Ep 537: Perplexity goes agentic, Google Gemini updates, NYT/Amazon team up & more AI News That Matters

Topics Covered in This Episode:

1. Perplexity's Agentic Mode Launch Overview
2. New Google Gemini Updates for Workspace
3. Perplexity Labs' Reporting Tools Explained
4. Robot Launch: Hugging Face Ricci Mini
5. AI Missteps in Government Healthcare Report
6. Amazon & New York Times AI Partnership
7. Eleven Labs' Conversational AI Update
8. Anthropic CEO on AI Job Displacement

Keywords:

Perplexity Labs, Perplexity agentic mode, conversational search engine, AI-driven research, Perplexity Labs Pro, Mac and Windows apps, customizable dashboards, deep research mode, Claude artifacts, Google Gemini, email summary cards, AI-powered email summaries, Gmail AI update, Google Drive video summarization, Google Workspace, US Department of Health and Human Services, generative AI citations, The New York Times, Amazon AI deal, Alexa update, AI foundation models, Eleven Labs, conversational AI 2.0, enterprise voice agent, AI turn-taking model, multilingual conversations, Anthropic CEO, AI job displacement, white collar jobs, Apple WWDC conference, Apple AI gap year, Black Forest Labs, Flux one context, generative AI image editing, financial predictions.

Podcast Transcript

Jordan Wilson [00:00:16]:
Perplexity is pivoting hard into agents, and we're here for it. The New York Times, even though it's fighting with big AI, just entered into a partnership with big AI. And Google Gemini is getting a lot better for workspace users, and you don't even have to lift a finger. There's a lot still going on in the world of AI news. After last week, experiencing our biggest week of AI news ever, a lot of smaller yet meaningful updates happening this week. So if you spend hours every single week trying to keep up with what's happening in the world of AI and how it'll impact you, your company, and your career, stop doing that. Just join us on Mondays for our AI news that matters segment. What's going on y'all? My name is Jordan Wilson, and welcome to Everyday AI.

Jordan Wilson [00:01:18]:
This is your daily livestream podcast and free daily newsletter, helping everyday business leaders like you and me not just learn AI, but how we can leverage it to grow our companies and our careers. So if that's what you're trying to do, it starts here with this unscripted, unedited, livestream and podcast, but where you're actually going to leverage what we learn is by going to our website at youreverydayai.com. There, first, you need to make sure to sign up for the free daily newsletter. We recap the most important points from each day's show as well as giving you everything else in that newsletter that you need to understand what's going on. Also, on our website, what's going on there is now more than 530 episodes. You can go listen to, read, watch the video. Everything you need sorted by category to be the smartest person in AI at your company. Alright.

Jordan Wilson [00:02:09]:
Like I said, most Mondays, we do the AI news that matters. So if you can only join us once a week or maybe you're just always spending way too much time trying to read and decipher and understand the AI news, and you're like, is this meaningful or is this marketing spin? Join us Mondays. I can't get you straight. Live stream audience, it's great to see you. Jose joining us from Santiago, Chile. Love to see it. Fred joining us, from Chicago. Live Stream Crew on the YouTube machine.

Jordan Wilson [00:02:41]:
Like, Keith, thanks for tuning in. Christopher from Kentucky. Brian, holding it down for Minnesota. Joe in Fort Lauderdale. It's a good day to be agentic, he says. Yeah. A lot of agentic news. Alright.

Jordan Wilson [00:02:55]:
Let's get straight into it. First, perplexity is going agentic. So perplexity has introduced a new tool or a new mode called the Perplexity Labs that enables users to create complex reports, spreadsheets, dashboards, and even web apps, all supported by extensive AI driven research from Perplexity. So this new agentic mode works by spending about ten minutes or more on self supervised tasks, leveraging third party capabilities such as deep web browsing, code execution, and generating visuals like charts and images or even slide shows and spreadsheets. So Perplexity Labs builds on the company's existing AI search products, including their flagship conversational search engine, and their deep research mode, which produces in-depth well sourced documents after extensive, data gathering. So the AI agents behind labs can organize data, apply formulas, generate charts, create text documents, spreadsheets, dashboards, and even small websites without requiring users to have coding skills or development tools. So users can access examples and templates through a project gallery showcasing use cases like interactive war maps, stock portfolio dashboards, comparing traditional and AI managed investments, and futuristic social media platform designs. So the new tool or mode features an app tab for building simple dashboards, slide shows, and interactive websites with all generated assets, images, charts, CSV files, and code accessible for download in the assets tab.

