EP 353: AI News That Matters - September 9, 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.

We are hours away from Apple's new iPhone announcements. Their first AI, first iPhone is Salesforce getting away from SA sales. And how did a three month old startup raise $1 billion to focus on safe AI? We're to be answering all of those questions today and more on Everyday AI. What's going on, y'? All? My name is Jordan Wilson and I'm the host and Everyday AI is for you. It's your daily livestream, podcast and free daily newsletter helping us all learn and leverage generative AI to grow our companies and to grow our careers. Before we dive in a little bit, I'm excited about bringing you the AI news that matters. But first I wanted to shout out our sponsor for today, the Work the Work Lab podcast from Microsoft. So why should you listen to the Work Lab podcast from Microsoft? Because it's made for leaders who know they must adapt to stay ahead. Work Lab is the place to find real life lessons and actionable insights to prepare you for the next phase of AI at work. That's W O R K L A B, no spaces available wherever you get your podcasts. All right, more on Work Work Lab here in a bit. But let's talk about the AI news that matters. If you are new here joining on the podcast. Thank you for tuning in. We do this almost every single Monday because you can spend literally countless hours trying to keep up with AI or you can just tune in here every Monday on the podcast on the live stream. So thank you for all of our live audience joining us. And as a reminder, if you haven't ready, go to your everydayai.com Sign up for the free daily newsletter. Yeah, it's a podcast, a live stream, unedited, unscripted and then also we put this out in our daily newsletter, so make sure you go check that out. So thank you for our live stream. Audience tuning in. Madan, Jay, Brian, Tara, Fred, Joe, everyone else appreciate y' all tuning in. So let's jump straight into it and talk about the AI news that matters for for this week. It's a big one, y'. All. We got a lot going on. So Apple is preparing for its Glow Time event with the new iPhone powered by on device AI. So in like three and a half hours, Apple is set to unveil a range of new products at its Glow Time announcement today, which should include the new iPhone 16 lineup. New Apple watches and new updates to AirPods. So the Apple event later today, like I said, just in a few hours, is significant for those of us who follow all things AI, even if you aren't interested in Apple at all, as many of the hardware announcements are tied to Apple's generative AI initiative, called Apple Intelligence. So the new iPhone 16 series is expected to feature edge to edge displays, with the Pro Max model potentially increasing its screen size further from 6.69 to 6.86 inches. Yeah, so apparently that's a headline, but there are also speculations regarding new Mac computers, But announcements for M4Macs may be postponed until November due to supply chain issues. All right, so let's talk about why this is pretty noteworthy here. Well, part of it is because there's been a recent trend of more minor upgrades in the iPhone models and that's led to a slump in sales, making the iPhone 16 innovative AI features even more critical for the company's future. The iPhone 16 is anticipated to be the first model specifically designed for advanced AI features which could enhance user experience significantly. So Apple's new iPhone 16 is expected to leverage both Edge AI or on device AI with a small language model approach focused on enhancing user experience while maintaining data privacy and then also connecting to open AI's GPT4O for more complex AI requirements. New features for Siri, including an updated interface and better app integration will be showcased emphasizing the role of AI in daily tasks. All right, so we will have this in our newsletter today. So it might go out an hour or so late, but just so you can keep up with everything that matters. And, and this one is pretty big and I've been saying this for more than a year. Apple is late to the party here when it comes to generative AI, but we will see if they can do it correctly, if they can implement Generative A AI correctly. So yeah, that is something at their WWDC conference back in June, not a lot of people realize is that Apple is actually working with two separate kind of techniques under the hood for its AI. So first is a small language model that is actually Apple built in house and they didn't even really talk about that at their keynote. It was just kind of a footnote which I found interesting. But it does seem like a majority of queries that users may ask about are actually just going to be tapping into OpenAI's Chat GPT app. So pretty interesting kind of two tiered approach here from Apple, kind of being the last major player to bring edge AI to a device. But this is going to be a big one for, I think, for the future of how we all interface with AI. Yes, it will still be available, I believe, on the last two most powerful iPhones, so the iPhone 15 Pro and the iPhone 15 Pro Max. But for everyone