15 INSANE Use Cases for NEW Claude Sonnet 3.5! (Outperforms GPT-4o)

Summary

This video showcases 15 innovative use cases for the newly released Claude 3.5 Sonet model, demonstrating its superiority over previous models like GPT 4. The host, a YouTuber with web design experience but no coding background, explores creating interactive dashboards, real-time object detection, animations, web apps, games, and more, all through simple prompts. The video also highlights the community aspect, inviting AI enthusiasts to join discussions and educational courses on the AI Foundation platform. The script emphasizes the ease of use and the powerful capabilities of Claude 3.5 for users to generate complex applications and visualizations without traditional coding skills.

Takeaways

• 🤖 The presenter introduces Claude 3.5 Sonet, a new AI model, and initially underestimates its capabilities but finds it superior to GPT 4 after testing.
• 🔥 The video showcases 15 diverse use cases for Claude 3.5 Sonet, highlighting its versatility in creating applications and visualizations without coding experience.
• 📚 The first use case demonstrates creating an interactive PDF dashboard with visual appeal and functionality, using natural language prompts.
• 🎨 Use case two features generating animations, such as a wound healing process, to visualize complex biological phenomena.
• 📊 Use case three involves creating visual diagrams and illustrations from static images, like a neural network flow, to enhance understanding through movement.
• 🌐 Use case four details the creation of web applications, such as a states of matter diagram, with dynamic simulations based on user input.
• 🎮 Use case five discusses the creation of games, from simple like Snake to complex 3D games, leveraging Claude's capabilities for entertainment and education.
• 🖼️ Use case six is about generating 8-bit or 16-bit styled images, showing Claude's artistic potential in creating retro-style visuals.
• 📈 Use case seven highlights creating 3D simulations and visualizations, such as a 3D solar system, showcasing Claude's ability to handle complex 3D modeling tasks.
• 📊 Use case eight focuses on data analysis, converting CSV to JSON and creating interactive charts with libraries like Plotly.js for better data representation.
• 🏗️ Use case nine is about creating system architectures for applications, like an SEO tool, to visualize and plan the structure and flow of data and services.
• 🔍 Use case ten involves using Claude for SEO, building a mini web application to analyze blog posts for keyword optimization and suggestions.
• 🕵️‍♂️ Use case eleven demonstrates building a real-time object detection system using TensorFlow.js, detecting objects through a webcam in a browser.
• 🃏 Use case twelve is about creating a card counting game to practice advanced reasoning skills, simulating casino card counting strategies.
• 🌐 Use case thirteen highlights creating mind maps in Claude, which are visually appealing and customizable for brainstorming content ideas.
• 📘 Use case fourteen discusses using Claude for deep learning research, creating interactive research posters from lengthy PDFs to enhance knowledge retention.
• 📝 The final takeaway is about creating prompt templates in Claude for specific projects, streamlining the process for future tasks and saving time.

Q & A

What is the main topic of the video?

The main topic of the video is to showcase 15 different use cases for the new Claude 3.5 Sonet model, highlighting its capabilities and comparing it to previous models like GPT-4.

How does the presenter describe their experience with Claude 3.5 Sonet compared to GPT-4?

The presenter initially had doubts about Claude 3.5 Sonet, expecting it to be similar to other 'claw models' and not very good. However, after testing it, they found it to be much better than GPT-4 or any model they've used in the past.

What is the significance of the statement 'this is the worst that this model is going to get'?

The statement implies that the current capabilities of the Claude 3.5 Sonet model represent its lowest point, and improvements will only be made from this point forward, suggesting ongoing enhancements and updates.

What is one of the use cases demonstrated in the video for creating interactive content?

One of the use cases demonstrated is creating an interactive dashboard from a PDF document, allowing for a more visually appealing and interactive way to view and learn from the information contained within the PDF.

How does the presenter create animations to visualize complex processes, such as wound healing?

The presenter uses the Claude 3.5 Sonet model to generate animations by providing it with a description of the process they want to visualize, such as the stages of wound healing, and then refining the prompts until the desired animation is produced.

What is the purpose of the AI community group mentioned in the video?

The AI community group is a platform for AI enthusiasts to discuss artificial intelligence, stay updated on the latest news, participate in live calls, and access educational courses related to AI.

Can someone with no coding experience benefit from the Claude 3.5 Sonet model?

Yes, the presenter themselves has no coding experience but was able to create various applications and visualizations using Claude 3.5 Sonet, indicating that it is accessible to users without a coding background.

What is an example of a web application created in the video using Claude 3.5 Sonet?

An example of a web application created is a states of matter diagram that dynamically reacts to changes in temperature and pressure, simulating the behavior of particles under different conditions.

How does the presenter use Claude 3.5 Sonet to create games?

The presenter demonstrates creating games by providing Claude 3.5 Sonet with specific prompts, such as creating a Snake game using React, and then refining the prompts to fix any issues and add features, like making the arrow keys control the snake.

What is the potential application of Claude 3.5 Sonet in the business context, as shown in the video?