Jordan Wilson [00:04:43]:
So right now, Labs is currently available to subscribers of Perplexity's pro plan on web iOS and Android with plans to extend availability to Mac and Windows apps soon. Livestream audience, have any of you guys used, Labs yet? So there's parts of it that I'm actually pretty impressed by. And then there's parts of it I'm like, alright. This isn't that great. One thing that I really like is what I just read there at the end is how it's all sorted. Right. So a lot of times when you're using, maybe other similar tools, I would say the most similar thing I can compare this to right now is Claude artifacts or similarly, Google, Google Gemini's Canvas or OpenAI ChatGPT's Canvas. Very similar in that it, combines kind of this research and, answer gathering from the web, but then it can create something new of value that's more than just straight text.

Jordan Wilson [00:05:48]:
So at least in my, initial testing, the web app capability was okay. I love being able to have that assets tab because that's something that, kind of the quote unquote competitive tools or modes don't necessarily have. So if it if it is something that is very, multimedia, that assets tab is really nice. But what I really think it does a pretty good job on is generating slides. Right? Kind of a small thing, that we don't talk about a lot, but yet we all spend so much time on. They're not gonna be fantastic looking slides. Right? Think if you open PowerPoint and you're looking at 10 templates, and there's two that are bad, there's two that are great, and then there's, like, six that are like, okay. These are okay.

Jordan Wilson [00:06:35]:
I'd say in your best case scenario, per perplexity labs will hit that like, oh, the okay, kind of, template. So this isn't something it's gonna be, you know, overly designed, but it actually does a way better job than I thought specifically on slide creation, which is something that the other, you know, those platforms, like I mentioned, aren't great at. And, you know, aside from dedicated slide tools like, you know, gamma or, you know, beautiful dot a I. Right? There's not a lot of AI research tools that also create slides. Yet, like I said, that's what so many business, you know, knowledge workers that's what we do. We create slides. Even right now, I have slides on my screen. Right? For our, podcast audience, I always have kind of a a, a screenshot of a news article, and then, I link to it, in the, newsletter so you can, you know well, first to give credit to the to the company that I kind of the read the article, but then you can go back and look to it.

Jordan Wilson [00:07:37]:
But I create slides almost every single day. So, I I try to get something, to the level of what I would use. It's not quite there yet, but it actually did a way better job than I thought when I said, hey. Here's my, you know, 10 news stories that I'm going over today. Go research them. Go bullet point them and create a kind of slideshow. So it's not at the point where I'd wanna use it yet necessarily, but it's definitely passable. Right? My use case is different.

Jordan Wilson [00:08:06]:
I like to put up my slides for, you you know, a lot of people to watch on the livestream, but pretty good. It's pretty good, and I was pretty impressed. Mahan here, says, perplexity is innovative, will grow big. I actually have thoughts on that, because I said back in January, I said, perplexity is going to have to pivot or they will get squashed. And what we've seen here as well, this is a pretty impressive pivot, I'd say. Right? And, yes, like, doctor Harvey Castro said, great, call out here, said Manus AI just added slides. That's that's big as well. So, kind of a very popular version if you aren't following the space very closely.

Jordan Wilson [00:08:55]:
Manus is, an AI agent, like, similar to OpenAI's operator, a computer using Internet using agent, that now they just roll out slides mode about a week ago. And the good thing about the Manus, AI slides is they're editable as well, whereas something like in perplexity, they're not necessarily editable how you would want to edit them without regenerating the whole thing. And if you regenerate the whole thing, it might change 10 things that you didn't want and only change the two things that you want. But, I think so far, pretty good. Douglas saying, I have made dashboards side by side comparisons, with Gemini Canvas. For me, I like the perplexity lab output better. Yeah. For sure.

Jordan Wilson [00:09:34]:
It depends on what you're trying to build. Right? A lot of times what I'm trying to build inside these, you know, artifacts or canvas, etcetera, is something a little more interactive in visual and not necessarily, asset based or research based even. It's more of, you know, building little mini websites or something like that or, generating certain types of code. So at least for my use cases, I would probably use perplexity labs a little bit more, for information gathering. So kind of combining, kind of deep research with slide formation. That's what I would use it for. And then when you need to be able to download those assets, that's what I would personally do, but I think a lot of good use cases. Also, great call out here I had in my notes for later.

Jordan Wilson [00:10:21]:
But also, yeah, perplexity does, have their comment browser that they've been slowly rolling out. They did the first rollouts about a week ago. So, pretty impressive so far. Alright. Our next piece of AI news, Hugging Face has released an open source robot called Riichi Mini, offering affordable AI hardware for developers. So the new open source robot is designed to help developers test and build AI applications, and it is priced between $250. Yeah. $250.

