Transcription
Welcome to another week of AI innovation. And this week is one of those that I love to do because we have a bunch of chatb updates. And I happen to know that most of you are still using chatbtd as their main used app. And let me just say a sentence about the fact that that is the case for me too. 70 to 80% of my personal AI usage is within chat GBT or claude. I play in all of these other apps and there certainly uses, but as somebody who doesn't rely upon coding on a day-to-day for me roughly twothirds of my work still happen in Chat GPT or Claude. Now mind you I say work I experiment and play a lot in the others and today we have a bunch of practical chachibility updates for you along with updates from Google and claws. The big boys all released a bunch of interesting things this week. We're going to talk about all that and more in this week's episode of AI news you can use show that pulls together all the AI innovations from across the internet. We filter for the stuff that matters and is usable and then I have the pleasure of presenting it back to you. Let's begin.
The big story this week is the story of OpenAI. They're raising a brand new round. And I guess to argue that and get some public attention, they also released a bunch of features inside of their core product. Along with that, they raised $122 billion at a $852 billion valuation. That's almost a trillion valuation for a private company. I believe that's the single biggest fund raise in venture history. And all of that happens with this mission to build a unified AI super app.
Now before we get into the features I just want to give you a little bit of context because as it has been recently revealed most of Chachibd's revenue around 3/3s of it come from chatb subscriptions whereas for a company like claude it's actually pretty much the opposite where most of their revenue comes from API usage developers and other applications using it that also includes claude code and open claw by the way and now that openi clearly stated that their goal is building this AI super app what they're referring to is expanding their consumerf facing application and really making this beautiful chat interface that we're all so familiar with into something that is AGI or the one unified application including everything. And I guess to prove to the public and to the investors now we can get into the features. They released a bunch of stuff this week. I'll start with my personal favorite which is the ability to share your location with chatt and ah yes yes yes you know privacy concerns I know it is very convenient and I'll show you that in a second but first let's look at how to access it go to your settings in charge data controls and now you have this location feature this is brand new as of this week and they didn't even have a proper announcement around this I find this so convenient though because when I have this on I can ask questions like what is a good co-working cafe near me This will also work on the phone, of course. And now that it knows I'm in Miami right now, it can make all of those recommendations. And as you can see, this is pretty much my location right here. So, I can use the interactive app here. So, beautiful pictures in here. Notice how in previous episodes I was saying they're really moving away from this wall of text at all means. And even a simple search like this like has maps, images, links. This is that direction of moving towards this AI supre away from just a chatbot. So yeah, this one looks really nice and I might check that out and you can use it for anything you do. It just makes it convenient and maybe even superior to using like a Google maps to finding these things. I mean the mostly fake reviews on there aren't worth much to me personally. This way of presenting it with the map and then the results in the images is just preferable to only having the map and then you know having to open each entry. And then finally I would add on the phone you can set so that it can only use a location while you're in the app. That might be the smart thing to do. It's a very convenient thing that Gemini already had and now Chad has it too.
Okay, so let's look at some other changes. A big one is that if you have a lot of text and you add it to chat. So let's say I just follow up my first location chat here with expand on each one of these. That's a thing I like to use. I also like to use the opposite of give me a more concise version when I need to get the high level. And the novel feature here is that if you give it text with over 5,000 characters, it won't just add it as text to a new chat or an existing chat anymore. It will add it as an attachment. There's a difference there. And I need you to be aware of this difference to get the most out of this tool. So, what I'll do is I'll just select all and copy that. And if I go into a new chat, you will see that if I add it, it's not added as text. It adds it in a .txt file as you can see here. Now, here is the deal. Attaching something isn't equal to adding it as text. That's why they give you this button here, showing text field. So, if I press show in text field, it will actually appear in my chat box right here. And all of a sudden, it's here. Now, you might be asking, Eigor, what's the difference between attaching it and having it in here? Well, they were never super clear on this, but if it's as pure text in the box, as you can see it right here, every single detail is added to the context window that you're limited by within chat GPT. Just as a quick refresher, every AI model has a so-called context window, which you can think of as a certain limitation on how much it remembers. It's basically a rolling window. So, if you give it more inputs than it has context, then the context window will roll down and forget what's in the beginning, and AI basically gets amnesia.
