21 Powerful Use Cases for Claude 3.5 Sonnet (FREE)

The AI Advantage
5 Jul 202427:39

Summary

TLDRThe video explores the capabilities of CLA 3.5 Sonet, a powerful AI tool that excels in code generation and interactive application development. It demonstrates how Sonet can transform prompts into functional projects like data visualization dashboards, narrative-driven websites, and even simple games, all within its artifact feature. The host shares practical use cases, iterates on prompts for improved results, and guides viewers on hosting their creations online, showcasing Sonet's potential to empower non-technical users to build and visualize complex ideas with ease.

Takeaways

  • 🚀 Sonet 3.5 is a powerful AI model that excels in code generation and has capabilities that surpass GPT-4 in various categories.
  • 🎯 Its standout feature is the ability to generate code and provide an interactive preview of the code in action, which is a significant advancement in AI technology.
  • 🛠️ Practical applications of Sonet 3.5 include creating websites, games, and data visualizations, offering a wide range of use cases for users to explore.
  • 💡 The model is available for use today, and unlike some AI announcements, it's not just a promise of future capabilities—it's a current reality.
  • 🔍 The speaker has researched and compiled various use cases for Sonet 3.5, demonstrating its versatility and potential through examples and custom prompts.
  • 📊 Sonet 3.5 is adept at creating dashboards for data visualization, with the ability to enhance these dashboards through iterative prompts for improvement.
  • 💻 Users do not need to be technical experts to utilize Sonet 3.5, as it can generate code and provide guidance for non-technical individuals to follow or recreate.
  • 🔗 There are limitations to the model, such as the inability to execute certain packages within the artifacts feature, but it still provides valuable previews and guidance.
  • 🎮 Sonet 3.5 can be used to create simple games, offering a starting point for game development that can be expanded upon with further coding knowledge.
  • 🔗 The script includes a variety of prompts and examples that viewers can use as a starting point for their own projects with Sonet 3.5.
  • 🌐 The potential for Sonet 3.5 extends beyond the examples given, as it can be used to transform various types of content, like PDFs, into interactive websites or visualizations.

Q & A

  • What is the main topic of the video script?

    -The main topic of the video script is the introduction and demonstration of the capabilities of a new AI tool, Sonet 3.5, which is highlighted for its code generation and artifact feature that allows users to see code in action as it generates.

  • How does Sonet 3.5 differ from previous AI models like GPT-40 in terms of code generation?

    -SoNet 3.5 is said to be superior in code generation and not just in generating code, but also in its artifact feature that allows users to visualize the code as it is being created, which is considered an unprecedented development in the AI space.

  • What practical applications are mentioned for Sonet 3.5's capabilities?

    -The practical applications mentioned include creating websites, games, and visualizations, with a variety of use cases demonstrated in the video.

  • What is the significance of the 'artifact feature' in Sonet 3.5?

    -The artifact feature in Sonet 3.5 allows users to see the code as it is being generated, providing an interactive preview of the code in action, which aids in understanding and refining the output.

  • How does the video script suggest improving the dashboard created by Sonet 3.5?

    -The script suggests using a simple follow-up prompt like 'make it better' to guide Sonet 3.5 in interpreting what improvements could be made based on the context of the conversation and the previously generated code.

  • What is the role of the community in enhancing the use cases of Sonet 3.5 as mentioned in the script?

    -The community is involved in a weekly challenge where members use a provided prompt to flesh out their own use cases, sharing their experiences and learning from each other to improve their prompt engineering skills.

  • What limitations are there when using Sonet 3.5's artifact feature?

    -The limitations include the potential for generated code to include packages that are not executable within the artifact feature itself, requiring users to adapt the code for external environments like Visual Studio Code.

  • How can users who are not technically inclined benefit from the video and Sonet 3.5's capabilities?

    -Non-technical users can benefit from the video by understanding how to use Sonet 3.5 to create various applications without needing to know how to code, following the provided prompts and examples to generate their desired outcomes.

  • What is the process for users to get started with Sonet 3.5 after watching the video?

    -Users can start by trying out the use cases demonstrated in the video, using the provided prompts, and experimenting with Sonet 3.5's capabilities to create their own projects, with the option to join the community for further guidance and support.

  • What additional resources are suggested in the script for those interested in more advanced use cases or game creation with Sonet 3.5?

    -The script suggests looking at tutorials by Chris from All About AI on YouTube for more advanced game creation, and joining the Advantage Community for support and shared learning experiences.

Outlines

00:00

🚀 Introduction to Sonet 3.5's Revolutionary Features

The video script introduces Sonet 3.5, highlighting its availability for free with limited usage and its superiority over GPT-4 in various categories, especially in code generation. The script emphasizes Sonet's unique 'artifact' feature, which allows users to visualize code as it's generated. Practical applications include creating websites, games, and visualizations. The video promises to explore various use cases that leverage Sonet's capabilities, which are available for immediate use, and contrasts this with GPT-40's promises of future capabilities. The script also mentions a separate video detailing the model itself, focusing this video on practical applications with examples and custom prompts.

05:00

📊 Data Visualization and Interactive Dashboards

This section delves into Sonet's prowess in data visualization, showcasing its ability to create real-time dashboards for tracking disease spread across countries. The script describes an initial prompt that leads to the generation of an interactive preview, even without specific data input. It also discusses the 'make it better' follow-up prompt, which results in an enhanced dashboard with additional features like a bar chart and a global map view. The script points out the limitations of the artifacts feature, such as the inability to execute certain packages within it, and suggests solutions for running the generated code in environments like Visual Studio Code.