In the business context, Claude 3.5 Sonet can be used to quickly create tools like an SEO tool that provides keyword suggestions, feedback scores, and ranking criteria for blog posts, enhancing content optimization and search engine visibility.

Outlines

00:00

🤖 Introduction to Claude 3.5 Sonet Model's Capabilities

The speaker begins by expressing initial skepticism about the new Claude 3.5 Sonet model, comparing it to previous models like GPT-4. However, after testing, they found Claude to be superior and emphasize that this model is only going to improve. The video promises to showcase 15 diverse use cases for Claude, including creating web applications and interactive dashboards from PDFs, and real-time object detection programs—all without any coding experience. The speaker also invites viewers to join an AI community for enthusiasts to discuss and learn about artificial intelligence, offering educational courses and live calls for further insight.

05:00

📊 Creating Interactive Dashboards from PDFs

The speaker demonstrates how to use Claude to create an interactive dashboard from a PDF document. They walk through the process of uploading a PDF about a language model technique called 'reflexion' and then instruct Claude to generate a visually appealing dashboard. The dashboard includes tabs for decision-making and a quiz feature to test knowledge. The process involves minimal coding and leverages Claude's ability to understand and manipulate the content for visualization.

10:02

🎨 Generating Animations and Visual Diagrams

This section covers the use of Claude to create animations and visual diagrams. The speaker first requests a visualization of the wound healing process in the human body, which Claude initially provides as a static image. After further prompting for an animated version, Claude delivers a moving representation of cells throughout the healing stages. The speaker also uses Claude to animate a neural network flowchart from a screenshot, adding movement to illustrate the transfer of information.

15:04

🌐 Developing Web Applications with Claude 3.5 Sonet

The speaker discusses creating web applications using Claude 3.5 Sonet, showcasing an example of a 'states of matter' diagram. They explain how to build the application by adding particles, adjusting temperature and pressure, and observing dynamic reactions. The application is interactive, with real-time updates and the ability to handle errors by simply pasting the error message back into Claude for correction.

20:06

🕹️ Creating Games with Claude 3.5 Sonet

The speaker explores the potential of Claude for game development, starting with a basic game of Snake created in React. They show how Claude can fix issues like non-functional arrow keys with a simple prompt. The video also highlights more advanced game creations by the community, such as a 3D Doom game and Tetris, demonstrating Claude's versatility in generating complex game logic and mechanics.

25:08

🎚️ Generating 8-bit and 16-bit Styled Images

In this part, the speaker uses Claude to generate 8-bit and 16-bit styled images. They request images of Pac-Man and Mario in different styles, showcasing the simplicity of the process. Claude provides the images along with the underlying code, allowing for customization. The speaker also experiments with a '128bit' high-resolution vector SVG of Mario, demonstrating Claude's ability to handle various levels of detail.

🌌 Crafting 3D Simulations and Data Analysis

The speaker discusses Claude's capabilities in creating 3D simulations and visualizing data. They highlight community creations such as a 3D solar system with physics and a 3D cube animation with bouncing balls. The speaker also demonstrates converting a CSV file into JSON for easier data analysis in Claude, using it to create an interactive dual-axis line graph with Plotly.js.

🏭 Building System Architectures with Claude

The speaker illustrates how to use Claude to create system architectures, using the example of an SEO tool. They show the process of generating a flowchart for the tool's architecture, which includes databases, external services, and API connections. This use case highlights Claude's ability to quickly conceptualize and plan complex systems.

🔍 Developing an SEO Tool with Claude

The speaker demonstrates building a mini SEO web application within Claude that allows users to upload blog posts and receive keyword suggestions and analysis. They show how to prompt Claude to generate the necessary code and interact with the tool to analyze a sample blog post, providing immediate SEO suggestions and a revised blog post based on the analysis.

🕵️‍♂️ Real-time Object Detection System with Claude

The speaker describes creating a real-time object detection system using Claude and Tensorflow.js. They show how to prompt Claude to generate a complete HTML file for the application and then demonstrate the application detecting objects held up to the camera, highlighting Claude's ability to produce practical and interactive applications.

🃏 Advanced Reasoning with Claude for Card Counting

The speaker uses Claude to create a card counting game to practice the skill used in casinos. They guide Claude through multiple iterations to develop a game that shows a random number of cards, asks the player to count the deck, and provides feedback. The final game includes a timer, random card display, and the ability to restart and continue counting.

🌐 Creating Mind Maps with Claude

The speaker shows how to generate mind maps in Claude for brainstorming content ideas, specifically for an artificial intelligence niche YouTube channel. They demonstrate Claude's ability to create a visually appealing and comprehensive mind map that categorizes different types of content ideas, making it easier to organize and plan video topics.

📚 Deep Learning Research with Claude

The speaker discusses using Claude for deep learning research by creating research papers and posters from lengthy PDFs. They provide an example of generating a research poster from a PDF on multimodal chain of thought reasoning in large language models. The poster includes interactive elements, expandable sections, and dynamic data visualization to enhance understanding and knowledge retention.