Jordan Wilson [00:10:56]:
Not $250,000. It's a little robot guy. Alright? For a podcast audience, it's this, you know, cute little, you you know, desktop sized robot, but it is making it significantly more accessible and affordable than other robotics hardware on the market. So Ricci Midi resembles a small Wall East style bust that can turn its head and interact with users through speech. So the product is positioned as a kind of rise, raspberry pi, for robotics targeting AI developers by providing affordable, customizable, customizable hardware for experimentation. Also, the open source nature of Ricci mini means developers can modify both its software and hardware, which I think means this is gonna be the thing that ultimately pushes humanoids and, like, robotics into the mainstream, into the home. Right? I think what's gonna happen is this new Ricci mini is probably gonna be extremely popular just because it's only 250, 250 to $300. It is open source, so you have to know, the basics.

Jordan Wilson [00:12:03]:
You gotta be you don't gotta be a geek, but you have to be dork ask, to be able to set this up and run it. But then after that, I do see a lot of households probably, using this piece of hardware to do it. And I need to double check on the price actually. Alright. Yeah. Sometimes sometimes I'm a little, tired when I put all of my, notes, together, and I wanted to double check on this price, because my screenshot here, from the article says 3,000, and my notes said something else. So, yeah, I made sorry. I made a mistake.

Jordan Wilson [00:12:42]:
It's not $250. Alright? Bad human. I just hallucinated, because I get tired sometimes. I got my coffee here though. Sorry. $3,000, not 300, 3 thousand dollars. Still, when you look at for comparison like the Tesla Optimus gen two, that's expected to cost, probably, like, $25,000, and it hasn't even been released. Right? Even though we're like, oh, we're gonna get these robots very soon.

Jordan Wilson [00:13:08]:
Right? And then you have other, more advanced humanoid robots that are costing hundreds of thousands of dollars. So for an open source, robot or humanoid to only cost $3,000, not bad. Not bad. And I do see that cost going down, maybe not to $250, like I accidentally said, missing a zero there. But maybe to a less than a thousand dollars, I could see that happening within a couple of years. So I do see this as a huge boon for not just robotics, but in home humanoids as well. You know, Marie said, oh, Jordan. I never thought you would hallucinate.

Jordan Wilson [00:13:48]:
Yeah. It's a feature, not a bug. Alright. Our next piece of AI news. Well, Google Gemini is getting a little more user friendly. So, Google has launched automatic AI powered email summary cards in Gmail, which now appear at the top of emails without requiring users to tap for a summary. So this update means that Gemini will proactively summarize long email threads and keep the summaries updated as new replies arrive. That's the part that I'm looking forward to.

Jordan Wilson [00:14:24]:
So right now, workplace admins can control whether users have access to these summaries via the admin council, giving organizations some oversight. So if you're wondering like, hey. Why haven't I seen this? That's probably because the your workspace admin maybe hasn't enabled it yet. So it is turned on by default, which, a lot of people are saying from a privacy perspective is, problematic, but I'll say this. It's not. Right? Because you're going to go in and use this feature anyways, whether it's turned on by default or not. So in certain countries such as The EU, UK, Switzerland, and Japan, it will be opt in or turn off by default. In The US, when this is fully rolled out, it will be, opted in by default, but you can opt out.

Jordan Wilson [00:15:11]:
So, again, once it's enabled by your workspace admin. So this manual option to generate a summary still does remain available as a what Google calls a clickable chip at the top of emails and in the Gemini side panel. Alright. Another kind of small but, still pretty big quality of life improvement for Google Drive workspace users is a new feature that Google has rolled out in Gemini that lets workspace users get quick summaries and insights from videos saved in Google Drive. So the same thing, saving a Google or sorry, saving a Google in a video drive, saving a video, inside Google Drive, Gemini will now automatically use its AI prowess and summarize that thing instantly without you having to go in and ask it to summarize, which is really nice having this available inside Google Drive. This is something I spent a lot of time using, in AI studio for this exact reason. So the Gemini, inside Google Drive now supports not only video files, but also by extending, its previous document and a PDF summarization capabilities. So users can interact with a chatbot interface inside of Google Drive to request summaries or specific details, such as listing action items from recorded meetings or highlighting major updates in announcement videos.

Jordan Wilson [00:16:40]:
This feature, though, does require captions to be enabled on videos and can be accessed through Google Drive's overview, previewer or a new browser tab. It's currently available in the in in English only right now for Google Workspace and Google One AI premium users, plus anyone with, Gemini business or enterprise add ons with a full rollout expected in the coming weeks. So you might not see this yet, but if you are on a paid, Google plan, you will be seeing this feature pretty soon. So I don't know. Are these Gemini features worth writing home about? Are you gonna use them? I know for me personally, I'm gonna be using them because these are features that, if I'm being honest, Gemini hasn't been that great at inside of its default workspace apps. What's been great is Google AI Studio. So I'm constantly bringing long email threads over to Gemini AI, like, super long email threads because of the context window and being able to use Gemini 2.5, Pro is is nutty, as well as video. So I've been using, Google's AI studio for a lot of these things now, but it's really good, to have these rolling out directly, to workspace apps.