else, you will have to, if you do want this Apple intelligence, you will have to be upgrading to the iPhone 16, which will be announced here in a few hours. So Apple putting a pretty big bet here that essentially this AI and Apple intelligence is going to be enough to kind of force people to upgrade to a new device when the last couple of updates have been kind of incremental. Right. I know a lot of people that, you know, have hung onto the same iPhone for, you know, three, four, five years just because the updates haven't been super significant. All right, so. Hey, hey, Penelope. Yes, thank you. I am feeling better last. Last. Last Monday. I was not feeling that great for the AI news that matters. All right, let's. Here's. Here's a big one that I don't think people are talking about. It's technically kind of small, but I think it's pretty big. So Google, Google has introduced finally, it's Ask Photos AI feature to enhance image search using AI. So Google is finally rolling out its new Ask Photo Photos feature, which allows users to interact with their photo libraries in innovative ways, marking a significant advancement in the use of AI for personal photo management. So this development is particularly newsworthy as it showcases how AI can streamline the process of finding and organizing personal memories. So the Ask Photos feature from Google enables Google Photos users to ask specific questions about their photos, such as, where did we camp last time we went to Yosemite? Or, you know, show me pictures of my son as he's learned to walk over the last couple of months. And then it's going to utilize Google's Gemini AI models. The photo app will then respond based on the content of the images and provide relevant pictures related to the inquiry. So users can also request summaries of, as an example, recent vacations, or if you attended a conference and took a bunch of photos, or select the best family family photos for shared albums, enhancing the overall user experience. So Google did announce and roll out a wait list available for users to sign up and access this feature on their website, indicating a phased rollout. So in addition to Ask Photos, Google is improving its classic search functionality, allowing searches using more natural language, like Alice and me laughing or kayaking on a lake surrounded by mountains. Right. So essentially just conversational search. So pretty. Like. Like I Said this isn't a, I don't think this grabbed a lot of headlines, but I think it's a pretty, pretty big announcement here from Google that they are finally rolling this out after teasing it back at their developer conference. What was that back in May? All right, so yeah, Google, I think Google has this tendency to announce things and then like, ah, four or five, six months later we might be seeing it roll out. Uh, so here's one in the Ask Google or Ask photos that we're finally getting from, from Google. And hey, like I said, just right there, big, pretty big news from both Apple and Google. Curious live stream audience and podcast peeps. Shoot. Shoot me an email. I put my, my LinkedIn in there. Put, put the email as well. Let me know if you're excited about any of these announcements. All right, next one. This one I think is interesting. So Replit has launched an AI agent to revolutionize software developments. Yeah, hey, even for non coders, even if you're not a developer, listen up, because I think this is changing things. Right? So Replit has introduced an AI engine, or, sorry, an AI agent. Essentially it's an engine that can autonomously build entire applications, marking a pretty significant shift here in software development. So the new AI agent operates with greater independence than traditional coding assistance, making decisions and executing tasks without constant user input. Yeah, it's pretty impressive. You just tell it, hey, build me, you know, build me an application that does this. And it just kind of does it, right, all of it. So it learns and adapts over time, improving its performance based on feedback and new information, which could dramatically change how developers approach their products. Also, Replit AI agents can handle the entire software development lifecycle, from writing code to managing deployments, thereby streamlining the entire process. Right. I think a lot of cases this can take it from multiple days, multiple hours, to at least getting something usable out in like a couple of minutes or at least less than an hour, you know, depending on kind of your current skill set in the area. So Replit has leveraged a vast amount of data from its millions of users to train the new AI agent, making it more effective in completing complex tasks autonomously. So one of the standout features is the Bounty service, where users can request software projects, providing valuable real world data for training AI agents. So Replit is obviously not alone in this space. Competitors like Microsoft's GitHub, Copilot and startups like Cognition Cursor has one that's been blowing up online lately in Magic. They're all also developing similar tools. So as AI agents are