📋 Creating Prompt Templates in Claude

The speaker concludes by emphasizing the importance of creating prompt templates in Claude for specific projects. They demonstrate how to ask Claude for a prompt template that can be reused to generate a particular type of interactive poster from any PDF. This method saves time and helps maintain consistency across projects.

Keywords

💡Claude 3.5 Sonet

Claude 3.5 Sonet refers to a new AI model that the video discusses extensively. It is presented as a significant advancement in AI capabilities, outperforming previous models like GPT-4. The video's theme revolves around demonstrating the diverse and powerful applications of this model, such as creating web applications, interactive dashboards, and animations, all without the need for traditional coding experience.

💡Interactive Dashboard

An interactive dashboard is a user-friendly interface that allows users to visualize and interact with data. In the context of the video, the dashboard is created using the Claude 3.5 Sonet model by uploading a PDF and prompting the AI to generate a visually appealing way to view, test, and learn from the information. It is an example of how AI can enhance data presentation and accessibility.

💡Real-time Object Detection

Real-time object detection is a technology that can identify and locate objects in a video stream or image in real time. The video showcases how Claude 3.5 Sonet can be used to create a program for this purpose. It demonstrates the AI's ability to generate code for complex tasks, emphasizing the model's versatility and potential applications in fields like security or augmented reality.

💡AI Enthusiast

An AI enthusiast is someone who has a keen interest in artificial intelligence and stays updated on the latest developments in the field. The video script mentions a community for AI enthusiasts, indicating that the content is targeted at individuals who are passionate about AI and are looking to engage with like-minded people, share insights, and learn about new AI tools and techniques.

💡Web Applications

Web applications, or web apps, are software programs that run on web browsers. The video demonstrates how Claude 3.5 Sonet can be used to create web applications with interactive features, such as a states of matter diagram. This highlights the AI's capability to assist in software development, enabling users to generate functional applications with minimal coding.

💡Animations

Animations in the video refer to the dynamic visual representations that Claude 3.5 Sonet can generate. The script mentions creating animations to visualize complex processes, such as the wound healing process in the human body. This showcases the AI's ability to transform static information into engaging and educational content.

💡Visualizations

Visualizations are graphical representations used to communicate information clearly and efficiently. The video discusses using Claude 3.5 Sonet to create visualizations from static images, such as a neural network flow, to better understand complex concepts. This illustrates the AI's role in enhancing learning and conceptual understanding through visual aids.

💡Data Analysis

Data analysis involves examining and interpreting data to draw conclusions. The video script describes using Claude 3.5 Sonet for data analysis, specifically converting CSV data into JSON format and creating interactive charts with plotly.js. This demonstrates the AI's utility in simplifying data analysis and making it more interactive.

💡System Architectures

System architectures are high-level designs that outline the components, work flows, and interactions within a system. The video shows how Claude 3.5 Sonet can generate system architecture diagrams for applications like an SEO tool, indicating the AI's ability to assist in planning and conceptualizing complex systems.

💡SEO Tool

An SEO (Search Engine Optimization) tool helps improve a website's visibility in search engine results. The video demonstrates creating an SEO tool using Claude 3.5 Sonet that provides keyword suggestions and analysis for blog posts. This highlights the AI's application in digital marketing, offering automated and data-driven insights for content optimization.

💡Prompt Templates

Prompt templates are pre-defined sets of instructions or questions used to guide AI responses. The video emphasizes the importance of creating prompt templates in Claude 3.5 Sonet for efficiency and consistency when generating specific types of content, such as interactive posters from PDFs. This showcases the strategic use of AI to streamline content creation processes.

Highlights

Introduction to 15 use cases for Claude 3.5 Sonet model.

Claude 3.5 is significantly better than previous models, including GPT-4.

Creating web applications and interactive dashboards from a PDF using Claude 3.5.

No coding experience needed to create complex applications with Claude 3.5.

Interactive community and educational resources available for AI enthusiasts.

First use case: Creating an interactive PDF dashboard for better data visualization.

Adding interactive features like quizzes and graphs to the dashboard.

Second use case: Generating animations to visualize complex processes, like wound healing.

Third use case: Creating visual diagrams and illustrations from static images.

Fourth use case: Developing web applications, like a states of matter simulation, using prompts.

Fifth use case: Quickly creating games, from classic Snake to more advanced 3D games.

Sixth use case: Generating 8-bit and 16-bit styled images and visualizations.

Seventh use case: Creating 3D simulations and visualizations, like a solar system with physics.

Eighth use case: Comparing data using interactive charts and dual-axis line graphs.

Ninth use case: Creating system architectures and flowcharts for web applications.

Tenth use case: Developing SEO tools for analyzing and improving blog posts.

Eleventh use case: Real-time object detection using TensorFlow.js with Claude 3.5.

Twelfth use case: Creating advanced reasoning applications like a card counting game.

Thirteenth use case: Generating mind maps for brainstorming and content planning.

Fourteenth use case: Deep learning and further research applications, including interactive research posters.

Fifteenth use case: Creating prompt templates for repetitive tasks and projects.

Community engagement and networking opportunities for AI enthusiasts.

Claude 3.5's capabilities in automating business processes and enhancing productivity.

