So Much AI News: New ChatGPT Tools, NotebookLM Updates, Big Claude Changes, Grok Bot Live, + More! transcript

This was a busy week in the world of AI. We got a lot of updates that you probably missed and it's all really exciting stuff. We got some major updates to Claude that make it well better. New features for Notebook LM coming, live voice for Grockbot, updates to lots of chat GBT features, and a way to get 1 billion Metamuse tokens for free. We're going to talk about all of that and a whole lot more in this video. My name is Paul Jay Lipsky and every week I make a video just like this one where I filter out all the noise and just talk about the AI news that actually matters.

And I want to start by talking about Claude because they just made a very interesting announcement. They've said that Claude Co-work and Chat are now merging into one. So, as I'm sure you know if you're a user of Claude, if you click on the top left here, there's two choices. There's chat and there's Claude code. But inside of the chat tab up here, there are two other choices. We have chat here and then we have claude co-work.

So, it's a little bit fragmented. And Anthropic is making this easier because very soon when you come into this, which I guess they're just going to call chat, it will not give you the option to choose between chat and co-work. You'll just ask your question, put in your prompt, and it will choose based on what you're inputting whether that's a chat query or a co-work query. This simplification is obviously a very good thing.

It's kind of confusing when you have to switch back and forth between chat and co-work and usually people end up just leaving it on one. I know that's certainly the case for me. I just left it on co-work all the time. So, it' be great to be able to use regular chat without actually having to tell it to use regular chat. They've also said that with this roll out, everything you already have set up is not going to change.

So, if you already have projects set up, for instance, none of those are going away. The experience should feel almost exactly the same except you don't have to switch between the two different options. I wouldn't be surprised if OpenAI follows suit on this pretty soon because they also have a slightly fragmented experience where inside of chat GBT you can manually switch between chat and work. Now, a lot of people have complained about that.

I actually think that for the initial roll out of these products that was a good idea to separate between chat work and even code because that gives consumers the the chance to learn about all the things that these tools can do. If all of this just rolled out as one product, I think there would have been some confusion as to who should be using this. Uh, originally these tools were really heavily focused on coders and so I think a lot of people would think that these weren't accessible to knowledge workers and it couldn't do everyday work that a knowledge worker would have to do.

But since we had the opportunity to play with co-work as a standalone product, we now know, oh, this is what it can do. And when it all comes together, it's going to be a lot easier for us to understand that. So, I actually think it was a good thing, but it is time I think for all of these different fragmented features to come together. The other exciting thing about Claude is that they've added a few new features in here.

So, if you click right here where it says output, you will see docs, slides, and design. I'm particularly really excited about design. Claude design has been one of my favorite products for creating things like motion graphics or websites. And I just find it's so easy to use and creates beautiful outputs, but it was always a separate website you had to go to. So now that it's built into Claw, it just makes a lot more sense.

So I don't have access to the full experience where everything is combined, but I do have access to all three of these. So I want to quickly try this out to see how this works. So I'm going to select design here, and I'm just going to give it a prompt that I ordinarily give it for different videos that I'm working on. So, first I can choose my design system. Oh, where are my design systems? That's weird. Why can't I choose my existing design systems?

I can open them up and I see my design systems here and I can open my design system inside of Claude design, but I can't use them in the new Claude. That's a real bummer that we don't have access to our design systems. If you're not familiar with design systems, these are templates for the look of different things that you create with Claude design. So, let's say you frequently make websites in Claw Design and there's a certain look you want all of your landing pages to look like.

You can create one of these design systems so that every time you create one of those landing pages, you don't have to start from scratch. It knows exactly what you want your landing page to look like. They're really useful, but apparently all the ones that I made with Claw Design, I can't use inside of Claw, the new Claw, which is a huge bummer. But let's try it out anyway. So, this will be a simple one. I'm just telling it to create a motion graphic where it takes the clawed logo and words mark here and just has it appear on the screen kind of just animating that image.

All right, so here it's done. You can open them up by clicking here obviously where it says open and there's our design. I think it looks great. This is exactly what I would expect out of claw design. So all of that is nice. I just wish they had the design systems brought in. I also tried out while I was waiting for that one of these new artifacts which was the slides. So just a reminder the way that works is when you come over to a new chat you can click on output here and I just selected slides and I asked it to turn this announcement about this update into a presentation.