taking over some of these more routine tasks. Software developers will need to adapt by focusing on higher level creative problem solving skills. So right now the Replit AI Agent is currently in beta for Core and Teams subscribers. So those are their paid plans suggesting that early adopters will and paid users will have the first opportunity to explore these new capabilities. So according to Replit CEO Amjad Massad, this innovation is not about replacing developers, but enhancing their creative potential in making software development more expensive, accessible for everyone. You know, I'm curious out there, you know, and when I created everyday AI, one of the reasons I set it up the way I did is because I wanted this to be a place for non technical people to learn about generative AI. But now here we are a year and a half after I started this thing and now it's like even people without a lot of tech skills can start to do these very complex tasks, right? If you would have told me even three years ago that you could develop an entire app or you know, an entire web project with a simple prompt and you could get it live, I would have been like, that's pretty wild. So, you know, I'm curious, would you all like to see a little bit more on this technical side, right? Would you like to see whether it's tutorials or dedicated shows, you know, on these AI agents, software development, Replit cursor, etc. Like I said, maybe if you are brand new, if you're not a technical person, this might sound super technical, but I promise you it's actually not. Right? Yeah, I'm a little bit of a geek, but I'm not someone that hand codes or hand develops, although I can, you know, I understand the basics of, you know, JavaScript, Python, CSS, HTML, all of those things. But you know, I'm curious to our podcast audience and live stream audience, hit me up with a message, let me know if we should be covering this a little bit more. Michael says yes, and someone from 1A Lee here says definitely more AI coding. All right, so yeah, let me know if you guys want to see more of that. I think, I think it should be, if I'm being honest, I. I think it's pretty interesting and it changes things. So speaking of changing things, my gosh. So Safe Super Intelligence Inc. Has secured $1 billion to develop advanced AI systems. So Safe Super Intelligence Inc. Or SSI, is obviously founded by a very notable figure in AI, which is former Open AI co founder Ilya Sutskever. So they've raised a billion dollars to advance their Mission of creating safe artificial intelligence that exceeds human capabilities. Yeah, SSI just as wild. So this significant funding illustrates the ongoing interest from investors in foundational AI research, even as the market shows signs of cooling. Reportedly, I don't think so anyway. So SSI was co founded, like I said, by Ilyatskiver, former co founder of OpenAI, and it has quickly gained a valuation of approximately $5 billion according to anonymous sources. So yes, this was some exclusive reporting from Reuters. So SSI plans to utilize the funds to enhance its computing power and attract top tier talent, aiming to build a small trusted team of researchers and engineers based in Palo Alto and Tel Aviv. So major investors this round include the usual who's who of VCs so Sequoia Capital, DST Global, etc. So the founders have emphasized the importance of surrounding the company with investors who align with their mission of ensuring the development of safe super intelligence while focusing on research and development before launching products. So this initiative from SSI comes at a time when AI safety has really become a critical concern, with fears that unregulated AI could potentially pose risks to humanity, prompting legislative efforts in California to impose safety regulations on tech companies. Yeah, we've talked about that the last last week when we went over the AI news that matters kind of that new California AI bill. So Set Skiver, who previously served as OpenAI's chief scientist and co founder, noted that SSI's approach to scaling AI will differ from his former employer, hinting at more innovative strategies without disclosing specifics. So like I talked about, the company right now is really focused on two things. Hiring some of the smartest people in the world and compute, Right. So it seems like anytime a company raises, you know, I don't know, ten million, a hundred million, or in this case a billion dollars, it is a race for two things. Who can we kind of lure away from? Who are some of the smartest people in the world that we can lure away from other companies? And how many GPUs from Nvidia can we get our hands on? Right? It seems like anytime a company raises, you know, you know, hundreds of millions or a billion dollars, that's what it's about. But hey, those are the, those are the resources that are scarce right now, right? So right now compute is almost, is becoming almost too cheap to meter, right? But you still need the GPUs to create that. So. So this funding round right now in the establishment of SSI highlight the ongoing demand for talent and resources in the