Emphasis on the importance of staying updated with AI advancements through community and education.

Encouragement to join the AI Foundation community for continuous learning and collaboration.

Transcripts

in this video I'm going to show you 15

insane use cases for the new Claude 3.5

Sonet model that was just released now

when this thing came out I was like oh

great another claw model this probably

isn't going to be very good but after I

tested it I quickly found out that this

is much better than GPT 4 or any model

that I've used in the past and the crazy

thing to think about is this is the

worst that this model is going to get it

only improves from here now we're going

to be doing things in this video like

creating web applications interactive

dashboards from a PDF I'm going to show

you how I created a realtime object

detection program all from Claude all of

this I did with Claude and I have no

coding experience all I am is a YouTuber

who has a little bit of experience with

website design now we're going to get

into those 15 use cases but first I

wanted to show you this because I know

if you're here right now you're

definitely an AI Enthusiast you like

staying on top of artificial

intelligence so this group is the

perfect thing for you I'm going to leave

this group in the description or the top

pend comment below and what this group

is is a community for this channel AI

Foundation

and we're in here all day talking about

artificial intelligence not only that

but we have a classroom filled with

multiple different AI educational

courses and in the calendar tab you can

see all of our live calls that we have

coming up you know Carter and I like

going live with the group and the live

calls have been amazing so far a lot of

interaction and a lot of unique Insight

found on artificial intelligence so if

that's something that interests you if

you love talking about Ai and want to

stay in the know and up to-date and

learn all the powerful techniques then

this is the place to be again it's in

the description below or the top end

comment now let's get into the awesome

use cases that collaud 3.5 Sonet has to

offer the first use case is to create an

interactive dashboard when using a PDF

so let's say I'm reading this white

paper on a large language model

prompting technique called reflexion now

this is kind of a complex topic but it

goes into a verbal reinforcement

technique in order to help large

language models have better responses

let's say I read this and now I want to

have a fun way of interacting with that

information I can upload that PDF by

hitting add content and then I can type

in a quick prop I said create an

interactive PDF dashboard to help me

view test and learn this information in

a more visually appealing way then I

said make it dark mode of course I'm

going to send that off and make sure

that you have the artifacts feature

enabled because what this is going to do

is it's going to open up a dual screen

view where it showing you the code that

it's generating for your PDF dashboard

and not only that but after it generates

the code we're going to be able to

preview and use this dashboard live and

we might get an error sometimes it

starts off with an error but it's very

easy to fix you just past the error and

CLA and it'll fix it for you but as you

can see nothing went wrong here as we

can see we have this amazing dashboard

that was just created based on the PDF

that I uploaded and that I was just

showing you we have different tabs for

decision making we have icons in here I

mean this is just beautiful reasoning

20% improvement over Baseline it really

helps you see this in a more visual way

I can go to programming and what I can

even do is I can add things to this as

we go I said add a tab for graphs

showing the examples and also a test

test tab where I can test my knowledge

with a 10 question quiz I can send it

off and it's going to keep the same

structure but just add onto the code

that I already have not that I already

have but that it already made me as you

can see we have two more tabs and the

graphs tab isn't currently displaying a

graph so we could just reiterate that

but we now have a quiz tab that has 10

questions that we can go through and

answer use case number two is to use

animations and create animations to

visualize anything I got pretty unique

with this one I said create a

visualization showing me how cells work

in a human body body in order to heal a

wound from getting injured to the wounds

being healed what it created me was this

wound healing process image but what I

wanted was an animation this was great

though I mean for 10 seconds just typing

out a quick prompt having this be

generated this is very very cool it

shows the injury the inflammation of the

injury the proliferation and then the

remodeling which is very cool you can

kind of see how that might go from a cut

to inflamed to Healing to remodeling now

this was pretty good I liked how it

showed the whole process and did it

quickly but I wanted it moving so I

asked can you make us a visualization

with cells actually moving around and

this is what it created me next an

actual visualization of cells moving

around within this little box and it

changed each part for me so it went from

injury to inflammation and as you can

see the cells are moving a little bit

faster to proliferation and then finally

to the remodeling stage but this still

is not what I was after right I wanted a

couple more things done so that's when I

just ask all you have to do is ask now

eventually I ended up just prompting

with Claude and getting this

visualization in this anim to look

something like this so that it looks

like somebody's actually getting cut

right here and so this would be the arm

process getting hurt and then we zoom in

on that cut I had it do all this just by

prompting as I showed you and now I have

actually inside the cut the entire

process with a progress bar and more of

a time-lapse formation which is just

amazing it even has some key points down

here in terms and then it zooms back out

and Loops the animation use case number

three is to make visual diagrams and

illustrations and

based on static images let's say I'm

reading this white paper on large

language models and I want to see this

in a different visual way what I'm going

to do is I'm just going to screenshot

this then I'm going to hit add content

and then just upload that screenshot of

this neural network flow and I'm just

going to say stimulate the movement of

information through this neural network

image using moving elements and react I

can send that off and based on the

static image it's going to create code

in order to stimulate the movement of