Jordan Wilson [00:17:51]:
Hopefully, the integrate, the integration and the rollout goes well. Sometimes it's super slow, especially when anytime this, this type of, utility comes to workspace users. So, unfortunately, sometimes more of the, kind of, great AI features don't roll out initially, to workspace users. So as an example, I'm on the new, you know, $250 a month, Google Gemini AI Ultra plan, but I can't access it via workspace, which absolutely stinks. Right? So I can't really use it with my business data. I have to use it with my personal Gmail account. So if I wanna take advantage of a lot of the more powerful AI features, I would have to set up an automation that sends all of my work emails and files over to my personal Gmail account. So at least it's good that Google is rolling these out, into workspace.

Jordan Wilson [00:18:44]:
So love to see it. Next. This is not a good look for the federal government, but a government health care, government health report led by Robert f Kennedy junior and released by the Department of Health and Human Services. It's causing a lot of controversy because this paper was found to contain a ton of fake and botched scientific citations, raising concerns over the use of generative AI in official policy documents. So this is according to reporting from the Associated Press in the Washington Post. So the White House responded to criticism by updating the report and correcting citation errors, but downplayed this issue as, this is just a minor formatting thing rather than huge, a huge mistake. So at least one of these studies, so it was the overprescribing of oral corticosteroids for children with asthma. It does not exist outside of the reports.

Jordan Wilson [00:19:52]:
Alright. So, in 37 of the 522 footnotes were repeated multiple times according to reports. Also, several URLs in the report included the tag o a site, which is a tag linked to OpenAI. So they did not even bother to go and try to say, like, this is an AI generated. They included those, like, citations that clearly show that this was generated by a large language model. So not a great look. Also, some of the studies that the report talked about, were cited correctly or they misrepresented or inaccurately summarized. So as an example, there was a claim in this study about a 40 fold increase in childhood bipolar in ADHD diagnosis tied to psych psych, I can't even talk, psychiatric, manual that wasn't published until years after the cited period.

Jordan Wilson [00:20:56]:
So, yeah, just a government report having to do with health that was released with just a ton of made up information. So not a good look when this is happening, not just at the federal government level, but for something extremely important like health. 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 GenAI. 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.

Jordan Wilson [00:22:10]:
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. We'll help you stop running in those AI circles and help get your team ahead and build a straight path to ROI on GenAI. Marie says, seems like the administration is hurrying and rushing and not checking data first. That is what it looks like. Alright. Speaking of checking data, you gotta check this story. I was not expecting this because Amazon, and the New York Times have struck an AI deal. And I'll tell you why that's important here after the details.

Jordan Wilson [00:22:53]:
So here's the details. The New York Times has reached a multiyear agreement with Amazon to allow the tech giants to use its editorial content across Amazon's AI platforms. So the deal enables Amazon to, incorporate real time summaries and short excerpts from the New York Times and its other properties like New York Times cooking and the athletic, the, sports, publication into Amazon products such as Alexa. Will Alexa finally get a little smarter? Hopefully. So Amazon will use the times content to train its proprietary foundation, foundational AI models, helping improve the quality and relevance of its AI driven services. So terms of the agreement were not disclosed, but the partnership reflects a growing trend of news outlets opting for licensing deals with tech firms rather than pursuing litigation over AI content use. Here's why this is pretty noteworthy. Well, it's the first big licensing partnership the New York Times has struck.

Jordan Wilson [00:24:00]:
That's because they have one of the most noteworthy and technically famous or infamous depending on how it lands, lawsuits against Microsoft in OpenAI. So the New York Times filed this lawsuit in December of twenty twenty three. It is still in, the courts where the New York Times is suing Microsoft and OpenAI for allegedly copying millions of the New York Times articles. So one of the things that the New York Times, asked for, in its lawsuit was for the GPT technology to be destroyed, which again, I think has, like, a next to zero chance of actually happening, because the whole, world now runs off of, like, the GPT technology. And, even if it were theoretically possible to destroy it, it's kind of too late, but it would essentially bring the world's economy, to a screeching halt. Yet, this is extremely noteworthy, because this is the first time that New York Times is entering into an agreement with a big tech company being like, yeah. Go ahead. Train your model on our data.