AI sector, which could influence job Seekers and startups looking to make their mark in the industry. Yeah, the ssi, y', all, you gotta keep an eye on it. I'm telling you, I don't know what they're gonna be coming out with, but even some of the talent that they've attracted so far, it is worth keeping an eye on. All right. Hey, worth keeping an eye on. Gotta shout them out again. Our sponsor for today's podcast, Worklab. All right, so let me tell you a little bit more. So the Worklab podcast from Microsoft is made for leaders who want to understand how work is changing. So effective leaders adapt, they stand ahead of or they stay ahead of trends. They embrace any edge that they can get. They also know that AI powered organizations, they also know that AI powered organizations will be better at spotting opportunities, creating new products and business models and maximizing value. So for real world lessons and actionable insights to help you stay ahead, check out the Work Lab podcast. The new season launches September 12th. That's W O R K L A B, no spaces available wherever you get your podcasts. All right, we're going to be talking a lot more more about Worklab in the coming weeks, so make sure if you haven't already to check them out. All right, back to the AI news. This one, we talked about it last week, but Anthropic has launched Claude Enterprise with GitHub integration kind of to enhance programming tasks. So Anthropic has officially launched Claude Enterprise, a business focused tool designed to offer enhanced security and administrative controls, making it a noteworthy development in the AI landscape. So the new cloud, Cloud Enterprise is featuring a native GitHub integration. So talking about companies that you know, have a lot of coding and development projects, this will be good. Although that's not being rolled out yet, it's being rolled out later this fall. So the new tool from, if you want to call it that, I guess the new software, right, Claude Enterprise is, is built with more robust security measures including audit logs and role based access with fine grained permissions, ensuring compliance and security for organizations. Pricing for Claude Enterprise is tailored to each organization. I'm not a fan of that. Allowing companies to work with Anthropic sales team to customize their plans based on unique use cases. So the launch of Claude Enterprise is part of Anthropic strategy to attract larger clients and tap into those enterprise budgets, a competitive necessity in the AI market. So this release coincides with a shift in the industry from initial excitement about generative AI towards a focus on return on investment and risk mitigation. So also Anthropic's collaboration with Amazon, right, that's a big piece there which invested $4 billion and the company aims to provide users with a secure environment through a, through Amazon's AI marketplace bedrock. So a couple things and you know we talked about this last week on the show. We went over kind of like what's new and does this matter? And I'll say this personally, I'm not a huge fan of how Anthropic rolled out Enterprise number one. I just talked about it there. There is reportedly no set pricing tier. You have to sign up and apparently pricing is based on how your company is going to use it. So like I talked about last week, not a fan of that. I would love if they set tiers. Right. So you know, yeah, you might not want to have a blank, you know, 30 or 50amonth per user if you have extremely high usage. So I would have liked to see from Anthropic more of a tiered approach. Right. So yes, if you're, you know, company does need to use this way more, you know, you might have to pay more but I would have at least liked to see them be a little more transparent with their pricing and set actual tiers or limits. So not a fan of that. Also talked about this last week and you know, you can go back and listen to that show right now. I would not advise companies, at least as it stands today for the most part to use Claude Enterprise unless you are a development heavy company or unless this is just for developers. One of the main reasons, well the pro is the GitHub integration once that does fully launch but it apparently will be slowly rolling out this fall. But the reason why I cannot really recommend this to anyone is because right now CLAUDE is the only major large language model that does not have access to real time information. So like we talked about Chat GPT does Microsoft Copilot does Google Gemini does Anthropic? Claude does not. And that is key because employees. Right. Even though yes, there are great ways within Anthropic to bring in your data. It is all static data. Right. So you know how quickly things in your business change, how quickly your the market, the niche that you compete in changes without access to up to date real time data. Could not recommend Anthropic Claude Enterprise as it is now to customers. I couldn't. Right. So yeah, people, you know, pay us all the time to consult and I'd love to be able to recommend that but you can't right now. So that's something to keep an Eye