10:01

🛠️ Enhancing Dashboard Functionality with Advanced Prompts

The script continues with a discussion on refining the dashboard functionality using advanced prompts. It details a process where error messages from the artifacts feature are used to improve the generated code, making it compatible with the feature's limitations. The script also touches on the ability to host web apps outside the artifacts feature, providing a step-by-step guide for implementation in a production environment. It showcases an example of a pendulum visualization created by a user in just five minutes, demonstrating the potential for quick iteration and improvement.

15:01

📚 Transforming PDFs into Narrative-Driven Websites

This part of the script explores the transformation of PDF documents into narrative-driven websites using Sonet. It describes a prompt that converts a PDF into a website with storytelling elements and unexpected visuals. The script provides an example of turning a tutorial PDF into a story-driven website, complete with chapters and interactive elements. It also discusses the potential for adding features like quizzes to enhance user engagement and the ability to iterate on the project by pasting prompts directly into the system.

20:02

📈 Interactive Data Visualization from Images and PDFs

The script presents a use case where an image containing data is used to generate an interactive dashboard, demonstrating Sonet's ability to recognize and utilize image data. It describes the process of uploading a screenshot of data from a website and generating a population growth visualization dashboard. The script also highlights the speed and efficiency of Sonet's visual capabilities and the potential for customization and iteration based on user feedback.

25:03

🎮 Exploring Game Creation with Sonet 3.5

This section discusses the potential of Sonet for game creation, from simple games like Pac-Man and Tetris to more complex projects. The script mentions a video tutorial by Chris from All About AI, which demonstrates building a fully functional game with assets and AI opponents. It also showcases examples from the community, such as a 3D first-person shooter game and a checkers game with AI opponents, highlighting the ease with which non-technical users can create interactive games within the artifacts feature.

🎨 Creating Custom Game Assets and Interactive Games

The final part of the script focuses on creating custom game assets and interactive games. It describes the process of generating an 8-bit style image of a cat with a hat and then using that asset to create a snake game where the cat chases hats. The script emphasizes the quick iteration possible with Sonet, allowing for the creation of simple games in a short amount of time. It concludes by expressing excitement for the future capabilities of such tools and suggests further learning through tutorials on building custom chatbots with code.

Mindmap

Keywords

💡Code Generation

Code generation refers to the process of automatically creating source code from a set of input requirements. In the context of the video, it is a core feature of the AI model Sonet 3.5, which excels in creating functional code snippets for various applications. The script highlights this capability as a significant advancement in AI, allowing users to create websites, games, and visualizations with ease.

💡Artifact Feature

The artifact feature mentioned in the script is a unique capability of the AI model that allows users to see the code in action as it is being generated. This feature provides a dynamic preview of the code's output, making it an unprecedented development in the AI space. It is used to demonstrate the practical applications of the AI's code generation abilities, such as creating interactive data visualizations.

💡Data Visualization

Data visualization is the graphical representation of information and data. In the video, the AI model is prompted to act as a data visualization expert, tasked with creating a real-time dashboard to represent the spread of a disease over time across various countries. This showcases the AI's ability to not only generate code but also to create meaningful and interactive visual representations of complex data sets.

💡API

An API, or Application Programming Interface, is a set of protocols and tools that allow different software applications to communicate with each other. The script mentions the potential to connect the AI-generated dashboard to an API to provide real-time data updates. This illustrates the practical application of the AI's capabilities in creating web applications that can interact with live data feeds.

💡Packages

In the context of programming, a package is a collection of code, usually in the form of libraries or modules, that can be used in software development. The script explains that while the AI can generate code that includes packages, some of these may not be executable within the AI's artifact feature but can be used in other environments like Visual Studio Code.

💡Prompt Engineering

Prompt engineering is the process of carefully crafting the input prompts given to an AI model to elicit the desired response or output. The video script describes how the speaker used prompt engineering skills to refine AI-generated use cases, resulting in more detailed and specific outcomes, such as a more advanced data visualization prompt.

💡Web Apps

Web apps, or web applications, are software programs that run in a web browser over the internet. The script discusses the AI's ability to generate code for web apps, such as interactive dashboards and narrative-driven websites, demonstrating the AI's versatility in creating different types of web-based applications.

💡PDF Conversion

The script describes the AI's capability to convert PDF documents into different mediums, such as websites. This is showcased through a prompt that transforms a PDF into a narrative-driven website, blending storytelling techniques with visual elements to enhance the information presented in the PDF.

💡Interactive Quiz

An interactive quiz is a type of web application that allows users to participate in a quiz with immediate feedback. The video script includes an example where the AI is prompted to add a tab with a multiple-choice quiz to a website, showcasing the AI's ability to generate engaging and interactive content.

💡Game Development

Game development refers to the process of creating a video game. The script explores the AI's potential in game creation, mentioning the ability to generate ideas and assets for games, as well as code for simple games like Pac-Man, Tetris, and Snake. It also references more complex game development projects that require additional coding knowledge.

Highlights

Claw 3.5 Sonet is available with limited usage and outperforms GPT-40 in many categories, especially in code generation.

The artifact feature in Sonet 3.5 allows users to see code in action as it generates, marking an unprecedented development in AI.

Practical applications include creating websites, games, and visualizations with a wide range of use cases.

Sonnet 3.5's capabilities are available today, unlike some AI models that are only promises of future features.