information through this static image

which is very cool it's going to create

its own little environment it's going to

use all the code show us a preview and

it's going to show us how information is

actually moving in the static image so

that we can see it in a more visual way

as you can see it says unsupported

libraries if it ever does that just copy

in your error code copy the entire

message and paste it in and it will fix

it for you as you can see it creates a

mock of the image and actually has

moving elements for how information is

passing through this but as you can see

I don't have any input values in here

and nothing's labeled properly so I'll

say add input values make sure

information is moving correctly and make

it bigger just simple English right just

simple natural language to this large

language model as you can see

everything's labeled now and we actually

have different lines that it's moving

through as you can see it's moving

through all these hidden layers based on

inputs and this is just amazing to watch

and it's cool that it can do all this in

a matter of seconds and you could keep

on going with this I mean you could go

as in-depth as you want you could make

it look however you want you could add

in multiple different inputs you could

really go through this and make it your

own use case number four is to create

web applications using Claude 3.5 Sonet

now this one's fun I created a whole

states of matter diagram showing how the

states of matter react when under

certain conditions let me show you how I

did that as you can see I literally

built this thing with a couple of promps

and clawed now this is just amazing I'll

show you the promps I used in order to

do this but here's the idea of it I

select a particle type let's say neon

and then I add my quantity let's say I

want a 100 of those particles in here I

can hit add and it adds it to the

simulation box of the neon particles

moving around and these things are

Dynamic so when I up temperature they

start moving faster if I update pressure

they go through this pressure line and

if there's too much pressure too many

particles and the temperature is too

high then it will explode because the

pressure is so much that it just blows

the lid off of this simulation and then

all the atoms go scattered we have a

dynamic information panel updating with

how many particles are within the

simulation this thing just gets pretty

wild so if I increase pressure once

again and temperature there they all go

seems like it be like very hard to code

and it would be if you doing it on your

own because if I go to the code I mean

there are 224 lines of code so yeah that

would take a long time and I just did it

in a couple minutes just prompting and

asking Claud some questions and whenever

there's an application error all I do is

upload it and it fixes it right away and

I ask basic questions like this when I

add particles I don't see them moving

nothing is there I want them moving

around and then it does that for me and

I just keep on going through perfecting

this until I get it exactly how I want

it and you end up with a final product

like this I mean this is just amazing

where you can create these little apps

there's an SEO tool that I'm going to

create a little later in this video that

I want to show you because that's an

example of how you could use it for

business use case number five is to

quickly create games whether that be

existing games or unique games that you

come up with or that you strategize with

and create with Claude let me show you

some examples of not only how to create

a basic game but also examples from how

some people on X used it to create more

advanced games yes we can do the classic

example of create the game Snake using

react and make it functional you could

even do this with python using vs code

I'm going to send this off and it's

creating the snake game in react for us

just like this and we'll be able to play

it within Claude which is very

interesting so now we have the snake

game in react but as you can see our

arrows aren't working so you can just

ask Claude and it will edit it for you

it just said make it so my arrows move

the snake and now as you can see my

arrows are actually moving the snake

after making that quick adjustment my

snake goes through the walls if it hits

itself then the game is over now that's

a basic game I've seen people even make

Tetris with this which is quite amazing

but this is one expost that I saw and

came across that I was just amazed by

created by this user he says I just

recreated a basic 3D Doom game using

Claude 3.5 Sonet with artifacts and just

three prompts this game is procedurally

generated Maps sound effects zombies

that come after you in a mini map to do

something like this took me two days

with GPT 4 when it first came out and if

you watch this it's just amazing it even

has sound effects as well but he's doing

all of this while in Claude in the

preview section he has a map he can stop

these things from coming at him by

hitting his Mouse button and he did this

with three prompts he created a game

like this that is just amazing so

congratulations to this guy and just to

show us what you could do in the video

game industry with something like Claude

3.5 Sonet use case number six is to

create 8 bit or 16bit styled images so I

could say something like create an 8bit

style image of Pac-Man when I send that

off Claude ended up giving us this fun 8

bit styled Pac-Man with him over here on

the right eating these white pellets on

the left next I asked it to do a Mario

photo on a white background and then it

gave me this 8bit styled Mario SVG which

is very cool and you can even see the

code that it did in order to do this

then I thought you know what let's do a

16bit styled Mario instead which gives

him a little bit more detail here you

can see the code it gets a little bit

more complex to generate this

SVG and then I did the 128bit Mario high

res resolution Vector SVG and there was

a lot more code to this and Mario turned

out looking very interesting I'm not

sure if this is even Mario at all it

kind of scared me but I do like the

16bit and the 8bit styled images a lot

so I recommend you give that a shot the

next use case for this new claw update

is to generate 3D simulations or

visualizations now for this one I just

want to take you back over to X because

I think there's some users there who

have done some amazing things with

actually generating 3D objects in a

single conversation so let's go take a

look at some of those

this gentleman right here ended up

making a 3D solar system with physics

and collision so if I open this up as

you can see he can hit add a random