on. Well, kind of speaking of Claude Anthropic or Anthropic Claude. So a little bit more on that. But right now, Salesforce is making a bold move into AI with Agent Force, and that is signaling a new era for the popular CRM maker. So Salesforce CEO Mark Benoist emphasized the company's shift toward generative AI, including a new product called Agent Forced, a platform designed to create autonomous AI agents for businesses. So this pivot comes as Salesforce aims to redefine customer relationships, CRMs in the competitive landscape of AI technology. Salesforce will be launching Agent Force at the upcoming Dreamforce conference next week in San Francisco, allowing users to build AI powered agents that can automate customer service tasks and make decisions. So the platform represents a significant advancement beyond traditional AI chatbots, positioning Salesforce to. To better compete against rivals like the leaders in the space, Microsoft and SAP. So Salesforce CEO described this moment as a quote, unquote, hard pivot for the company, likening it to starting over as an AI agent startup after 25 years in the CRM business. So he's invoked this concept of show sin or, or the beginner's mind to encourage innovation and adaptability within the company as it embraces new technology. So right now, Salesforce annual revenue stands at $38 billion and the company employs 75,000 people, highlighting its significant presence in the CRM space. Yeah, it is the most significant. So Salesforce has also, like I just talked about, partnered with Anthropic to enhance its new Einstein AI capabilities by integrating the Claude models. So that's Claude3.5 Sonnet and then Claude3 Opus and Claude3 Haiku, available through Amazon Bedrock, allowing customers to customize AI powered applications across Einstein AI. So yeah, Salesforce's AI agent Einstein. So despite strong second quarter earnings from Salesforce, analysts remain skeptical about Salesforce's ability to maintain its competitive edge in AI, noting that the company must deliver on its promises. So this is a pretty bold transition by Salesforce and it's probably going to impact a lot of professionals and businesses because it's highlighting they are doing this hard shift. So it should be interesting because so many companies have really built the backbone of their business on Salesforce's traditional CRM products. So it's going to be interesting to see in the coming quarters and years how much ongoing attention there is to its kind of traditional core business versus this reported hard shift. Right. So literally the CEO talking about now a hard pivot and, you know, focusing almost like a startup, even though they are one of the largest companies in the world. So yeah, we'll be, you know, in the newsletter, we'll be covering Dreamforce next week. All right, so last but not least, speaking of big companies, big price hikes. Question mark. So, yeah, OpenAI is considering a high priced subscription for upcoming AI models. So according to reports from the Information, OpenAI is reportedly exploring the possibility of offering high priced subscriptions for its next generation language models, which could have significant implications for users and the competitive landscape of AI technology. So internal discussions, according to the information at OpenAI have considered subscription prices reaching up to $2,000 per month for advanced models in the future like Strawberry and Orion. And hey, thank you to our audience, you know what this is? It's live, unedited, unscripted. So you know, a lot of times I mispronounce something. So thank you all for telling me I was mispronouncing Orion. Right. So although these two thousand dollar a month figures are just reports and obviously still under consideration, there is skepticism about whether the final pricing will be that high. So the potential pricing strategy reflects the growing demand for sophisticated price AI tools as companies like OpenAI, Anthropic, Amazon, Microsoft, everyone seek to enhance their capabilities and services and to get a return on the investments they're making in these large language models and generative AI systems. So competitors such as Microsoft, Google and Anthropic are also likely monitoring these developments closely as pricing decisions could influence their own offerings in the future. Also, the willingness of customers to pay for these AI services will play a crucial role in shaping the future of AI subscriptions and accessibility. So according to sources familiar with the matter, the outcome of these discussions could redefine how businesses and individuals interact with advanced AI technologies in the near future. So are we actually going to see these kind of, if it is called Orion GPT5, are we actually going to see a $2,000 a month model? I would say probably not. Right? I think companies are very intentional about information that they leak to, to the media, to journalists. Right. I was a former journalist for seven years and you know, you had to take with a grain of salt, you know, what these big tech companies are leaking out there. They probably wanted to get