So this is what it created for me. You can see what that looks like. And what's cool about this is it's over here on the side. So you can come in here and actually make changes to this. So if I wanted to change any of the text, everything can be edited. I can move anything around. And I can even present this by clicking here where it says present. Or of course you can download it as a PDF, a web page or a PowerPoint file.

And then finally, if you choose as an output docs, then when you generate it, you'll see the experience is very similar where it creates the dock over here on the right. What's interesting I've noticed is that it will create the outline for the dock first. You can see where its cursor is and then as it thinks through and plans out the document, it will begin to fill in each of these sections. There you go. You see it's writing it out right there and then it'll move on to the next section.

And just like with the presentations, you can come in here and edit anything. So I could like bold this for instance or add text or delete text. So very interactive. This feels like a full circle moment. For those of you don't know this, a little over a year ago when I started using Claude for the first time, you used to be able to create these artifacts very similar to this where on the right you could have a document that you and Claude could work on together and you could change any of the text in that document and then ask Claude to make any changes to it.

Just kind of go back and forth. They took away that ability and it really made me mad because it was one of the best features about Claude and there was no reason for them to take it away. So I'm glad to see that that is basically back now. So overall I think these new features are very well implemented and will make Claude a lot more usable. Before we move on, let's quickly hear from today's sponsor, Whisper Flow. Whisperflow is one of those tools that has completely changed the way that I use my computer and my phone.

Whisperflow is a voice to text tool that works anywhere you can type. You just hold down one button, it listens, you let go, and it writes exactly what you meant to say. For instance, let's say I'm writing an email. These days, I just hold down one button and talk. Hey, Catherine, I'm just touching base about the trip next week. We still need to book a hotel. Um, let's try to keep it under uh 150, actually $200 a night.

And make sure there's a what's it called? Early check-in. Thanks, Paul. Notice how it removes the ums, fixes my corrections, and formats the email. They also have snippets, which allows you to save long text you often write and drop it in instantly just by saying a short phrase. For instance, PS, if you want to check out my YouTube channel, you can find it here, YouTube snippet. Notice how we replaced the words YouTube snippet with the actual link to my YouTube channel.

My favorite part about Whisper Flow, though, is that it works the same no matter where I use it. Hey, if you want to check out my um YouTube channel, you can find it right here. YouTube snippet. So, you see it works exactly the same and even carries over my snippets. Whisper Flow is free to use, but you can get one month of Whisper Flow Pro with unlimited words by clicking on the link in the description or using promo code Lipsky in your plans and billing settings on desktop.

And now back to the news. Oh, and there's one more update for Claude. Sorry, the news moves quite quickly in this space. And this literally just came out as I was editing this video. What they said is that now you'll be able to chat with Claude inside of one task and then it will create these other tasks to actually delegate that work out. So you can just use one of them as a sort of chief of staff. This is currently rolling out for Clawed Code, but it says it's starting in Claw Code.

So, this seems like this will be available inside of Claude chatcowork in the future. There are also some updates to Gemini notebook that are coming out that I'm very excited about. This is something I've been asking for for a long time. So, the first one is the ability to record directly into a notebook from your phone. Currently, when you open up Gemini Notebook on your phone in the app, there's the ability to add sources.

And there is an option here for audio, but you have to upload existing audio. There's no way to record audio directly into Gemini notebook, but that is changing. You can see right here, you will have a microphone button and then you can record directly into the notebook. This would be great for students who have to record lectures or really anyone who just wants to jot down ideas on the go, maybe record meetings. This is something that I've been saying for probably close to a year.

I wouldn't love to see inside Gemini notebook. So, I'm very excited that they're adding this one. The other thing that they are adding is the ability to have live conversations with your notebook. You can see that right here. You can click on the live chat button and it will listen to you and talk back very similar to the way it works inside of Gemini, but now it's particular for your notebook. So, that will be an excellent addition.

I think this will help people be able to learn the material inside their notebooks a lot faster and a lot more naturally. And all this should be rolling out over the next few weeks. Google also announced two new live voice models, Gemini 3.8 Live and 3.8 Live extended thinking. So with 3.8 Live, it's supposed to be able to handle interruptions better. You're supposed to be able to transition languages easily and it should understand visual context better.

And with the extended thinking model, it will both think and speak in parallel, leading to a lot more of a seamless experience. So, I was using this all day yesterday and noticed some big improvements with it. Let me just give you a quick demo of it, though. Hey, Gemini, can you hear me? >> Yes, I can hear you clearly. >> So, um I'm going rock climbing with some friends on Friday. I've never been before. This is like with the one with the ropes.

