AI News: Claude 2.0 Dropped, Grok Bot Voice & Claude Code Projects (+ more)

0:00 We have another huge week of AI updates. Enthropic is finally combining Claude chat and co-work. They also just released two new products called Claude Docs and Slides. And they announced a

0:11 new update to Claude projects that manages multiple tasks at once. Higsfield just went open source and allows you to generate images and videos with their brand new API and this is

0:21 going to save some people a lot of money. On top of that, we have a few more updates we're going to cover in this video. If you're new to this series, I break down the top AI updates

0:29 every single week so you can stay informed without feeling overwhelmed. So without further ado, let's dive right in. All right, so we have some new updates from Claude. We have a

0:37 completely new interface. Claude Co-work and Claude Chat mode are now merged together. There are a few new products called Claude Docs and Claude Slides that are now inside of the interface. So

0:47 I'm going to break this down right now and show you exactly how this works. Enthropic dropped this onbrand claude launch video and what it shows is a merging of the chat and the co-work

0:57 interface. One thing I want to point out is right now inside of my account I am not seeing this new change. So if you remember maybe one or two months ago they merged the chat and the co-work tab

1:08 into this simple chat bubble icon. So we now have claude chat as well as co-work and then we have claude code if I come right here. But I still see in this interface the chat and the co-work

1:18 toggle button. But I believe over the next couple of weeks, this is now going to be one option and we don't need to toggle between these because Claude should be able to understand what it is

1:27 we're trying to do and route it to either chat or co-work. And I know this solves a lot of the problems that you've probably had because for me, I sometimes run into the issue of whether or not I

1:36 should be using Claude chat mode or Claude co-work. And this should hypothetically solve the decision fatigue that I feel on a day-to-day basis when using Claude. Now, the next

1:45 thing that I want to point out is the new output section. So, this is where we could use Claude Docs, Claude Slides, and Claude Design directly inside of this interface. I do want to point out

1:56 we're going to see the exact same thing inside of the Claude desktop app. So, if I toggle on co-work, you can now see that we have the ability to select a specific project. We can then change the

2:05 approvals to manual, automatic, or skip all approvals. And then we also have this output section right here, which I'm going to break down right now. Now, let's say I want to create a document

2:15 from a CSV file. I'm just going to select docs. As you can see, this is toggled on now. And I'm just going to take a big demo client CSV file. And I'm simply just going to say, hey, I want

2:25 you to turn this into a document that I can edit. And then what it's going to do is pull up this artifact on the right hand side of my screen. And this is actually working through and building

2:33 this in action. And I could see it building this from scratch. So this basically looks like a notion document. We have a title up here. We have a date along with Brock Message. There is a

2:44 summary of everything included in this CSV file. It's even showing a quick little table here. Let's let this do its thing and then I will show you the full final output. All right, so we now have

2:54 this document right here that we can go and actually manually edit ourselves. So, this is very similar to like an Excel file mixed with like a notion document. I kind of really like the way

3:02 that this looks and this is technically a clawed artifact here that we could go ahead and share. I'm going to show you exactly how to do that later on in this video. Let's go through real quickly and

3:11 see how this looks. So, we have all of our different clients here. It shows the status of these clients based on the CSV file that I uploaded. And then it has all this information. It has a name,

3:20 company, email, and again, I could come here and change the status of all of this. So, if I want to change the status to pass due, I can go ahead and make that change right here. Another thing we

3:30 could do is I could just click this checklist button and turn this into a checklist. Very similar to notion, but this is now built directly into Claude. I could also create multiple different

3:39 tabs of this document. So, let's say I want to create a new one. And I'm going to call this one past clients. I could then just come here and make the changes myself, or I could just have Claw do it

3:49 itself. Now that we tested out the outputs, and I showed you the documents, let me show you exactly how the slides work. It's very, very simple, and we could even customize the specific design

3:58 system that we've configured inside of Claude Design if you've ever used that platform. So, what I'm going to do is I'm going to upload that exact same CSV file. I'm simply just going to come back