The video provides examples and prompts crafted by the author for various use cases of Sonet 3.5.

The AI is adept at creating dashboard-type websites, as demonstrated by a real-time disease spread visualization.

A simple prompt 'make it better' leads to significant improvements in the dashboard, showcasing the AI's iterative capabilities.

The AI can create code that may not be executable within the artifact feature but can be run in other environments like Visual Studio Code.

The video includes a step-by-step guide on how to host the generated web apps, useful for non-technical viewers.

The AI can transform PDFs into narrative-driven websites with unexpected visual elements and storytelling techniques.

The video demonstrates adding interactive elements like quizzes to web pages with simple prompts.

AI can analyze financial reports and create interactive dashboards with stock recommendations.

The AI's image recognition allows it to generate interactive dashboards from screenshots of data.

The video shows how to create a web app that functions as a summarizer tool linked to the GPT API.

CLA 3.5 Sonet enables the creation of simple games like Pac-Man, Tetris, and Snake within the artifact feature.

The video includes a tutorial on building a custom chatbot with code without any coding knowledge.

The AI can generate 8-bit style images and integrate them into games, as demonstrated with a snake game.

Transcripts

play00:00

all right so as you probably heard claw

play00:01

3.5 Sonet is freely available with

play00:03

limited usage and it beats chat GPT 40

play00:07

across many

play00:08

categories the one that is absolutely

play00:11

undebatable though is code generation

play00:13

and not just code generation their

play00:15

artifact feature allows you to see that

play00:17

code in action as it generates it making

play00:19

this a unprecedented development in the

play00:22

AI space what does this mean practically

play00:24

you can create websites you can create

play00:26

games you can create visualizations and

play00:27

so much more there's a ton of use cases

play00:29

and actually really took my time with

play00:30

this video and researched this across

play00:32

the past week since Sonet came out and I

play00:34

compiled all of them for you in one

play00:36

video the beautiful thing here is though

play00:38

as opposed to the GPT 40 use case video

play00:40

that I did a lot of these use cases you

play00:42

can try you can replicate you can use

play00:45

for yourself today because Sonet 3.5 and

play00:47

its capabilities are not some sort of

play00:49

announcement or promise of things that

play00:51

are yet to come it's available today and

play00:53

we'll be looking at all the various ways

play00:54

you could use it today if you want more

play00:56

details on the model itself I created a

play00:58

separate video for that this video is

play00:59

going to be all about how you can put it

play01:01

to work which is going to be a

play01:02

combination of examples of how other

play01:04

people put it to work and prompts that I

play01:06

crafted myself some simpler some a

play01:08

little more intricate all of the prompts

play01:10

that you'll see here I'm providing to

play01:11

you for free in the description of the

play01:13

video if they're too long in the

play01:14

description I'll probably just have to

play01:16

link this notion page that you can

play01:17

freely access just like a website okay

play01:19

so let's get into it there's a lot to

play01:21

cover and I could have actually quite

play01:23

easily made a video titled the 50 most

play01:25

mind-blowing use cases but the thing is

play01:27

not all of them are created equal some

play01:29

are more impressive than others and some

play01:30

are more practical than others so I use

play01:32

my personal judgment and experience to

play01:33

pick the use cases that I figured you

play01:35

might find really interesting here

play01:37

because we're going to be visualizing

play01:38

data as an animation this is something

play01:40

the code interpreter aka the data

play01:42

analysis tool in GPT 40 cannot do so

play01:44

namely The Prompt I'm using here is act

play01:46

as a data visualization expert create a

play01:48

real-time dashboard to effectively

play01:49

represent the spread of a disease over

play01:51

time across various countries and this

play01:54

is one of the first capabilities I want

play01:55

to highlight it's really good at

play01:56

creating these dashboard type websites

play01:59

we'll be looking at more example shortly

play02:00

here but look from this simple little

play02:02

prompt it writes the answer and then

play02:03

here you have the artifact that

play02:05

essentially is all of this code but you

play02:07

don't even see the code in the first

play02:08

view it just writes that in the

play02:10

background and it gives you this

play02:11

interactive preview and you can look at

play02:13

the little app look how great this is I

play02:15

didn't even give it any data it just

play02:17

came up with some mock data for me and I

play02:19

can switch between the different

play02:20

countries and it has several features

play02:21

like it simulates real time updat so you

play02:23

could hook this up to API with code that

play02:25

provides it with real-time data and it

play02:27

will be updating as it goes quite simple

play02:30

right but if we go further in this

play02:31

conversation I'll present you with by

play02:33

far my favorite follow-up prompt here

play02:34

it's so simple yet so effective by

play02:36

simply saying make it better it will

play02:38

interpret what you mean by better it

play02:39

will look at the context so far at what

play02:41

it created and come up with a list of

play02:43

iterations that would make it better in

play02:45

claud's eyes that is but it turns out

play02:47

it's really good at figuring out what

play02:49

would make it better so by simply saying

play02:51

make it better it came up with this

play02:52

dashboard with way more information at

play02:54

the top you have the base data it added

play02:56

a bar chart option and there's a global

play02:58

map view and here we get to the first

play03:00

important caveat this will write code

play03:02

for you that might not be executable

play03:04

within the artifacts feature but if you

play03:06

bring this over to something like Visual

play03:08

Studio code and run it you can have

play03:09