some sticker shock out there, if I'm being honest. Will there be a higher tiered model once we see these much more powerful, whether it's called GPT5, Orion, you know, we've heard about Strawberry and how, you know, technically this Strawberry Mode, according to reports, is right now being vetted by the federal government and will reportedly be used to create synthetic Data for open AI's next model. So yeah, will the kind of standard, you know, twenty a month, you know, chat GPT plus or the standard, you know, thirty dollar chat GP teams, teams subscription. Will they get access to these more advanced models or might we see a much pricier tier? Or maybe those tiers are going to face much stricter limits. So we're not sure. But this announcement comes on the heels of right when the CEO of OpenAI Japan announced at the KDDI Summit last week that the upcoming GPT Next model will be 100 times more powerful than, than GPT4. Also an OpenAI spokesperson did confirm that the terminology GPT Next. So yeah, we've seen all these different acronyms and names being thrown around. So an OpenAI spokesperson did confirm that GPT Next is not the actual name of the model but just a nod toward what is happening next in the GPT kind of family. So this, yeah, so we did see this kind of these, these two stories come out back to back where essentially it's saying that hey, the next version of GPT could be 100 times more powerful than the current version. And then we saw coincidentally, I don't know, a price tag that is a hundred times as much as the current $20 a month. Right? So might we see a 10x increase? $200 a month, $2,000 a month. I don't think it's going to go that high. Right, but at least these discussions are being made, right? Which you know, I think businesses need to keep an eye out on these stories, you know, because when you are building whatever AI solutions for your company, number one, you need to avoid vendor lock in, but you also need to be able to test different models and to find out which one is best, best for your business. Yeah, Joe saying yeah, wow, 2k a month, that's greedy. I, yeah, I mean who knows though, right? So we've also seen reports, you know, Open AI a couple of months ago kind of had this, you know, five step plan toward AGI leaked. Right. But maybe, who knows, maybe this isn't coming out for another two years. Right. And if that $2,000 a month model, I don't know, right. Maybe if we've quote unquote achieved AGI by then even though it's, it's a moving target. I don't know. I think a lot of companies, even small businesses, right, might pay $2,000 if that's the only way and if it is, you know, very advanced agentic like AGI platform, I don't know, companies might pay $2,000 a month. You know, if it is an incremental increase to a model where it's at now, I don't think so. But you have to think about this as well. The GPT4 family is now almost, what is it, more than a year and a half old, right? It's, it's, it's going to be two years old soon. So yes, there's been big upgrades in big upgrades in both performance and speed and cost in GPT4, Turbo GPT4O, GPT4, oh, Mini, which is great for developers. So we've seen these pretty big, dare I say, incremental updates, but we haven't seen a really seismic shift from OpenAI. And whether you're a fan or not, OpenAI is the leader in the space and it's not even close, right? Yes, companies like Meta and Anthropic and Google and even Microsoft working on their own models have kind of closed the gap. But right now it's still not close. Right? OpenAI in their model that is still technically right, based on the framework of a year and a half old is still by far the most powerful large language model in the world. So I don't know, we might see a pretty sizable shift in terms of capabilities. I would be flabbergasted if it's actually 100x. I don't think that's possible in the next year or two. But who knows? Who am I? I'm just a guy here talking to the smartest people in the world and bringing all of the news to you. All right, speaking of bringing the AI news to you, if this was helpful, you need to bring this to others. We spend hours every single day cutting through the fluff to only bring you the AI news that matters. So both, yes, on our Monday series, but also through the rest of the week when we do, you know, kind of dive into deep topics when we bring in expert guests. Got a lot of guests in the next coming weeks that I think you're going to be excited about. So if this is helpful, please share this with others and also go check out. I think we're going to wrap this up this week, our Thanks a Million giveaway to celebrate a million downloads for everyday AI. So if you haven't already, make sure to go sign up for that giving away a year of the baseline, the $20 a month large language model plans from ChatGPT or if you prefer, Gemini Claudian Thropic Copilot Pro from Microsoft. It's your choice. Go Sign up now and go sign up for the daily newsletter if you haven't already. Your everydayai. Com. 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 everydayai.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.