4:07 to Claude, select slides, choose a design style, and I'm going to click Bramble right here, which is one that I've preconfigured. And I'm just gonna say, "Hey, I want you to turn this CSV

4:16 file, these demo clients into a slideshow that could visualize all of this data and information for me." All right, so here is the final output. We now have this slideshow that it

4:25 generated using that design style that I crafted previously inside of Claw Design. So you could import any design system that you've created previously and begin using that right here inside

4:36 of this interface. So let me just quickly go through this. So obviously it's taking that CSV file data and turning it into a slideshow. I think that this is pretty solid. Obviously it

4:45 doesn't look like the most beautiful thing I've ever seen in my entire life. It definitely looks like AI generated this, but I think this is pretty cool. The fact that we could not only use

4:54 this, but I think this is pretty cool. Not only can we generate this inside of Claude without leaving the Claude ecosystem, but we could even come here and edit this very simply as if we were

5:03 using something like Google Docs. So, let's say I want to change the color of this text here. It's as simple as just coming and literally changing the color of it. I could edit any of the text. I

5:13 can add things. I could remove things. I can add an image to this as well. So, let's say I want to drag on this image. I can go ahead and do so. Now, once we perfect this to look exactly how we want

5:22 it to look, I could go here and present this. So, obviously, it's just going to pull this up in present mode. I could then read speaker notes. So, if I want, I could add different speaker notes if

5:31 I'm actually giving a real presentation. And then, let's say that I want to share this with a team member or share it with a different computer or something like that. What I could do is I could come up

5:39 to where it says share. I could either copy the link to the actual artifact or I could click on export. And there's a couple of different ways that we can export this. I could send this to these

5:49 different applications here, Canva, Gamma, Miro, or other different AI platforms. Or I could download this as an HTML web page, a PDF, or a PowerPoint if I want to edit it there. So, let's

6:00 just click download PDF. I'm going to save it to my desktop. And here we go. We now have that slideshow downloaded as a PDF that's very easy and accessible for us to access. Next up, we have a big

6:11 update. Higsfield just open source our platform. They now have an API accessible to the public that we can use. And this is much cheaper than paying for the subscription that we

6:20 previously had to use inside of the Higsfield platform. Now, I want to quickly talk about an analogy that should help you understand exactly how this works because I think this could be

6:28 a little bit confusing, but this should really help explain it even if you don't know what an API is. I want you to think of this new product as Costco for Higsfield. And what exactly do I mean by

6:38 that? So, let me break it down right now. We get access to use these different image and video models that Higsfield has natively inside of their platform, but we get them at a very

6:48 cheap cost. So instead of having to pay for these plans right here, we could pay as we go with the specific image and video models that we want to use instead of being locked into a subscription. And

6:58 you're probably thinking, Brock, can't I go straight to Cling, Seance, and all these other image and video models? And the answer is yes. you could use their API, but this is a way to use it for

7:08 even cheaper because Higsfield essentially acts as this wholesaler for these image and video models. I would imagine they're able to get these cheaper prices because they're buying in

7:17 bulk because they have so many users on the platform. We can now take advantage of these lowered rates that they get as opposed to going direct to the image and video model providers. And another

7:26 amazing part about this is we could access all of these in one place. So instead of us needing to go to each of these individual model providers, we can now go direct to the Higsfield API and

7:37 all we need is one API key that we could generate with Higsfield and we get access to all of them. It's almost like we have our own Costco card in order to go and get cheaper supplies and

7:47 groceries instead of us going and buying all these different products from other grocery stores that are much more expensive. With this one access card right here, we now get discounts on all

7:56 these different models. All right. So, now that you know exactly how the Higsfield API works, I'm going to show you how to set this up yourself and begin saving money. So, there will be a

8:04 link in the description to this page right here, which is the Higsfield API. And all I'm going to do is come here and click on create API key. It's then going to bring us to this page right here. And

8:15 this is where we could select all the models that we want to use. For example, we could use Cance 2.5 at a 15% discount. We could use Cling 3.0 at a 15% discount. And some models are

8:25 discounted more than others. So go and just take a look at which model gets the biggest discount and which one that you would like to use. Then what I'm going to do is come over to dashboard. So

8:33 right here we could select three models and save up to 50%. So I'm just going to click on this and choose my models. Now we have Cance 2.5, Cance 2.0, Cling 3.0, Cling 2.6, we have one 3.0 Prime, we

8:45 have marketing studio image. We have tons of different models that we could select for video and image. So we could toggle between these because we might only want to use image models and other

8:55 people might want to use video models for example. And if I click on this where it says compare price with others, we could see how much cheaper this is compared to other platforms like this.