additional packages that are not

play03:11

available in here just as a recap for

play03:13

non-technical people which by the way is

play03:14

what I assume for this video you don't

play03:16

need to know code to follow or recreate

play03:17

these use cases that I show you a

play03:19

package is a set of functions and

play03:21

instructions that do a specific thing

play03:23

with code usually they're more advanced

play03:25

functions that by using a package you

play03:28

don't have to bother with so in this

play03:29

case it's a package that represents maps

play03:32

of the world really well but this

play03:34

package is not available in here yet

play03:36

Cloud will still create this for you

play03:37

because it assumes it'll probably take

play03:39

the code and host it elsewhere this is

play03:41

just a preview right so this is

play03:43

something to be aware of because

play03:45

sometimes the more advanced prompts that

play03:46

I'll show you will use packages that are

play03:48

not accessible in here but if all you

play03:49

want to do is stay in here and use the

play03:51

artifacts feature you can absolutely do

play03:53

that we'll talk about the nuances of

play03:54

that once that comes up but as you can

play03:56

see by simply saying make it better it

play03:58

really improved this dashboard it added

play04:00

this global map view that is not

play04:01

accessible in here because it's a

play04:02

package that's not in here as I outlined

play04:04

and that's the very first use case but

play04:05

it doesn't stop there I actually went in

play04:07

and used some of the prompt engineering

play04:08

skills that I've been teaching on this

play04:10

channel and I crafted a prompt that

play04:11

helps improve some of these use cases

play04:13

and the result of it is this it's this

play04:15

larger prompt that goes into all the

play04:16

details of the data visualization that

play04:18

I'm looking for this is something

play04:20

similar to Sam The Prompt creator that I

play04:22

showed you on the channel about a week

play04:23

ago but this is for the use case of

play04:25

creating projects using artifacts and

play04:27

sonnet so what I'll do in this video is

play04:28

I'll give you all the use case prompts

play04:30

but this Improvement prompt we're

play04:31

refining that in the community we're

play04:32

actually hosting a weekly challenge

play04:34

right now where all the community

play04:35

members get to use that prompt to flesh

play04:37

out their own use cases so in the spirit

play04:39

of the advantage Channel I'll say what I

play04:41

always say is that these videos are

play04:42

always there to give you fish but then

play04:44

the community exists to teach you how to

play04:46

fish so you can catch them yourself and

play04:49

in this case the fish that I'm giving

play04:50

you is this flesh out prompt so if you

play04:51

copy paste it into here and simply run

play04:54

it you will see that it does a few

play04:56

things first of all you have manual

play04:57

control over the various features in

play05:00

here okay so instead of having a quick

play05:01

oneliner and it assuming a lot of things

play05:03

you tell it exactly what you want and if

play05:05

you're building a visualization like

play05:07

this that represents the spread of

play05:09

disease over time across various

play05:10

countries then it might be really good

play05:12

to have manual control over these things

play05:14

because here you can go into the prompt

play05:15

and you can actually edit you can

play05:16

actually see what decisions have been

play05:18

made rather than telling Claud bro

play05:19

figure it out yourself and here is the

play05:21

result and look I kept it like this on

play05:24

purpose because here we ran into our

play05:25

first problem there is an error message

play05:27

right this generated code includes

play05:28

unsupported life lies date FNS and

play05:31

calendar as calendar icon from Lucid

play05:33

react you don't need to know these

play05:34

packages you don't even need to know

play05:35

what this means you just have to know

play05:37

that this is an error message and what

play05:38

you can do in cloud and in gbt 40 every

play05:41

time is just copy the error message and

play05:43

paste all of this back into it you could

play05:45

also copy just the error message in the

play05:47

middle but this works like a charm just

play05:49

paste it in there and it apologizes to

play05:51

you and it will rebuild the app without

play05:53

those packages now you want to do this

play05:55

if you want to stay inside of the

play05:56

artifact feature if you want to bring

play05:57

this outside if you want to host this

play05:59

yourself on the internet you don't mind

play06:01

these packages will probably do

play06:02

something even better than you can do it

play06:04

here in these artifacts but if you want

play06:05

to stay in this web interface just take

play06:07

the error messages that it gives you

play06:09

throw them back in and there you go

play06:11

here's the improved app you can see

play06:12

there's various improvements different

play06:14

views you have an option to share the

play06:15

page you have an option to export the

play06:17

data and all we did is lead with the

play06:19

more advanced prompt and then if you

play06:20

want to stay inside of the artifact

play06:21

feature we just paste the error message

play06:23

and it readjusts it so that it works in

play06:25

here just be wary that will usually

play06:27

result in worse code objectively so if

play06:29

you want to host this elsewhere take the

play06:30

first version and also the obvious

play06:32

question eigor how do I host this

play06:33

elsewhere well this is not a technical

play06:35

tutorial on how to host websites or web

play06:37

apps on the internet but what I did do

play06:39

and in my prompt I included the block

play06:41

that tells it to give you a step-by-step

play06:43

guide on how to host it so if you look

play06:44

at the bottom of its output it doesn't

play06:46

just give you the artifact that the

play06:47

website that works but also here at the

play06:49

bottom it tells you to fully implement

play06:51

this dashboard in a production

play06:52

environment you would need to add a

play06:53

proper date range picker Implement real

play06:55

API calls add proper error handling

play06:58

implement the sharing and data expert

play06:59

