day, bringing you the brightest minds in AI from all over the world, every single sector, and giving you direct access to them. So. So if you're a normal podcast listener, I suggest you try to schedule it into your day to come to the live stream. It's great to be able to ask questions of some of the world's leading AI experts and get answers in real time. All right, I hope this was helpful. Make sure to shout out, shout out. Let me know if you're going to be at Microsoft Ignite in Chicago this week starting tomorrow. My birthday. What a great birthday present, right? You know, the keynote and the kickoff off of Microsoft Ignite. So make sure if you want to, you know, connect chat AI, maybe tell me a little bit about what your company is doing in the AI space. Make sure to reach out to me. We'll set up a little meeting. Thank you for tuning in. Hope to see you back tomorrow and every day for more Everyday AI. Thanks y'. All.

And that's a wrap for today's edition of Everyday AI. Thanks for joining us. If you enjoyed this episode, please subscribe and leave us a rating. It helps keep us going for a little more AI magic. Visit your everyday AI.com and sign up to our daily newsletter so you don't get left behind. Go break some barriers and we'll see you next time.