Now, what happens if you add something as an attachment is that it won't be directly added to the context window. There's some sort of summarization step in between this document and it adding it to your prompt as you would if you're just writing in here like so. In other words, in attachments, you sometimes tend to lose certain details. Now, not always, but if you really wanted to retain every detail, adding it as text versus an attachment is superior, but it will limit the amount of other context you can add. Now, if you don't believe me, they even highlighted this in the documentation that they never promoted publicly, though. Yeah, if the text is over 5K characters, CHP will convert it into an attachment. This keeps the composer cleaner. This is the composer right here. And this is the critical part, helps prevent large pastes from consuming your entire context window. As a user, I wanted to consume my context window. They as a company don't want us to consume the context window because they have to pay for all of those tokens, for all of that context running through their system. Especially if you continue the chat, they'll have to run all of those tokens through the system again and again and again and you're just paying for $20 subscription. So, they don't want that. We as consumers usually do, unless you really have a lot and you don't want it to be consumed by the context window. So, my recommendation would be if you're doing quick work and you add text in there and press this button and you'll get more precise recall on what you actually added in there. Good to know.
Few more chat with the updates here. And these are quick ones cuz they're just single features that most people aren't using. For example, Google Drive had multiple apps. So, you had Google Docs, you had Google Sheets, you had Google Slides, and then you also had Google Drive. Now, they unified it into just Google Drive. When you connect Google Drive, gosh, if I say that one more time, I'll have a seizure. If you connect the app that I just mentioned so many times, it will also connect Google Docs and all the others. If you go to my account settings and under apps, you will see that it's all unified here. And you might have to click enable sync once, reactivate everything. If you connected the others before, you don't have to worry about this. It's all connected. But if you're using this regularly, I would actually go in here and make sure it's reconnected. So, it's all clean under one connector. If you're watching the show, you'll know I'm not a big fan of these connectors cuz they're very leaky. they do some of this like summarization in between bringing in stuff and putting it in here. So, it can't really fully rely upon it. It can be useful though, especially if you have shared context with other teammates. Anyway, it's simpler now all under one connector. It's a step in the right direction towards having a AI super app. And as we're talking about apps, they actually updated some of them. So, they have updates to Box, notion, linear, and dropbox. And basically, you can just do more stuff with the apps within CHBT, which can be nice. I don't know about you, please leave a comment if you are, but me personally, I don't really use these much. I just find them to always do the first step in the workflow inside of the app and then I have to go out and go into the app anyway. Might as well do that immediately rather than doing one step in chat and the next three steps in Dropbox. Anyway, if you use them, you can now upgrade them now to the new versions like so.
They also made some updates to the mobile application. So, if you check that out here, basically created a full screen view of the menu and you can kind of swipe to it which is nice. Everything is larger here. You have your chat history, GPTs, image generation or the apps on top. I like it. Basically, there's two screens now for menu and chat. And another quick chat change, and I'm excited for this in my car, is chatbt voice mode in CarPlay. So, you can, you know, talk to the assistant via CarPlay. I wonder how deep the integration runs with other apps. It would be kind of really nice if it would actually act as a Siri that works for my car. I like that idea. You need to have iOS 26.4 or later, and you need to have a iPhone and iOS running that supports CarPlay. Another thing that they improved is the shopping experience within chat. Remember they introduced partnerships with several brands and now you could see the products directly inside of chat. Interface wasn't the best. They overhauled it now and actually now it looks fantastic as you can see right here. It also looks at the shops more frequently. So it has the freshest products from the various providers. And then finally they added some plugins and codecs if you use that instead of claw code which I don't know many people if anybody who do that. Everybody that I know is in cloth code or open claw when they're messing around with that type of stuff. If you are using codeex, there is new plugins. And that is pretty much it. Updates on the road to the super app. Interesting. We'll keep you posted every week here. And let's see what else we have this week.