functionalities it tells you all the

play07:01

steps that you could then communicate to

play07:03

a developer or work with Cloud to guide

play07:05

you through it and also what you'll see

play07:07

in the first result it gives you a

play07:08

step-by-step guide on how to host this

play07:10

yourself so this app goes offline as

play07:12

soon as you close the tab and you can

play07:13

share it with the world for others to

play07:15

use it but if you follow this

play07:16

step-by-step guide that comes from the

play07:17

initial Advance prompt I'm confident

play07:19

that if you take some time and use cla's

play07:21

assistance in it you will be able to get

play07:22

this up and running it's not easy but

play07:24

with the assistance of something like

play07:26

Cloud it's not that hard either and then

play07:27

beyond that I have a more advanced

play07:29

example Le of a visualization here on

play07:31

Twitter by Lu batala and it took him 5

play07:33

minutes and a couple of iterations

play07:34

that's what it takes sometimes right if

play07:36

it doesn't work you got to follow up you

play07:37

got to tell it what you don't like what

play07:39

it should change and he just visualizes

play07:41

a pendulum with a bunch of random

play07:42

numbers in between it then it generates

play07:44

a bunch of other pendulums and look you

play07:45

get this visualization I'm not exactly

play07:47

sure how you practically use this but

play07:49

you can do things like this and while

play07:51

the options might not be infinite

play07:52

there's so much you can do which just

play07:54

wasn't possible a month ago if you want

play07:56

the final prompt on the Tweet links are

play07:57

in the description as per usual all

play07:59

right next up we'll be using a new chat

play08:01

and this is a very simple prompt yet

play08:02

again create a narrative driven website

play08:04

from a PDF that weaves together

play08:05

unexpected visual elements and

play08:07

storytelling techniques to reinforce the

play08:09

information in the PDF this one comes

play08:11

from my personal prompt Library I have

play08:12

various interesting and esoteric prompts

play08:14

that I like to play around with here and

play08:16

there this was a variation of one of it

play08:18

and I used its ability to convert PDFs

play08:20

into websites and combined that with

play08:22

these really powerful llm keywords like

play08:24

storytelling techniques and unexpected

play08:26

visual elements and what do I need to do

play08:28

here of course I need to Pro provided a

play08:30

PDF otherwise it will be a little lost

play08:31

this is a PDF I shared on the channel

play08:33

before it's one of Now 60 or 70 AI

play08:35

Advantage Community guides and it's this

play08:37

m Journey technique where you can

play08:38

generate screens with green screens in

play08:40

them which makes it really easy to

play08:41

create custom marketing materials for

play08:43

your very own brand because it's really

play08:44

simple to replace the green screens in

play08:46

either with a chroma key effect in a

play08:48

video editor or by selecting that one

play08:50

color inside of Photoshop canva or

play08:52

similar anyway the point is that this is

play08:54

a stepbystep guide and we're going to

play08:56

try to visually represent this and just

play08:58

run this prompt and see what we get get

play08:59

now this one is so interesting because

play09:02

the concept here is not the idea of

play09:04

turning a journey tutorial PDF into a

play09:07

website it's the idea of transforming

play09:09

any PDF into a different medium llms are

play09:12

essentially Transformers that generate

play09:14

various outputs right to generate text

play09:16

image even video now with Gen free and

play09:18

dream machine or with cloudon it they

play09:20

create web apps websites and games or

play09:23

interactive visualizations now look this

play09:25

is a very simple prompt and I just gave

play09:26

it the PDF without any additional

play09:28

context and it creat this website for me

play09:30

and as you can see it is narrative

play09:31

driven chapter one The Discovery and it

play09:33

turns my little tutorial into a story

play09:35

driven website isn't that interesting

play09:37

you could enter the mid Journey prompt

play09:38

here generate the image obviously you

play09:40

would need to hook this up to the API to

play09:41

actually work with mid journey and mid

play09:43

Journey doesn't have a API so you would

play09:44

need to do stable diffusion or something

play09:46

else but you get the point here there's

play09:48

placeholder images in the background

play09:49

look at these scrolling elements this is

play09:51

not your typical website powerful little

play09:53

prompt right there that you can play

play09:54

around with yourself here's the formula

play09:56

of how to do it and then here at the

play09:58

bottom we should have the different use

play09:59

cases we always make a point out of

play10:00

including use cases at the bottom of the

play10:02

community guides and there you have a

play10:04

few of them there's an epilog with a

play10:05

little motivational message very

play10:07

interesting isn't it and if you take

play10:08

this a step further and you flesh out

play10:10

this prompt into this yet again The

play10:12

Prompt is available right here it

play10:14

fleshes out all the details of what we

play10:15

want on the website and in this case

play10:17

it's interesting because this is

play10:18

actually less creative and less

play10:20

unexpected the more guidelines you give

play10:22

it in the prompts the less room it will

play10:24

have for interpretation so in this case

play10:26

maybe going with the more open-ended

play10:27

prompt of narrative driven web it that

play10:29

weaves together unexpected visuals and

play10:31

storytelling and then figure out all the

play10:33

details of that cloud might be a better

play10:35

option than going with this Advanced one

play10:36

nevertheless I just wanted to give you

play10:38

the options this looks certainly more

play10:39

formal but does the trick this menu on

play10:41

top jumps to the various sections of the

play10:43

website and it's just a very powerful

play10:45

way of turning any PDF into an

play10:47

interactive website and there you go you

play10:49

created the website from a PDF in

play10:51

seconds and here's another interesting