Jordan Wilson [00:25:11]:
So I'm not saying they're the the the last, media giant, to fall in line, but I've been saying this all along. Yo. I was a journalist for seven years. Right? So I you could say I have bias either way. Right? But there's no other option for news organizations. Either there's three routes. Route like, when it comes to AI. Right? Because, like, a lot of news organizations are like, oh, well, we'll block all these, you know, web crawlers so they can't scrape our information.

Jordan Wilson [00:25:44]:
That won't work because number one, the crawlers don't always listen to your instructions and your robot stock text files. Number one, if you wanna show up on Google, you don't have a choice. You can't opt out of Google AI training, but opt into Google search. So if you wanna be discoverable, you have to opt in. So you either slowly die because people aren't gonna find you anymore. If you wanna opt out, you're like, I don't want large language models to scrape my content. I'm gonna opt out. Well, number one, good luck.

Jordan Wilson [00:26:12]:
Number two, you're gonna die because you're not gonna get new users. So you either slowly die, you sue. Right? You sue all these companies, which is what a lot of media outlets and news organizations are doing, or you enter partnerships. There's no other way. Right? There's literally no other way. And even if you are trying to, you know, block all these, you you know, AI, scrapers from getting the information from your website and, you know, putting it into their, their their training set. Even if you do that, there's other scrapers out there that scrape the entire Internet. They make these third party datasets, and then the large language model in AI companies train off those datasets anyways.

Jordan Wilson [00:26:52]:
So there's no way around it. So, pretty interesting and newsworthy, announcement there from the New York Times. But can we please finally, when are we gonna get this Alexa that's powered by Claude and has all this information? It was supposed to be rolling out as a paid service. I still don't have it. I will pay for that. I will pay. Right? It's obviously great to be able to talk to, you know, ChatGPT or Gemini live, right when I have my phone near me. But I still am am calling out to Alexa or Siri a lot of the times, and I could get better answers from talking to a brick.

Jordan Wilson [00:27:33]:
It is mind numbingly stupid. Right? Alexa, Siri. So please, can we get this smarter version already? I'm not counting on it from Apple, but at least, from, Amazon and Alexa. Can we please just roll it out? Alright. Doug was saying, Jordan's saying Alexa right now and setting off thousands of Amazon devices during the tech podcast. Yeah. I've heard that happens quite a bit. Sorry.

Jordan Wilson [00:28:05]:
I need a code word. How can I talk about these things without saying that? I can spell them out. Right? But that's a lot of work. Alright. Next. And I don't usually talk about a lot of, a lot of announcements on the creative side. Right? So, you you know, photo and video tools or, audio tools, but I have two this week that I think are pretty big or from a quality perspective are worth talking about. But, you know, livestream audience, if you wanna see more updates on these visual tools, let me know in the AI news roundups.

Jordan Wilson [00:28:39]:
Generally, I'm focusing on large language models. So, generally, I'm focusing on, OpenAI, Google, Anthropic, Microsoft, Meta. Right? Some of the big tech trillionaire companies because I think those are the ones that impact most of us business leaders. So if you want more news on kind of the the the multimedia or creative side, let me know. But I think this one, from Black Forest Labs is big enough to talk about because the quality is very impressive. Alright? It is already on par. If you wanna say state of the art AI image generating, it's there. So Black Forest Labs has launched Flux one context.

Jordan Wilson [00:29:19]:
So podcast audience, that's contact with a k. Alright. A new family of image generating and image editing AI models. So the most advanced model is called the flux one context, and it can generate images from text prompts in optional reference images, delivering results up to eight times faster than leading competitors. That's according to, Black Forest Labs. So the suite includes two main models, Flux one context, which allows for multistep image refinement while preserving style and character, and flux one context, max, which emphasizes speed and prompt accuracy. So unlike previous models, these new versions are not available for offline download. Alright.

Jordan Wilson [00:30:03]:
So, yeah, a lot of people are like, oh, wait. I can you know, this is an open source. This one's not. Alright. So you can't, you know, download this and use it right now. In that way, it's only available right now in a private beta for safety and testing purposes according to Black Forest Labs. So they're also launching a model pray, playground giving users, 200 free credits to try the models online on their website. So this release comes as competition heats up in the AI image generation sector with Google and OpenAI recently releasing their own advanced models.

Jordan Wilson [00:30:41]:
So the company Black Forest Labs reportedly sought $100,000,000 in funding in at least a $1,000,000,000 valuation last year, and they are based in Germany and founded by former members from the StabilityAI team. So why the heck am I talking about this? Well, we've talked about, not at great length, but many times how good specifically Google Gemini is with editing images with simple text prompts. And then, obviously, we covered OpenAI, their GPT four o ImageGen, right, which, went viral, kind of online many times. But I would say at least today, Flux one contacts from Black Forest Labs is on par or even better than those. And here's why I think it's important enough to talk about on a, you you know, on a new show. Don't trust anything anymore. You can't. Right? Anything you see online rather than thinking, is this is this AI generated or is this real? Right? You should probably assume from now on, right, and think this is going to add, which is why they weren't making this available, kind of, as an open source model.