9:04 So compared to FAL AI, this is a 15% discount on all these different models right here. And this is going to save us a lot of money if we're generating images and videos at scale inside of the

9:14 Higsfield API. Next up, what we need to do is claim the models that we want to use. Personally, I want to use Seedance 2.5 because I think this is the best video model on the market. I'm also

9:25 going to claim Seedance 2.0 because you get up to 4K resolution. And then for image models, I'm going to select marketing studio image right here. So, now that I have the three claimed, I'm

9:36 going to click on claim selected models. Click confirm and claim. And there we go. Next up, I'm just going to add my card here to get $15 in credits because since we do pay for this as we go, we do

9:46 need to have a card on file that it can bill. But there's a couple ways we could do this. We could autorenew a specific balance. Or we could have a set amount and once it hits that amount, it's not

9:54 going to allow us to generate any more images and videos, but then we don't continue paying and rack up a big bill. Since I added my card, it automatically topped up my balance with 15 free

10:04 credits. So, let me show you exactly how to add this to your AI agent. And I'm going to do this with codeex right now. We need to first create an API key. So, what I'm going to do is click on create

10:13 API key. So, I'm just going to give this a name. Click on create key. And then what I'm going to do is I'm going to copy this prompt that it gives us right here. And then I'm just going to bring

10:22 this over to and then I'm just going to bring this over to codeex. I'm going to select a specific project. I'm just going to paste this in. Now it's going to basically set this up for me. It's

10:30 going to ask me a few things after it runs this. I'm going to have to paste in my API key into a separate ENV file, but I'm going to show you exactly how to do that. So what I did was basically just

10:39 tell Codeex to open the ENV file for me. And it opened it up right here. So what I'm going to do is click on open in. I'm going to select default app. So I'm going to click open in. I'm just going

10:48 to use cursor for this. And then all you need to do is copy your API key that Higsfield gave you that you generated and just paste this in right here. I'm not going to do this cuz I don't want to

10:58 expose my API key. But once you add this, just simply save this file. Come back to codeex and say done. All right. So now what our codeex agent is doing is it's running the cedance 2.5 model via

11:10 the API. So it's actually going and generating this video. So let's give it a second. We should get an output here shortly. And then what I'm going to show you is the exact cost breakdown of how

11:18 much it costs to generate a 5-second video from this example. And here we go. Here is a Cance 2.5 model video output. This is a 5-second video, 720p, and it's in 16x9 aspect ratio. So, let's take a

11:30 look at this. Okay, so that's solid. That's obviously Cance 2.5 generating that and it's using the Higsfield API in order to do so. What I did is I just said, "How much did

11:42 this cost?" It looks like Codex said it was approximately $2.31 before any of the discount price. But in order to check the actual price, I need to go to the Higsfield console. So, it pointed me

11:52 there. Now, inside of here, if I click on analytics, this is where we could basically see our monthly spend and how much we're spending on all of these video generations. So, in total, I've

12:02 generated two videos so far and spent a total of $3.24. So, I believe that this was about half of that. So, it was about $160 per video that I generated with Cedance 2.5. And

12:14 that is for a 5-second video. So, I think that this price could still add up. But, we are getting a discounted price when using the Higsfield API. And this is a good solution to people that

12:22 want to generate videos inside of their own applications, and they don't necessarily want to have to pay for a Higsfield subscription because they're already paying for tons of recurring