play10:52

capability you have this menu at the top

play10:54

right here and you just got to realize

play10:56

that you can always add to these web

play10:57

pages add to these projects it's just

play10:59

because I show you how to do things in

play11:00

one prompt doesn't mean that you can't

play11:02

have a conversation with it I'll just go

play11:03

in here and paste this prompt here add a

play11:06

tab with an interactive quiz I featured

play11:08

this many times on the channel creating

play11:10

little quizzes questionnaires flashcards

play11:13

it's all LMS are really good at that and

play11:15

we can use that capability and our

play11:17

awareness of that capability in a little

play11:18

prompt like this by adding a tab with a

play11:20

quiz that consists of five multiple

play11:22

choice questions I'll just run this and

play11:24

it will update this project here for me

play11:26

all I have to do is wait for a few

play11:27

seconds the funny thing is with gbd4 I

play11:30

usually had to stop and wait for a

play11:32

minute until it's finished and then we

play11:33

cut the recording but here I'm just

play11:35

talking over this happening a few

play11:39

moments later okay fair enough we had to

play11:41

fast forward a little bit just because

play11:43

generating the code for all the multiple

play11:44

choice questions took a while but look

play11:46

at that we have a little tab with a

play11:48

multiple choice quiz you could submit

play11:49

that quiz you can ask Claude how to

play11:51

implement this practically and there's

play11:53

some console messages that I could now

play11:55

take and paste back and this is the way

play11:57

you want to work with it you want to add

play11:58

your ideas and something doesn't work

play11:59

and you get error messages you just

play12:01

paste them back in and then it will

play12:02

improve on it by itself and here I want

play12:04

to highlight another use case here from

play12:06

Sabo shubman on X and what he actually

play12:08

did he gave it the Tesla earnings report

play12:10

for the first quarter of 2024 and

play12:12

because that is very well formatted he

play12:14

gave it a simple prompt like this come

play12:16

up with a detailed analysis and summary

play12:17

with stock recommendations by carefully

play12:19

reading the attached earnings report put

play12:21

together a nice interactive and highly

play12:22

detailed but visual appealing dashboard

play12:25

and here you have the dashboard keyword

play12:26

again very useful if you just want to

play12:27

create overviews of something and here

play12:29

it is he transformed a PDF into a

play12:31

website this is just another use case

play12:33

after the step-by-step guide that I

play12:35

showed you you could do this with any

play12:36

PDF just try it out I mean it's free to

play12:39

use isn't that crazy sure there's a

play12:40

limitation of only a few messages but

play12:42

you could just try all of this yourself

play12:44

today all right next up I'll be using

play12:46

another one of my prompts create an

play12:47

interactive dashboard with editable

play12:49

sliders and text fields to visualize

play12:51

population growth in the US over time

play12:53

and mind you we're not giving it any

play12:54

data here so far right we're just

play12:56

telling it to create in another

play12:58

dashboard but what I did here is I took

play12:59

a screenshot from this website worldom

play13:01

meters. info you can take your data from

play13:03

wherever you want but because it can

play13:05

recognize images I want to show you this

play13:06

capability of you feeding it images and

play13:08

using the various dashboard website game

play13:11

generation capabilities that we have

play13:13

here to create something that might be

play13:15

useful to you so all I did is

play13:16

screenshotted a part of this data that

play13:18

is relevant to what we're doing here

play13:20

here's the screenshot and I'll just drag

play13:21

and drop it onto Cloud now now I

play13:23

uploaded the content and as soon as it's

play13:25

uploaded I can run this and it will

play13:27

generate this dashboard based on the

play13:28

data data that I gave it right you could

play13:30

also run it without the data then it

play13:32

will just invent some random stuff for

play13:33

tell you to connect it to API later on

play13:35

so this time we're not doing PDF plus

play13:37

prompt but we're doing image plus prompt

play13:39

and we're creating an interactive

play13:40

dashboard from that and again this

play13:42

should happen so fast that maybe we

play13:44

don't even need to fast forward or cut

play13:47

too much it's already working on it it's

play13:49

already generating code it's going to

play13:50

open up this little IDE over here and as

play13:53

you can see here I love this you can see

play13:54

how it populates every single year you

play13:56

can always double check that but the

play13:57

vision capabilities of cloud as I talked

play13:59

in my summary video they're also Best in

play14:01

Class when it comes to the benchmarks

play14:02

and practically that's the case too so

play14:04

now it packages all of that it finishes

play14:06

up the prompt and we should be presented

play14:08

with a visualization any second here

play14:11

let's have a look there's a little bit

play14:12

of loading oh and I'm out of messages

play14:14

okay I'll have to wait another 20

play14:16

minutes that is okay but look at this

play14:18

wonderful dashboard that it created in

play14:19

the very first try here I can actually

play14:21

adjust these sliders like what is the

play14:22

range they want to go from 9055 to 2024

play14:25

now here I would maybe like a second

play14:27

button here where I can readjust it I

play14:29

could just follow up prompt this if I

play14:30

had more messages this very second but I

play14:32

think you see the point here by the way

play14:34

I'm on a I'm on a PID plan and I've been

play14:36

testing like crazy over the past days

play14:38

and especially today as I finalized all

play14:40

the research for this video but no

play14:41

worries in 20 minutes we'll be back and

play14:43

as you can see this works as expected

play14:45

and we have an interactive graph that we

play14:47

easily generated from a screenshot and a

play14:49

simple prompt like this pretty amazing

play14:51

and I want to highlight one more thing

play14:52