Jordan Wilson [00:32:08]:
This is gonna add to the deep fake problem, right, to the misinformation, disinformation epidemic that's sweeping the nation at least here in The US. As these models get better and better and you can get character consistency, and you can go in and edit photos, and you can't even tell. And then those can be the base for videos in a very powerful tool like Google v o3. Right? It's scary. You should assume everything you see online now. Right? There there's been a lot of recent stories. Maybe I'll do a a an entire episode on this, but people are launching, fraudulent GoFundMe's. Right? People are are obviously using this for, blackmail in a lot of bad ways.

Jordan Wilson [00:32:50]:
This is gonna be a problem. But for our audience, you should always assume from here on out, Everything is AI generated. Everything you see, unless you know otherwise. Right? Even probably me, I would tell you. Right? But you should always assume because that's how good these models are now. And I think it took the general public maybe twenty years or at least ten to fifteen years to come up with this concept of, oh, things could be photoshopped. Right? Things you see online or, you know, something in, in a magazine, something, you know, you're you're seeing on TV. Oh, something could be photoshopped.

Jordan Wilson [00:33:33]:
Right? Something could be digitally altered. Now I think you have to start with that as default with anything you see. Right? Whether it's an ad. Right? These UGC ads, whether it's something you're seeing on TV, you you know, a a pre roll message on YouTube, something you see from a celebrity that you're like, oh, this is interesting. I wouldn't have thought this person would have this take or even your favorite news anchor. Assume everything is AI generated until you are able to confirm it's not. So when in doubt, assume it's AI generated. And adding to that, our next AI news story, Olin Labs has launched their conversational AI two point o.

Jordan Wilson [00:34:17]:
So this is a pretty significant updates to its enterprise voice agent platform. And this is just four months after the original launch of their conversational AI platform. So if you don't know, Eleven Labs, I would say they've traditionally been the leader, or a leader, probably one a or one b in text to speech. Although now you are seeing a lot of, competitors both open source and, from the big tech conglomerates in OpenAI and Google. But the new Eleven Labs text to speech conversational platform, these new updates in their two point o version feature state of the art turn taking model. And that's great. So and you might be wondering, okay. Why does this matter? Well, just like how I said, assume everything that you interact with is AI generated.

Jordan Wilson [00:35:07]:
I would assume, probably within a year, anytime you call a call center, you are going to be calling or talking to at first, in AI voice. Right? Or an AI voice tree. Okay? Which isn't necessarily a bad thing. Right? Versus talking to somebody in a in in a call center, in a different country where you can't really hear anything or talking to one of those, you you know, computer generated ones and you're just like, you know, human human operator operator. Right? And you're just screaming things to get past, the robotic prompts. So I don't think it's necessarily a bad thing, but I think it is, things like Eleven Labs conversational AI two point o. This is gonna be the future of how we interface, with the rest of the world, especially on the phone and on websites. So this turn taking advancement, I think, is what makes it kinda noteworthy.

Jordan Wilson [00:36:01]:
So it does, this new model does a little bit better job of, identifying when the human on the other end is done talking or not. Right? So they have some demos where someone is talking, and it seems like, oh, maybe they're just thinking or taking a pause or retrieving some information where normally a, voice AI might instantly interrupt, or there might be a huge latency when you are done. So that's what, is kind of pretty impressive here at least from the demos that I looked at, at eleven Labs. Also, now it has integrated language detection, that allows for seamless multilingual conversations. And the update also introduces a built in retrieval on many generation or RAG system for companies to use enabling voice agents to instantly access external knowledge bases while maintaining low latency and privacy, which is especially useful for industries like health care and customer support. Multimodal communication is now also supported, so agents can interact via voice, text, or both. So, yes, companies, can go build these conversational AI agents and embed them pretty simply on their website or, you can use this, as a traditional, kind of phone operator or phone operating system. So right now, enterprises can automate large scale outreach with batch outbound calling, allowing multiple calls to be made, at once for surveys, alerts, or personalized messages.

Jordan Wilson [00:37:34]:
So right now, subscription prices range from a very limited, free tier with, limited minutes to a tier that is more than $1,300 a month for a business plan. So it just depends on how often or how much you're gonna be using it. Those are the base plans. So if you're a true enterprise and you wanna scale this out, right across, like, call center type, it's gonna be way more than $1,300 a month. Those are just the base business plans. Alright. Our next piece of AI news, a job apocalypse is coming. Well, at least according to Anthropic CEO, Daria Amodi, who warned that AI could eliminate up to fifty percent of all entry level white collar jobs within five years.