12:31 subscriptions every single month. Grockbot just released a brand new voice mode inside of the platform. You can now speak with each of your individual AI agents inside of Grockbot. And I am

12:41 extremely excited for this release. So, voice is rolling out on desktop and mobile over the next couple of days. I will say I have tried updating my Grobball app on my phone a couple times

12:52 now, but I still do not have access to this, which is probably the thing I'm most excited about. I'm very excited to pull up a Grockbot on my phone, literally just speak with my AI agent

13:01 and have it go and orchestrate different tasks across the different AI agents I've configured inside of Grockbot. Now, if you want to test this out and you don't already have Grockbot downloaded,

13:11 there will be a link in the description to sign up and download this for the desktop app, but you could also download this on your phone as well. Now, let me dive right into Grockbot. And as you can

13:20 see, I have a bunch of different AI agents configured. What I'm going to do is come to my main orchestrating agent. And what this does is this coordinates with all of my other AI agents. And this

13:30 is the one that I spend the most time talking to. And it's able to go and delegate different tasks to the other ones. If you're new to Grockbot, this is basically how it works. So you could

13:38 have you at the top basically coordinating with your different executive AI agents. And then these AI agents are coordinating with the operators or I like to call them

13:46 specialists. They are specialists in a given task. and I have about like eight to 10 different agents running for me throughout the day. All right, so inside of Grockbot, we're now going to see on

13:55 the bottom right hand side of our screen this new voice icon. So all I have to do in order to trigger this is to click on this button here. I'm going to set myself to mute and just explain very

14:05 simply how this works. So first of all, we have our main agent right here that we're speaking with. We can speak with each of these agents individually. So for example, I have my brand partnership

14:16 demo agent right here. I could always go into this specific agent and have a conversation with this one instead of my main AI agent. So each of these we can conversate with. Then we can click on

14:25 settings and there's a couple of different things we could change. We can change obviously the voice of these AI agents which I've only played around with a couple as well as it shows this

14:34 cool little icon for the different AI agents that you have with that voice. This is cool because you could obviously have specific voices to each of these individual agents to differentiate

14:44 between them so they don't all sound the same, which I think that is pretty cool. As well as I think this is really great. We could change the speed of our voice agent as well. So, we could change this

14:55 up to 1.5x, which I've found very helpful since testing this out because it seems like they're very slow at communicating. I like my voice agents to speak very fast and get to the point and

15:06 when it takes forever to say something, it's just very annoying. So, I speed mine up to 1.25 or 1.5x. And then we could also select the language. So, we have a few different languages here. We

15:17 have about like 12 to 15, maybe 20 different languages. Or we could just select auto detect and then it's automatically going to detect which language you were speaking in. So, it

15:26 could then speak in that language back to you. Now, what I'm going to do is I'm going to take this off mute and I'm going to speak with this agent. Hey, what's going on? What's your name?

15:34 >> I'm your main agent. The one who keeps you pointed at the stuff that actually moves YouTube AI for non techies. School and Lumi. What's pulling at you right now?

15:44 >> So, as you can see, that is very, very slow and annoying. So, I'm going to turn this up to 1.5x and then go from there. Yeah. So, my main thing is I just have a lot of emails I need to reply to

15:54 tomorrow. >> That's exactly the kind of thing I root out rather than dig through myself. >> All right. So, that's my main agent. Let me go to a separate agent now and test

16:02 that one out as well. Hey, how's it going? >> Hey, things are moving. Rivermark's contract is still sitting unsigned on five AI audits worth 21,000.

16:10 Brightfold's three projects are unsigned at 24 and Stackmir has been stalled on 148 for a while now. Want me to nudge any of those today or are you focused somewhere else?