Uh what should I know before I get started? For your first time, you'll be doing top rope climbing. >> Oh, I should say I also have uh some shoes already. They were given to me by a friend. Some climbing shoes, but they're a little tight. >> Climbing shoes are designed to fit very snugly to give you stability on the wall. >> Yeah. So, the interruptions work a lot better. Just the whole back and forth experience feels very, very natural.

And like I said, it's also a lot better at visual context. So, I was doing this yesterday. I was asking it how to fix something. and when it told me apart, I would put my hand on it and say, "Do you mean this part?" And it said, "Yep." And knew exactly where my hand was. So, all of that worked great. But comparing Gemini against what has been my favorite voice chat, which is the one from Chat GBT, I still feel like Chat GPT feels a lot smoother and more natural.

Gemini 1 [snorts] is great because it's fast and it's definitely a step up from what we used to have with Gemini. Here's another interesting new tool that Google announced this past week. This is CC. Now CC itself is not new. It's been out for months now as a Google Labs experiment. It was this tool that would email you every morning and tell you what's on your calendar and the important emails that you have to look at for the day.

It's kind of like a morning report, but now it's more of a family agent. So you can connect up to five members to your CC agent and you'll still get that your morning ahead type email, but this time it not only includes your calendar, but everyone's calendar from your family. So it's everything you need to know about your family coming into this morning report that's going to be emailed to you every morning. It also creates a shared Google calendar and Google tasks and you can coordinate in Google chat with CC.

So you can talk with it in Google chat and give it relevant tasks like tell it to put together a weekly meal plan or put together a shopping list for school supplies. And it looks like you can even delegate paperwork to CC as well. I think this is pretty neat and solves a real problem. You know, a lot of times parents, they will put things on their calendar and then the other parent doesn't see it on their calendar. So, being able to bring an entire family's calendar into one place and get a morning report like this, it's really smart.

So, if you have a family and you're having trouble keeping everything organized, check out the new CC. It's at labs.google/cc. And let's just finish up the last bit of news from Google. A couple of small things. So, the first one is that Google Flow, which is a tool you can use to generate images and videos, they have released the iOS app. They've had the Android app out for a while, but now they have one for iOS. So, if you really like using Google Flow, try using it from your phone.

It makes it a lot easier to do things like take photos and just add them into Flow. And then finally, if you're a fan of Gemini, make sure to come over to connected apps and see what's new here because they are adding more connected apps here all the time. The newest one that I just noticed is Whisper. So, this is from the same company as Whisper Flow, and they have a note-taking tool as well. So, now you can bring your notes from that note-taking tool uh from the meeting note-taking tool into Gemini with this connector.

And you can see there's bunch of other ones as well. and I know they are adding more. Next, let's talk about what Open AI has been up to. And we're going to start with what wasn't released this week. We did not get GPT6 Soul, which is something that feels imminent at this point. So, just to sort of recap this because this is getting a little bit confusing, here is the current layout of the models that we have from OpenAI.

We have Luna, which is the weakest model. We have Terra, which is supposed to be like a middle ground. and then soul and then we have Astra on top of that. Astra is the most powerful model. So that's four different levels of models, which is admittedly a little bit confusing, which is why when you switch over to regular chat, you really only get to choose between two different models if you click all the way through, which is the current model available in chat, which is 5.6 Soul, or the last model, which is 5.5, and that one's going away soon.

So really, you don't get a choice with regular chat, which makes sense. They're trying to keep it as simple as possible. The only thing they really want you adjusting is the thinking level right here. But inside of work, you got a lot more options. You can choose between all four of those different levels or even 5.5 if you wanted to. Now, like I said before, GPT6 Astra is the most powerful model available right now, but it's almost too powerful in the sense that it is too expensive and is burning through people's usage quite quickly.

I know for myself, I switched back to using 5.6 Soul for a lot of tasks. All my automated tasks now are using 5.6 6 soul and I'm only using GPT6 Astra where I really see a difference and where I'm willing to spend the extra usage on it. So, we need something to bridge this gap. People have felt the power of GBT6 Astra and they want something closer to that without the expense. So, I believe GBT6 Soul is going to be that.

And when it does get released, which should be in the next week, that is probably going to be most people's default. My bet is that it's going to be a very, very capable model. 5.6 Soul is already a fantastic model. And if we get a bump up from that, it's going to be incredible and a lot more cost-effective than GPT6 Astra. And my guess is I'm just going to use GPT6 Soul for everything. And I'm not even going to use Astra anymore just because I don't want to have to keep switching back and forth and it's just too expensive.