Jordan Wilson [00:38:28]:
So the Anthropic CEO kind of made a short media tour last week shortly after their clawed four product was announced and, you know, kind of sounded the alarm on, hey. This powerful AI that we're all building, it's gonna take a lot of jobs. So Amodi predicted that US unemployment could spike up to 10 to 20% due to AI driven automation, a dramatic jump from the current 4.2 unemployment rate. So he says the threat to white collar jobs is not being acknowledged by society or lawmakers and that the impacts will hit faster than most expect. So a lot more on this story. Actually, I'll go ahead and tease it tomorrow. So I won't spend any more time talking about this because we're gonna do our hot take Tuesday on this exactly because I think this one, needs a little bit more exploring, and I have some takes on this, particularly the timing of all of this little media tour. Alright.

Jordan Wilson [00:39:33]:
So, make sure to, tune in tomorrow as we go over that. And last but not least, on our AI news that matters. So according to reports from Bloomberg reporter Mark Gurman, Apple's AI conference or sorry, Apple's WWDC conference is gonna be not really about AI this year, and that's probably for the best. So at w w d c twenty twenty five, Apple is expected to reveal only minimal advancements in artificial intelligence a full year after going all in on AI and falling flat on their face. So this is according to reports a signal that Apple realizes they are very far behind in AI. And, also, they've been facing a lot of class action lawsuits, right, because all of these, AI features and they even went as bold, my gosh, to try to rebrand AI, call it Apple Intelligence. Last year at their worldwide developer conference, they, you know, hyped up all this, you know, AI that was gonna come out to iPhones. They even had a bunch of marketing commercials, and now they're facing a lot of class action lawsuits because most of what they announced and marketed never came to fruition.

Jordan Wilson [00:40:49]:
So according to, this report, the biggest AI news expected, and this is next week, June ninth. So one week from today is the kickoff of Apple's w w w DC conference. So the biggest AI news will be Apple opening its on device foundation models, which have around 3,000,000,000 parameters to third party developers. So doesn't look like there's gonna be a lot of AI announcements, which is probably for the best, because Apple is I can't wait for someone to make the movie about how this is probably one of the biggest failures in modern business history, right? Apple's inability to successfully roll out any really usable piece of artificial intelligence when they are multiple years behind their biggest competitors, like, Google Samsung on the device side, Microsoft, right? They are so far behind. It is laughable. So, a couple more things from this report. So Apple plans to introduce several smaller AI related features in their new operating system, which they're also renaming the operating system. It'll be called iOS 26 now.

Jordan Wilson [00:42:14]:
So they're, aligning the, the operating system numbers to the years. Apparently, they spent, like, millions of dollars to work with consultancy, and that's what they got. Right? Instead of I don't know. We're on, like, iOS, like, '17 or 18 right now. Similarly on the macOS. So instead, they're just gonna change it to the year that it's released. Anyways, Apple does plan to introduce several AI related features in the next operating system, including a new battery powered management mode, a revamped translate app integrated with AirPods and series, and labeling some app features in Safari and Photos as AI powered. So Gurman, the, famed Bloomberg reporter who gets this right almost every single time it is the one breaking all the Apple news, has described this as a gap year for Apple, which is hilarious.

Jordan Wilson [00:43:07]:
Right? Like, yeah. We're gonna take we're gonna go ahead and sit this one out. Right? Anything that Apple tried to announce between what was announced, two weeks ago between Microsoft and Google, what they announced at their, respective, build and IO conferences. If Apple tried to end it, like, do what they did last year with AI, they would get laughed at, and their stock would go in the tank. I would expect right? I don't I'm not I'm not your financial analyst, but I would expect, Apple's stock is not gonna be looking great for the next, month or two, because of this kind of gap year. So I love, I love that, you know, this is how it's being described. But Apple is, according to reports, actively developing more advanced AI projects, including a large language model version of Siri, a redesigned shortcut shortcuts app, project Mulberry focused on health in a ChatGPT like competitor with web search. But for the most part, what we're gonna see is they're essentially you know, Apple has their own edge AI models, and they have larger variations of their own in house large language model that they built.

Jordan Wilson [00:44:18]:
They're essentially gonna be opening this up to third party developers. So, developers using their apps, can take advantage of the on device AI, which in the end might actually just be better than Apple trying to do it because they failed, and they failed miserably. So, that is a wrap. Yeah. Allison just saying, seriously, Apple? Yeah. Seriously. Big bogey here on YouTube just throwing about Apple intelligence with a bunch of face palm emojis that's saying it nicely. Alright.