There's a few more interesting ones. Starting with the Google releases. So, they added a bunch of new models. And usually new model releases m not that interesting, but Gemini 3.1 Flash Live brings kind of a new experience to using Gemini because it's so darn fast and it works with voice mode. So, basically, this is an upgrade to Chat GPT's advanced voice mode. you know, the mode where it talks to you and you talk back and it talks back again. But inside of Gemini and now it's faster than ever with a larger context window, so it remembers more. But here on my phone, there's a button and now I can talk to Gemini. Hey, what up? Give me a comprehensive intro of yourself and what you can do for me. >> Sure. I'm a large language model trained by Google. I can understand and generate text. >> All right. Can I interrupt you easily? Okay, that was pretty good. Not bad. Now respond instantly. You bet. I'm ready when you are and I'll do my best to keep the pace going. >> Okay, great. So, basically, I can share video and stuff like that, too, right? So, if I give video access, what's up? I'm just filming right here. What kind of setup is this? Uh, what platform do you think I'm recording for? >> That's a pretty professional setup. I see a microphone, camera, and lighting, plus a script or notes on your laptop. The layout suggests you might be recording for a YouTube channel, online course or >> nice. It got that well. So, it is snappier. Definitely faster than the previous one. Very comparable to chat GPT. But with the various features like sharing your screen, sharing a camera, it being fast, larger context window, but they've been kind of quiet about this category, right? We had all of these fancy demos out of Chat GPT with the voice mode. Did all the fancy demos out of Google about the voice mode, but the stuff didn't really ship. But now it's coming together with the new model, ability to share video and all that stuff. larger context windows. Gosh, it almost seems like Google is also trying to build AI super app. Who would have thought? Who would have thought? Not me. So, that's Gemini 3.1 Flash Live.
Now, also, they released a family of open- source models, which I'm actually becoming increasingly interested in because, by the way, I have this new MacBook here, and this is me laughing because, you know, the broke 10-year-old kid living with a single mom moving around all the time kind of looks at this and is like, "No way." But it's the maxed out one. And I don't say that to brag. I say that to show you what you can get for $8,000 today. Apple M5 Max 128 GB and it can basically run 200 billion parameter models. This is super slow with models like Kimmy, but if you run something like Quen 3.5, gosh, it's faster than Chat GPT and the quality of it is actually close to chat. Anyway, I just briefly wanted to talk about that. Maybe I'll make a dedicated video about local models and this machine soon. If you guys are interested in that, leave a comment if you are, by the way, or leave a thumbs up on a comment that says it is. It's kind of hard to gauge if people actually want to hear about that stuff. Anyway, Google released new open source models which could run on this machine. That's why I pointed this out. The models are the GMA 4 family of models. And I don't know much about these yet. Usually takes a few days for people to run their benchmarks and to get experience with these to make a judgment. But from the benchmarks, gosh, even the 31B models seem as capable as the Deep Seek 3.2 two models and almost as capable as the Kim K models which for many people is the golden standard for running them with their open claw Mac minis. So yeah, Google threw their hat in the ring. They have new models. Really impressive scores. If you use offline models, look into them.