planet and it adds planets certain

planets Collide it's very cool and I

love how all of this works Dan here

actually created this 3D cube animation

with balls bouncing around inside the

cube using claude's new artifact feature

he said this took him only 30 minutes to

do with all of these functions added but

that's a very cool example of showing

how he did that 3D bouncing ball

simulation here's another animation

created by Max zyell and it's another

particle Cube animation with a bunch of

different factors that they have the

ability to edit which is very neat and

this is just amazing at what Claud is

capable of doing now and here's another

example of somebody creating a 3D gaming

CLA I mean this just looks amazing and I

guess you could put this in 3D

simulations category this created by

coffee vectors said it took him 90

minutes it's crazy how it can

incorporate sound effects 3D elements

and so much more the next use case in

Claude is to actually compare data now

with this new artifacts update I found

that the data analysis side of Claude is

much more powerful here's an example so

if I just upload a data set I'll just

upload my sub based platform data set

this is a very very basic Google sheet

showing month revenue for that month and

social media post for the month and

first what I can do is I can just say

convert this CSV into Json and then it's

going to give me that Json

representation of the data just like

this and this is going to allow us to

actually create an interactive chart

using something from the JavaScript

library next one I'm going to do is say

create an interactive dual AIS line

graph of these two data points revenue

and social media post with the xais

being the month use a plotly.js chart so

we can hover over these elements on the

chart so it is interactive and then I

can send it off and because we have this

Json it will be able to create this

interactive dual AIS line graph much

easier I've tried this by just using the

CSV file and it definitely doesn't work

out as much so if you do want to do some

data analysis in clae 3.5 Sonet highly

recommend you convert it to Json before

you start doing this and this actually

looks very similar to chat GPT except

they do have their minor differences I

like how we get some image settings up

here where we can download the plot as a

PNG and we can take a snapshot of the

plot we can zoom out we can zoom in I

think this is pretty cool the next use

case for Claude 3.5 that I think it's

amazing at is creating system

architectures let me show you how that

might work for some something like an

SEO tool so I typed out a quick prompt I

just said create a system architecture

for an SEO web application that allows

users to upload blog posts get keyword

recommendations feedback score and a

ranking criteria right I just want that

flowchart that system architecture for

this web application that way we can see

it and Claude can see it better if we

actually go through and build it with

Claude if it's possible now I'm going to

send this one off and as you can see

it's going to get coding and it's going

to use mermaid in order to create the

system architecture which is very very

cool I really like how it breaks it into

the databases down here on the left we

have a user database blog post database

and a keyword database external services

and SEO ranking API backend client front

end so if you're trying to do

automations or you're trying to build

applications this is going to be huge by

building out these system architectures

in Claude it does it so quick it's

interactive and it's a lot better than

doing it in something like chat GPT

right now use case number 10 is actually

to use Claude 3.5 for SEO now we just

got that Mach system chitecture created

for a web application for SEO but now I

actually want to show you how you can

build a mini web application in Claude

3.5 use it in the preview section in

order to create an SEO tool that will

allow you to upload your blog post get

keyword suggestions and more so I'm

going to start a new chat and Claude and

I can just paste in this prompt that I

used earlier what I have here is create

an SEO tool that allows me to upload my

blog post industry and what keywords I'm

trying to rank for then I say after I

upload all that stuff I want to be able

to hit a button that gives suggestions

on what to change including a reset

button that starts to process over and

also add a plus button to keywords so I

can add multiple different keywords I'm

trying to rank for in the SEO

suggestions area after analyzing the SEO

give me specific numbers and stats for

my blog post as well as caution signs to

show what needs attention I can send

this off and within seconds it's gets

generating the code and this is just

amazing not only for blogs as I've been

saying but for anything you need a tool

for and take a look at this we now have

an SEO analysis tool and keep in mind we

can edit the styling but right now this

is pretty good for about 15 seconds of

generating something I'm going to go to

chat gbt and grab this blog post that I

have on artificial intelligence taking

over the world I'm going to paste it in

here in clae and now that I have this

blog post in here what I can do is I can

actually just use this tool within

Claude I don't even have to upload this

to some website or make an app out of it

I can just test it out in here what

industry I can just say technology and

then I can type in some keywords here so

maybe I want to rank for the keyword AI

maybe I want to rank for

artificial

intelligence and then maybe I want to

rank for something like large language

models something that might not be as

prevalent in this post in specific and

then I can just hit analyze SEO then it

says word count 800 it gives me an

analysis result keyword density 5.88%

artificial intelligence

0.25% large language model 0% so if I'm

trying to rank for large language models

I haven't even put anything regarding

large language model models within here

so this is very cool I can hit my reset

button this was built with one prompt in

about 15 seconds this is amazing this is

crazy I haven't even added anything as

you can see I added one extra prompt I

said change the styling change the name

and then I told it to add a revised blog

post section after I upload my initial

blog post and as you can see it has the

new name the new styling the SEO mro

with the wizard this is starting to look

like my own little app I can paste in

the text right now the text yellow but I

can make that whatever I want whenever

type in my industry technology type in

my keywords again then when I hit

analyze SEO what it does is it actually

gives me a revised blog post suggestion