Jordan Wilson [00:44:50]:
That is our recap of what's happened, but got a new little segment. Hey. If you're still here in the live stream, just drop a yes or no. If you if you want this little segment to be tagged on at the end of our future, AI news that matters. Right? So I'm just calling this rumors and what's next. Alright. So live stream audience takes two seconds. You can even just say a y or an n.

Jordan Wilson [00:45:17]:
Do you wanna hear this? So these are some rumors and what's next, what you should be expecting in large language model developments. So a lot of these things could happen as soon as this week. A lot of them are expected to happen this week, or they might happen later in June. But here we go. OpenAI's o3 pro could be announced very soon. Perplexity's Comet browser, that we already talked about, which kind of doubles as a computer use agent, maybe getting a wider release, but it's already been starting to roll out last week to people who are very early on the wait list. So that should be rolling out to everyone else soon. OpenAI's GPTs, which have been pretty much widely ignored for the last year and a half, could finally be updated and get new features as well as being able to use the o3 model, which I can't emphasize enough.

Jordan Wilson [00:46:13]:
That could be some of the biggest large language model news in, like, six months if OpenAI actually updates their GPTs. Could be huge. Grok sorry. Grok 3.5, could be coming out, any day now, but this is also after, Elon Musk saying it's almost here for multiple weeks, but we're seeing some reports that it could be rolled out within the next week or two. Claude maybe getting an artifacts studio, which is an easier way to save your different Claude artifacts generations as well as kind of an inspiration gallery that you can see, and learn from other people that are using the Claude artifacts feature. We are and this has been confirmed by Logan Kilpatrick from Google. We will be seeing a new version of Google Gemini 2.5 Pro within two weeks, And I would assume that any benchmarks that Claude, was able to achieve, with, Claude four Opus or Claude four Sonnet, I'm assuming many of those are going to be absolutely erased with this new version of Gemini 2.5 pro. Sorry, Infrappy.

Jordan Wilson [00:47:23]:
You're not gonna be able to keep up with Google. Not this new Google. Sorry. And then like we said, next week, we will have WWDC happening on June 9, and this is gonna be a nothing burger for AI. So, we won't be able to cover that next week on the show because it's actually gonna happen about three to four hours after, the live stream happens, but we will be covering it later in the week even though it's gonna be about nothing. Alright. So that is a quick recap of what's going on in the world of AI news. So again, we had Perplexity AI launching their labs in agentic mode or tool.

Jordan Wilson [00:48:02]:
Hugging Face launching their open source robot, which cost $3,000, not $300, but I think it is still going to be extremely big news, regardless for robotics and the humanoids. Google has rolled out two pretty new, pretty useful features inside Google Workspace, both auto summarizing long emails in Gmail as well as being able to summarize, videos inside Google Drive, but Google Workspace admins do have to enable those for paid users. We had a pretty bad look for the federal government as, reports saw that they kind of hallucinated or fabricated a bunch of things in a major US health report. It was not a good look. Amazon and The New York Times struck a partnership to bring AI content to Amazon's AI platforms, including Alexa. Please get that going sooner rather than later. We saw some new creative, labs and features come out both from Black Forest Labs and their new image generator suite, which I think is on par or better than the, Gemini, image editing and the OpenAI image editing. We saw a pretty big update from Eleven Labs launching their conversational AI two point o with some major upgrades for enterprise voice agents.

Jordan Wilson [00:49:17]:
The anthropic CEO warned that AI could eliminate half of entry level white collar jobs within five years. The AI jobs apocalypse, we're gonna be talking about that one tomorrow on hot take Tuesday, so make sure to join us. And then last but not least, according to reports, Apple is gonna release essentially nothing in AI at their WWDC conference next week, and it's gonna be an AI gap year. Alright. I hope this was helpful. If so, please go to youreverydayAI.com. Sign up for the free daily newsletter, where we're gonna be recapping these stories as well as keeping you up to date with everything else happening in the world of AI. If this was helpful, don't be a jerk.

Jordan Wilson [00:49:56]:
Share this with someone. Right? Click that repost if you're listening here on LinkedIn. We'd appreciate that or on Twitter. Tell someone about this. Even if you think everyday AI is your little secret, you can't keep it to yourself. Share with others. Like, the whole point of why I do this, why I keep it free, why I do it every day is because I know artificial intelligence and generative AI is extremely hard to keep up with. And when we talk about job displacement and all of these things, everyone needs access to free, unbiased generative AI education.

Jordan Wilson [00:50:25]:
That's what I do. I do it all for you. So please return the favor, by sharing this with your friends. If you're listening on the podcast, I'd appreciate if you could follow, subscribe to the show, leave a review. Thank you for tuning in. Please join us tomorrow in everyday for more everyday AI. Thanks, y'all.