Okay, and here's a bunch of quick stories that I wanted to mention, but maybe we don't need a full segment for them. So there's two music generators that came out this week that were updated. One of them is So 5.5. And honestly, I'm you know, I think Sununo is fantastic. It creates probably the best quality songs in the market. But the latest releases all sort of sound the same to me and others people seem to be sharing this opinion. They got so good previously that it's kind of hard to top it. It's similar to the majourney situation. Cool, new release, but the bar was so high that the new bar is kind of at the same level. Their main focus was more on personalization and user control. So if you use that tool, check out the new version of it. Google has their own horse in this race. It's called LIA and they updated the newest LIA LIA free pro that can create up to three minute tracks. In terms of quality, it's definitely the main competitor to Sunno. The fun thing is they're starting to integrate this into other tools. For example, Google's Gemini. You can use it right in Gemini. You can ask for a track and it just creates it, which again, AI super app to bring everything together. That's what's happening here. Here's a quick snippet of a track created in LIA. Trade in the ocean for a frozen high. Slide the rail, feel the wind under the flipper. Slide the rail, feel the wind under the flipper. >> Compared to the same track created in Suno >> frozen air in breath. If he lands it or face plants, he's still the coolest.
Next, there's a claw code update where they give it the ability to use your computer. Yep. Basically, it can click buttons on your computer now, which goes beyond what it was able to do before. Huh. Almost seems like they're also gunning for this idea of an AI super app, but at a deeper, way less userfriendly level than Chacht is. So, that's that. Usually, computer use is still so flimsy that you can't fully rely upon it. And also, usually it works well with things like sheets, but then you do have to add a loop for it to double check what it did, which also doesn't always work. Worth the experimentation and headaches in some cases. Then there was a leak around this claude model called Mythos. Myth mythos. It's one of those two pronunciations. And I don't even know today. Some people were saying it was just a early April Fool's joke cuz there was also a Claude code leak that was April Fool's joke. I'm not sure. I think they are definitely working on another model. This is supposed to be the successor to Opus 4.6, their best model right now. It's basically just a rumor, but I wouldn't be surprised if we see a new flagship claude model within the next month, which this was supposed to be one. There's also Claude Tools on mobile. We tested these. We wanted to create a dedicated video around these, but from testing them, we arrived at the same conclusion that we get every time we try these connectors and tools and apps. And it's just the fact that they're not reliable. They get the job done once. Second time something happens, then it works again, then it doesn't, then it forgets something. It's just like working with a person that every time you tell them something, there's like a 40% chance of them forgetting it or doing something different. That's no fun. And neither are these claw tools on mobile. Can't recommend this wholeheartedly. And then Microsoft Copilot added a council and critique feature which I thought was worth a quick mention. Basically the council is something we already saw in Perplexity's product. We talked about this in a previous episode where it consults multiple models and takes the best of those and presents it to you. This is an advanced workflow that some people are successfully using for various things including investment advice in some cases. But at the end of the day it's not as simple as just asking multiple models and then throwing it in somewhere and saying hey summarize this. For example, Microsoft Copilot takes an approach where it's two AI partners. One is optimized for deep exploration and a second one is there to validate all the claims, improving the presentation and strengthening the structure. Creates a feedback loop that is runs through many times. It's kind of a cool product that I would expect we'll find in all of these AI super apps eventually.
And that's really it. Another week on our journey to AGI, AI super app, whatever you want to call it. That's what we're moving towards. And this show is designed to keep you up to date and to keep you at the frontier of all of that. And as per usual, you don't have to be using all of these things, but knowledge is power and so is awareness and now you're aware of all of this stuff. And even though you might be spending your days in chat, that's completely fine. As per usual, I would push back heavily on this notion of, oh my god, you're falling behind so heavily in AI just because you're not experimenting with open claw and oh, you're not building your own command center and clot code. What are you even doing? No, that sells well. That makes a good YouTube video. But even some of these pioneers in terms of AI usage, even if they use their open clauses, most of the time they use them similarly to how people use Chetch. They ask questions, manage context, and then they keep working with those results. My hope is that between using this stuff regularly and you following the show, you'll be able to stay at the frontier, which hopefully should allow you to focus on the things that you do best and the people that you love the most. That's the true mission of why we spend time together here every week. And it's my pleasure to be your guide on this journey. All right, I'll see you next week. Uh-huh. >> One stop, one spot for the clues. >> Every week, new views. This AI news you can use.