changes around the blog post a little

bit says here's why I made these changes

it added llms to the beginning added

large language models to the beginning

it added some things that I was lacking

within my analysis result and again I

could change the styling but for two

prompts in a matter of a couple minutes

this is awesome this next use case I

want to show you is how I built a

realtime object detection system using

clae and I'm actually going to run it in

my browser and show you how this works

firsthand I'm just going to come back to

Claud and actually paste in this prompt

that I used earlier but what we have

right now is create a single HTML file

for a realtime object detection web

application using

tensorflow.js and this model I'm not

sure if those are zeros or O's the Coco

SSD model this application should and

then I just have a bunch of parameters

for what this thing should do in order

to detect objects on my screen then I

tell it to provide the complete runnable

HTML file with inline CSS and JavaScript

make sure I'm in Sonic 3.5 send it off

and this is just amazing what this can

do and how it can actually generate this

for me an object detection web

application it's hard for me to even say

that but you can generate this for me

and I'm going to show you how it works

and actually detects objects and what

they are real time using this Library

says error over here but watch this as

soon as I hit this little download

button the bottom right hand corner and

then I actually open up this file up top

what this going to do is pull me to this

screen and it's going to ask me if this

file is okay to use my camera I can just

hit allow this time and since I'm

actually using the video to record it

can't access the camera but I'll upload

some footage of what that looks like and

as you can probably see it completely

messed up my camera right now so this is

great as you can see it's detecting

objects right now and it's detecting a

person and when it detects an object it

labels it here this entire thing was

built with Claud remember downloaded it

let me hold up something to my camera

like a cup and see how it does look for

the detected objects down here

underneath my face when I hold up a cup

see if it gets it cup it thought it was

a cell phone at first but then it ended

up no actually it thought it was a cup

and then it ended up going to a cell

phone so this is very interesting let me

hold up a water

bottle bottle detected isn't that just

insane how it can detect objects let me

try holding up something else maybe I'll

try holding up my laptop and see what it

says can it detect what a laptop is

laptop first detected so this entire

thing was actually built using Claude

the download file I just got was built

and sent from Claude in about 15 seconds

using tensorflow and some other things

so I'm already starting to build up my

Tony Stark lab right this next use case

really demonstrates claude's new

Advanced reasoning I want to curate a

card counting game so that I can

practice counting cards let's just say

and I just to see if Claude can do this

I actually know it can do this because

I've done it previously in Claude 3.5

Sonet and it's just amazing how much

Advanced reasoning goes into this

because you have to think about all the

different cards in the deck it has to

remember which cards it showed me and

which cards did not show me again it has

to know the card counting system and

implement this all into one application

that I can use similar to a game but I

think card counting has a lot of

advanced reasoning especially when you

start dealing with decks of cards things

can get very complex so that's why I

want to show this one to you so this one

took a couple more props but what I told

it to do is make a card counting game

the card should show the correct number

in suit show the user a random number of

cards ask what count the deck is the

count should always start at zero if the

player types in the right count say good

job if wrong tell them the actual deck

count and restart the entire deck and

this thing kind of evolved over time

this is what it started as right and it

kind of looks okay and it just told me

to enter the count right away but I

wanted it to run through some cards

before I got asked the count of the deck

now if you if you're unfamiliar with

card counting it's something that people

do in casinos well I'm not sure if

they're supposed to do it but but you

can basically predict what's coming next

in the deck whether it's a low card or a

high card based on adding up a certain

number within your head from the cards

that are shown to you whether it's plus

one Z or minus one and so I said I want

ERS to only submit a guess after you've

showed them the cards and I wanted it to

show five or 10 cards before we got to

submit a guess and then this is what it

came up with something which is pretty

cool so now it's counting down but it's

only showing the same card right it has

this little timer that I asked it to

create at a 3se second Pace but now it's

just showing one card and not changing

so after multiple times of reiterating

that I want to show different cards

rather than the same card and ended up

coming up with this so now it's actually

changing cards right this is pretty

Advanced has a timer it's changing cards

it's not showing the same cards in the

deck which I think is very cool and it's

randomizing it's a different one every

single time so if I restart now it's at

nine if I refresh it's at six of

diamonds so it's starting with a new

card every time and it's randomizing it

and so this was pretty good and after

just styling it a little bit more I

ended up coming out with this final

result this took me about 20 minutes to

create the casino card counter I can

change my speed that I want so maybe I

want the cards coming in every 4 seconds

when I hit start game I can now do this

right I can now practice my card

counting and it goes on my 4C timer in

this fun styling and then I can just

make a guess after five to 10 cards are

shown and I'm going to show you what

that screen looks like in a second now I

haven't been counting while doing this

so I'm not sure what the count is right

now but I'll just make a little bit of a

guess I have no idea I'm just going to

put minus one and submit guess the count

was one okay so that's kind of cool that

I can practice like that so I can hit

continue and then just keep going

through the entire deck I just wanted to

show you that example in specific to

show you the power and the advanced

reasoning of Claude 3.5 Sonet the next

use case with clae 3.5 Sonet is to

create mind maps with it now chat GPT

used to be better at this in my opinion