16:18 >> Yeah, I think uh if you could nudge the first one, that would be great. >> Drafting a nudge for Rivermark on those five audits. >> Can you let me know when that nudge is

16:27 sent? I'll flag you the moment it lands. Anything else you want queued up while that one's cooking? >> Yeah, I want to see if you have the

16:35 ability to use computer use. Is that possible? >> Yeah, I can get into your computer, browser, files, mail, all of it. That's how I'm pulling your inbox and drafting

16:44 those replies right now. What do you need done? >> Yeah, I want you to pull up I want you to pull up the actual computer use screen and go and try to log into my

16:52 Instagram. I can open the browser and head to Instagram, but I won't type in your login. >> Perfect. Let's do this. >> Open on Instagram's login page. Nothing

17:03 typed in. You'll need to handle the sign in yourself. Want me to watch the screen while you do it or move on to something else? Instagram's up on the logged out wall. Username and password fields. Log

17:14 in. Forgot password. Login with Facebook. Create new account. No signin happened. >> Thank you. That is all. Okay. So, that's pretty cool. Obviously, I sat here and I

17:24 told it to go and use computer use to go and log into my Instagram or at least try to sign in and it went and used the computer to pull up Instagram. That's pretty cool. Not only can it do that,

17:34 but it could also flag and basically delegate different tasks to my other specialized AI agents. So, I think that this is pretty cool, actually. Now, if you've never used Grock before, it's

17:44 pretty cool how we can connect the different applications that we use. So, for example, I'm having it go and pull from my email inbox, and I'm able to use voice mode in order to go and pull that

17:54 information. So, it is almost like we have our own Jarvis that can work across the different applications that we use. We're able to do this with the plugins inside of our Grockbot. So, for example,

18:04 I have 24 different applications configured, which I could go through and scroll through all these. I have X, I've got Granola, I have Gmail, Clay, ClickUp, all these different platforms

18:14 that I've connected to. But let's say that there's an app I want to connect to that I don't see right here. Well, you could do that with the Zap Your MCP. I personally have this Zap Your MCP server

18:23 configured. And for example, I use Drip for my email marketing campaigns. And there's no plugin inside of Grockbot. So there's no way for me to connect Drip inside of Grockbot unless I'm using

18:33 something like the Zapier MCP. So now what I could do is I could come here and toggle on my voice mode. And I'm just going to ask it to go and pull information from Drip since it's using

18:43 Zapier to bridge that gap. Hey, can you go do me a favor and pull up my latest campaign data from Drip via the Zapier MCP? >> Pulling your latest Drip campaign data

18:53 through the Zapier MCP now that's in motion. I'll have the numbers as soon as they come back. Anything specific you're hoping to see in there or just the full picture?

19:03 >> No, I just like to see kind of the main details and main statistics there. >> Main details and stats it is. That's exactly what the poll is set up for. Should be landing shortly.

19:15 Now, I will say this is probably going to take a little while for this to pull this information because it does need to do a bunch of tool calls on the back end because you can see at the bottom it's

19:23 saying connecting to Zapier with our little agent icon dancing. Man, I just really like Grockbot. It's so simple to use. I love the interface and I honestly see myself using this more and more

19:33 every single day instead of using Claude because Claude used to be my main driver, but I'm using Grockbot quite a bit now. >> 260917

19:41 Claude 2.0 0 subject claude just added docs and slides went out yesterday at 3:34 p.m. Lisbon time to subscribers from before September 9th. About 14,831 active subscribers.

19:53 >> Perfect. Could you actually >> One heads up, open and click rates aren't coming through this drip API. >> Perfect. Can you actually just paste that information inside of the chat and

20:02 not give me a full lecture here? >> Latest drip send yt-260917 claude 2.0. Subject: Claude just added. >> All right. All right. So, it's not very good at listening to instructions cuz I

20:14 said I wanted it to quit talking, but I wanted to just send a text overview. But, it did actually pull this information from Drip via the Zap Your MCP. So, that's pretty cool. The fact

20:23 that we could have it work across the different applications and give us these overviews and also go and perform actions across them as well. Next up, we have an update to MetaMuse. If you don't

20:32 know what Muse is, this is Meta's new personal AI assistant that is very similar to Grockbot. You can now use Muse on the Mac as well as this can work across your computer. So it could

20:42 organize your download folder. It could find a file that you've lost track of and summarize your messages and notes. So it now has access to different folders and files on your computer just