Next, ChatGpt Sites got a pretty big update. If you're not familiar with sites, this is a tool inside of chat GBT that you can create and deploy websites with very easily. So from within any chat, just type in the at sign and type in sites. This will be it right here. And then you type in the website that you wanted to create. And it not only creates a website, but it actually deploys it. So you'll get a URL that you can share with anyone.

It's live on the web. So you can use this to create landing pages, create a personal website or anything for your business or a game if you wanted to, whatever you want to make. So one of the updates is now you can share this with teammates so they can work on the site with you and you can also share it privately to only specific people without making it public to the entire internet. And you can add a custom domain.

So, that was probably the biggest request from people is that they wanted their own custom domain and that is now available. There's also been an update to chat GBT pets. Pets are these little avatars on the bottom. Well, they can be anywhere cuz you can move them around, but they stay on top of your screen. So, no matter what app you have open, it will let you know when the Chat GPT desktop app is finished with certain tasks, and you can return back to that task just by clicking on it.

So, they've added a few new things here. So, first you have the option to click right here to start a new chat directly from the pet or you can create a or start a live chat right here by clicking on this button. They've also added something called mini. So, if we come over to the settings and we click on the left here where it says pets, you'll see mini right here. This is like a lightweight version of pet. It doesn't have the full avatar, so it doesn't take up as much screen real estate, but it'll still show you the notifications.

You can still start a new chat, or you can start live voice. Personally, as much as I think the pets are super cute, I think I'm going to be using Mini, I just really appreciate how little screen real estate it takes up, and I don't notice it at all until I actually need it. So, I really like Mini. That's the one that I'm going to be using from now on. OpenAI has also announced chat GBT for Microsoft Word and Excel and PowerPoint for that matter.

So you can see here when you install this then inside of Microsoft Word you will have chat GBT over here on the right and you can ask it to do things in the doc for you and this works the same with Excel and with PowerPoint. You can also connect apps and build skills that you can use inside of the tools as well. All right, I have one Grockbot update for all of you and this is a coming soon. Grockbot will be able to talk with you with live voice.

It looks like this works pretty much the same way that chat GBT live voice or Gemini or Claude works where you can ask it questions, chat back and forth and give it assignments. And this is currently rolling out on desktop and mobile over the next couple of days. So maybe by the time you're watching this, you will already have access to it. I also want to talk about some stuff that's happening with Meta's Muse. I introduced Muse last week and said that surprisingly this may end up being the consumer AI agent that actually takes off because not only is it very good, but it's also very easy to use.

It's not as powerful as something like Grockbot or Chat GBT or Claude, but it's great and it's simple and it's free. Right now they have a little bit of a special going on where if you share your invite code with a friend and they use it within 48 hours of signing up for Metamuse, both you get 1 billion Muse tokens. That's on top of it already being free because there is a limit to how much you can use it every day. And after that, you do need tokens, but you can get a billion of them if you sign up for Muse and within 48 hours, go into your settings and enter in this code.

So, you can share your code with up to 30 people. I've already shared it with 11 different people. So, there are 19 uses left. And I am going to post it. So, there may be even less left by the time you're watching this. So, if you want 1 billion tokens, make sure to sign up for Muse. And in your settings, enter in this code within 2 days of signing up. And feel free to share in the comment section down below your code so that if all of mine run out, you can grab someone else's and you can both get 1 billion tokens.

There's also a rumor that very soon Muse will have the ability to make phone calls for you built in, which will be pretty wild. So you can say to it something like, "Hey, can you call this business and make an appointment for me or order this food for me or whatever and I'll actually be able to do that." So stay tuned for that. That is coming from what I hear. So that's it. That's everything I got for you from this past week.

A lot of exciting news updates in the world of AI. What are you most excited for? I think the direction that Anthropic is moving with Claude, bringing it all together and bringing more tools into it is very exciting. I want to see OpenAI do the same thing. And I'm also really excited by those little updates to Grockbot and Muse. I want to play around with those once they roll out to really see how useful those are. And hey, if you enjoyed this, do me a quick favor and subscribe to the channel because I'm like this close to a half a million subscribers.

And I know it really doesn't matter, but I'm really looking forward to seeing 500,000 subs on my channel. Either way, thanks for watching. I really appreciate it. And I'll see you next week for the next weekly AI news recap. Bye for now.