in the original chat it cut off some

play14:53

numbers here for me I didn't like that

play14:55

so I just told it the numbers are cut

play14:56

off if I go into billions make it more

play14:58

visual appealing you unique and better

play14:59

in this example I didn't even give it

play15:01

the data so it just invented some but as

play15:02

you can see by following up with your

play15:04

feedback with your human feedback as it

play15:06

would be giving it to your assistant or

play15:08

another human that works for you or

play15:10

another human that works with you it

play15:11

makes these adjustments on the go but

play15:13

now there's a graph missing so I say

play15:15

include a graph and then it creates a

play15:17

graph but what is this graph this is

play15:19

more graph than website the heck this is

play15:21

some mutant graph I don't like it I

play15:23

simply told it this graph is crazy fix

play15:25

it and then it interprets Crazy by

play15:27

itself okay this prompt might be a bit

play15:29

ambiguous going against what I teach on

play15:30

this channel but hey you got to have fun

play15:32

sometimes too and then eventually after

play15:33

telling it that it should fit on the

play15:35

screen being more specific than just

play15:37

telling it it's crazy that didn't work

play15:39

it does this for me which I think is a

play15:41

fantastic end result considering I give

play15:43

it what like three sentences in total

play15:45

incredible okay here's the same use case

play15:47

but with the flash. prompt that we show

play15:49

you how to do in this week's Community

play15:50

challenge so by copy pasting this prompt

play15:52

into here you get this artifact

play15:54

generated for you and as you can see

play15:56

from The Prompt straight up I didn't

play15:58

even edit the prompt right I just ran it

play15:59

from my little workflow there I get

play16:01

something that is immediately usable

play16:03

there's a lot of details like export

play16:04

data and it's just the more advanced

play16:06

version with more customizable building

play16:08

blocks in here but this is pretty great

play16:10

I get multiple ways to visualize this I

play16:12

export button and so much more and it

play16:15

all works in one shot so this is what I

play16:17

say in some of my other videos if you

play16:18

dare to give it feedback and if you know

play16:20

what you expect it's fine to have short

play16:22

and simple prompts but if you want

play16:24

reliable results or you're going to

play16:25

automate it away or you're going to

play16:27

delegate the prompt to somebody else you

play16:28

want want one prompt that contains it

play16:30

all because that eliminates room for

play16:32

error and provides you with consistency

play16:34

the results from this prompt will be way

play16:36

more consistent then from a prompt like

play16:38

this just because we're nailing down all

play16:40

the various variables and telling it

play16:41

exactly what we want rather than letting

play16:43

it figure out everything itself all

play16:45

right for the next use case we'll be

play16:46

creating a web app and concretely we'll

play16:48

be creating an AI powered web app now

play16:50

this is something that won't fully work

play16:52

in here just because you can't link to

play16:53

API endpoints from within the artifacts

play16:55

feature this you would need to copy over

play16:57

to some IDE or host it yourself on

play16:59

something like versal but by simply

play17:01

telling it create a summarizer tool that

play17:02

links to the claw API include a field to

play17:05

paste text and a field to add the prompt

play17:08

and you get this here you paste the text

play17:09

you want to summarize here you have the

play17:11

option for a custom prompt and all you

play17:13

do is hit summarize and it will give you

play17:15

the summary good but then I'll follow up

play17:17

with the good old make it better what do

play17:19

we get well we get a better version it

play17:21

figures out what better means in this

play17:23

context itself concretely here it adds

play17:25

an advanced mode where you can actually

play17:27

adjust the max tokens and the

play17:29

temperature and beyond that it gives you

play17:30

the free different clot options to

play17:32

choose from so if you want to be using

play17:33

sonnet you can pick that it's going to

play17:35

be faster it's going to be cheaper again

play17:37

this won't work by itself as it outlines

play17:39

here in the prompt you will need to go

play17:41

into the code and you'll need to replace

play17:43

your API key here with your actual API

play17:45

key if we head on over to the code this

play17:47

shouldn't be hard to identify you can

play17:49

always just command of API uncore and

play17:52

here it is this is what you would need

play17:54

to replace with your very own API key I

play17:56

showed you that about 20 times before on

play17:57

this channel just look up up one of the

play17:59

chatbot building tutorials to retrieve

play18:01

your opening ey API key or just ask

play18:03

Claude where you can find it now beyond

play18:05

that I continued in the conversation by

play18:06

telling it to make it more aesthetic and

play18:08

look at these simple little prompts make

play18:09

it more aesthetic make it better or in

play18:11

some cases make it more unique also

play18:13

works really well and it does that it

play18:14

makes it more aesthetic look at that it

play18:16

adds this universal symbol for AI that

play18:18

the internet seems to have agreed on at

play18:20

this point it's kind of funny and then

play18:21

here we have the advanced features now

play18:23

it's not perfect there's a bit of

play18:24

padding missing on this but again we

play18:26

could follow up on this and I just want

play18:28

to highlight this is not supposed to

play18:29

build 100% perfect apps it's just the

play18:33

best llm we ever had for building

play18:35

applications like this and it allows

play18:37

non-technical users to actually build

play18:39

things which before wasn't really the

play18:41

case you need to download IDE install

play18:43

python make sure you have all the

play18:44

packages that are needed I covered this

play18:46

on the channel before matter of fact

play18:47

there's one video that teaches you how

play18:49

to build a basic chatbot it's the first

play18:50

one in the chatbot building playlist on

play18:52