20:51 like Grockbot, just like Claude Code, Co-work, and all these other AI agents. If you guys want to try out Muse, there will be a link in the description for you to do so, as well as I broke down

21:00 exactly how to use it in last week's AI news episode if you want to check that out as well. Next up, we're going to talk about this one briefly. The openi just dropped chatubt law which you can

21:09 now leverage the power of chachet Astra 6 for specific law use cases. Right now this is only available to specific law firms. So if you're not a law firm this probably does not apply to you but this

21:22 is interesting that chatbt is now getting involved in specific lawyer tasks. And if you take a look at this demo right here, you could actually see this model, GPT6 Astral Law Ultra, which

21:33 is the first time we've ever seen any taskspecific model. So this is an interesting product release not only if you are in the law space, but also for adjacent industries like Chachevt

21:43 Finance. I can imagine they're going to drop a similar model. For example, ChachevT6 Astra Finance. I'd assume that they're probably working on a similar model that is specific for that sector

21:53 and industry. Again, this doesn't necessarily apply to you if you don't work inside of a law firm, but I think this is just an interesting product release and maybe kind of a direct road

22:01 map of where OpenAI is headed with these specialized models based on given tasks and industries. If you want to learn more, I will link the article below in the description. So, you can go ahead

22:11 and check this out. It's pretty detailed, but yeah, pretty interesting. Now, for our next update, Enthropic released a brand new way to use Claude projects inside of Claude Code. So, let

22:20 me break this down very simply how this works. So as you see in the interface inside of cloud code you now have a projects button and inside of that project what we could do is have

22:29 multiple different instances of claude code running. So we could give it a specific task and it will go and route that to a different sub aent. This way we could run multiple tasks in parallel

22:39 at the same time with inside of one given project. Then what this does is it separates them into different threads. So multiple instances can run at the same time. For example, you could send

22:48 off one task right here, and it's going to go and create multiple different conversations in order to get this done. This is what it's going to look like inside of the interface. You can see

22:56 that there's three different ready to preview threads running at one time. And this is from one given chat. So, basically, you could send off this task to Claude and it's going to work on

23:05 these simultaneously. There are a few caveats at Enthropic made. So, let's go through that right now. Right now, projects inside of Claude Code are starting with limited cloud beta access.

23:14 I personally don't even have access to this inside of my account. on so I'm not able to demo it for you guys. Anthropic said starting today updated projects are available in beta to select cloud pro in

23:24 Mac subscribers who use cloud sessions in cloud code and don't have any existing projects on the web or desktop which is pretty interesting to me. I don't fully understand this because most

23:34 people probably do have existing projects on the web or for example inside of the desktop app which I personally have going. Now here's an example of what this interface is going

23:43 to look like. So, inside of this project, we have the ability to chat with this in one conversation. And right here, it shows start a thread that reads what customers are saying about

23:52 scheduled reports. And then it pulled up this specific thread here called launch feedback. Then on the right hand side of the screen, we're going to see a couple things. We see the library with all the

24:01 different artifact files. You can see all the document files. We could add specific files as well. Without having access to this, I only have limited perspective on this. But I will say it

24:09 does feel like Claude is trying to make it much easier to use Claude in one simple chat interface. If you think of platforms like Grockbot or Metam Muse, you can basically speak with one

24:21 individual AI agent and have a persistent conversation over time without you having to go to different chat threads or claude code instances that you have open. So, I think this is

24:32 a pretty interesting release. I'm very excited to test this out myself. And if you guys want to do more research on this, I'm going to link this article that Enthropic released right here. So,

24:40 you can do your own due diligence and see exactly what you think about this. I do want to preface and say this article is pretty vague. It doesn't really give us too much information. So, I guess

24:48 we'll have to wait and see once we get access to test this out to to fully understand the capabilities and how this functions. And there we have it. That is this week's weekly AI update. I do these

24:58 every single Friday breaking down the top releases for you to stay up to date as a non-technical person. So if you want more episodes like this, make sure to subscribe for more. Thank you for

25:07 staying to the end and I will see you in the next video. Cheers.