using some gpts but after being in

Claude I realized how much better Claude

actually is at this currently I'm not

sure when GPT 40 is going to come out

completely when it does we'll see how it

Compares up against this but right now

mind maps and Claude are a little bit

more visually appealing I feel like I

can edit it more and I have a little bit

more fun generating them so what I said

in my collaud prompt bar was create a

mindmap brainstorming content ideas for

the artificial intelligence nich YouTube

I can send it off and similar to the

system architecture it generates this

huge mind map and it just looks awesome

I can even edit this if I'd like but we

have ai YouTube content and it breaks it

up into more categories tutorials

explainers interviews news and updates

case studies debates and discussions and

then it goes into even more categories

and you can use generated knowledge and

then put the generated knowledge into

here as well you don't just have to get

the knowledge generated uh in this chat

thread you could have something else

generated from another chat thread throw

it in here you could have a Google sheet

with your ideas throw it in here and

then get your visual like this Claud 3.5

use case number 14 the second to last

use case is to use it for deep learning

and further research now you can do so

many research things like create

research papers with this with

interactive graphs categories sections

uh headings subheadings so many

different things you can go crazy on the

Deep learning aspect side of this

similar to the interactive dashboard I

showed you in the beginning of this

video but let me show you an example of

how you might be able to create a

research paper using clae 3.5 it just is

amazing how it does this maybe I'm

reading this paper 25 page PDF on

multimodal Chain of Thought reasoning in

large language models or in just large

models pretty hefty and let's say I want

to create a research paper based around

that I'm going to download this behind

my head and then I'm going to hit save

and then I'm going to head over to

Claude start a new chat and then I'm

going to drag that PDF in here and I'm

just going to say create a research

poster SLP paper on this PDF use unique

visuals in a more

complex way of viewing this data then I

said make this dark mode with yellow and

blue accents create a section of this

poster for actionable steps to take and

remember this is all based off that PDF

that's 1500 lines right this entire PDF

it's taking all this information it's

reading that PDF and creating a research

paper/ poster for me so this is going to

help me when it comes to actually

viewing this data in a better way maybe

if I find a template that I can use then

we can use tip number 15 which is about

to come up in order to do this for every

PDF we get take a look at how this did I

think this is just amazing how it breaks

the PDF into this fun interactive poster

as you can see we are missing an image

right here it just says placeholder but

let's say you want this to be a little

more interactive you can try something

like make this an interactive poster and

see what it does if it does anything at

all it say it's going to add expandable

sections and Dynamic data visualization

and clickable elements so I think making

this poster not only a poster with

complex viewing compared to the PDF but

also turning it into an interactive

poster it's going to be powerful so look

at this we now have drop downs so we can

hit these dropdowns and kind of go

through everything that we want and of

course the spacing isn't all that good

but I think this is pretty sweet how it

has interactive graphs and performance

comparison uh for the multimodal chain

up thought prompting then we have even

more things that we can kind of click in

here and just edit and manipulate

actionable steps that's awesome so this

is another way to use Claud Sonic 3.5 is

by using it for deep learning research

and just furthering your knowledge

retention by creating more fun things

like this from long drawn out boring

PDFs tip number 15 is going to be an

important one especially if you stuck it

out through this entire video because I

know that a lot of you are going to

create things that you want right you

you're going to create things that you

want to recreate and show people so I'm

going to show you the best way to create

a prompt template for your specific

project right CU after I get done

creating a project this is what I'm

doing now this is how I got this video

done in a fairly decent amount of time

is because I had a lot of these prompt

templates that you've been seeing me use

throughout the video so I'm going to

show you how to do that now let's take a

look at that example that I just did by

making this research paper and let's say

that I want to be able to do this well

once I complete it right and once I get

it exactly how I want it let's say that

I want to able to do this for every PDF

I have I could reply to Claude within

the same chat thread and say give me a

prompt I could use it would generate

this type of interactive poster every

single time in Claude how would I ask

you for this after uploading my PDF I

can send it off and then what it's going

to do is give me a prompt it says you

could use a prompt like this and then in

the quotes it gives me a beautiful

prompt that I could use in order to

generate something like this every time

and you can just use it whenever you

have a new PDF so creating these prompt

templates for your projects in Claud is

going to save you a lot of time it's

going to help your message limit stay

down so you don't end up buying two

accounts like me all right that's all

that I have for this Claude 3.5 update

again if you want to join a community of

AI enthusiasts who are doing stuff like

this all day finding AI news hopping on

live calls together you know you even

get this classroom section where you're

learning about large language models

there's tons of different modules in

here on every sector of artificial

intelligence we even have automation

systems that go very very in depth on

how to automate your business and we

really are just staying on top of AI in

here people are in here all day talking

we're making friends connections

networking so again if you want to join

this community I'll leave a link in the

description or the top pin comment but

with that being said that's the Claud

3.5 Sonet update I hope you enjoyed this

if you enjoy this type of content please

like it subscribe for more I'm going to

be staying on top of this mainly in the

community but also here on YouTube all

right with that being said I appreciate

you watching and I'll see you in the

next video