the adantage channel and that teaches

play18:54

you how to set up a local environment

play18:55

for you to copy this code into where it

play18:58

will work it's about building a chat GPT

play19:00

chatbot with python but most of the

play19:02

setup steps and retrieving the API Keys

play19:04

Etc are all the same so check out that

play19:06

video if you want a detail step-by-step

play19:07

tutorial on how to use this code

play19:09

yourself anyway it's time to move on to

play19:11

the next use case which in this case is

play19:13

something that me and the team found on

play19:15

Reddit and this one is incredible let me

play19:17

tell you now look before you even look

play19:18

at this I just have to clarify this is

play19:20

not something you will get from one

play19:21

prompt this must have took a lot of

play19:23

iteration and as you can see this is not

play19:25

hosted inside of the artifacts features

play19:27

so you have to bring the code out host

play19:29

it somewhere and then you can get

play19:31

something like this as you can see this

play19:32

is hosted on the internet these more

play19:34

advanced prompts that I include in

play19:36

tutorial always give you a step-by-step

play19:37

guide on how to host plus there's a

play19:39

plethora of tutorials across the

play19:40

Internet it's not too hard almost

play19:41

everybody can figure that out in an

play19:43

afternoon but as you can see this is a

play19:44

web hosted app and it's this travel

play19:46

planner and I think it's absolutely

play19:47

incredible it includes images which were

play19:49

obviously added it includes a link to

play19:51

Google Maps and all of this was set up

play19:53

by Claude is there additional steps

play19:55

after Claude does the generation yes

play19:57

absolutely but can CLA son it guide you

play19:59

through every step of the way also yes

play20:02

you can provide screenshots of your

play20:03

progress as you work for something and

play20:05

tell it what kind of problem you have

play20:07

and it will act as a tutor for you in

play20:08

implementing it too but look at that you

play20:10

can search for various locations across

play20:12

B with custom images it gives you the

play20:14

different points on the map Incredible

play20:15

use case right there and here's another

play20:17

one from Sabo Schuman where he actually

play20:19

took a screenshot of the chat GPT

play20:20

interface and he let CLA 3.5 Sonet

play20:23

recreate chat GPT kind of rude isn't it

play20:26

everybody stay calm

play20:29

but it did exactly that it wrote the

play20:31

entire code for a cat gbt 40 clone with

play20:34

the prompt presets with all of that you

play20:36

could hook that up to the GPT 40 API and

play20:38

have yourself built chat GPT 40 clone

play20:41

just like that again the links to all

play20:43

these posts are below you can check out

play20:45

more details there if you're interested

play20:47

in that but this brings me to my last

play20:49

category which is a bit of a tricky one

play20:51

because it is game creation and I do

play20:53

have some very basic prompts here but

play20:55

look I just have to sayate I'm not a

play20:56

game developer I might be an avid gamer

play20:58

always have been but I've never created

play21:00

games but with CLA you can get your feet

play21:02

wet and you can build your own

play21:04

interactive little games yourself and

play21:06

they actually work in artifacts featur

play21:08

so you don't have to bring them out you

play21:09

don't have to host them anywhere they

play21:11

just work in there now is there

play21:12

limitations of course you won't be able

play21:14

to build GTA 6 in here let me tell you

play21:16

that but you can create simple little

play21:18

things like Pac-Man Beaman Tetris or the

play21:21

good old snake so we'll do that in a

play21:22

second here and I'll even show you how

play21:24

to create custom game assets yourself

play21:25

that you can then use in your very own

play21:27

game but before that I want to dive into

play21:29

the deep end of the pool and Shout out a

play21:30

friend of the channel Chris from All

play21:32

About Ai and this video that he posted

play21:34

on YouTube I think this is an absolutely

play21:36

incredible one if you care more about

play21:38

game creation and you want to see what

play21:39

you can do if you're ready to bring the

play21:41

code out and to do a little bit of code

play21:43

edting which does require some coding

play21:45

knowledge then check out this video you

play21:46

can also follow his project step by step

play21:48

and rebuild this game that he did but

play21:50

here in under 30 minutes he actually

play21:52

went through a step-by-step tutorial

play21:53

through all the asset Generations he

play21:55

generated these little flappy bird

play21:56

assets in eight bit that I'll show you

play21:58

something simil in a second here himself

play22:00

or maybe not himself maybe rather in

play22:01

tandem with Sonet 3.5 but that's the you

play22:04

doing it by yourself I guess and then he

play22:05

just goes for all the step of adding

play22:07

various Sprites adding AI opponents yes

play22:10

that's something that Sonet can do

play22:11

really well adding music and sound and

play22:13

background and finishing up the game and

play22:15

then he is a fully functional game that

play22:17

he can actually play and he uses claud's

play22:19

new project feature to build it as he

play22:21

generates more assets and more code he

play22:23

adds it to a project and then he keeps

play22:25

talking to it and Sonet is aware of all

play22:27

the different code and all the the

play22:28

different things that have been added to

play22:29

the game so if you're interested in

play22:30

building something more complex and you

play22:32

try the various use cases in this video

play22:33

then Chris's video here might be a

play22:35

fantastic next step and let me just note

play22:37

I forgot my table mount for this mic so

play22:38

I keep handing it off from one hand to

play22:40

the other cuz it's really heavy anyway

play22:42

let's get back to the video here so here

play22:43

you have a full project this is kind of

play22:45

as far as you can take it with a tool

play22:46

like this when it comes to game

play22:47

development today I mean I'm sure you

play22:49

could do a little more but this really

play22:51

shows you what is possible very well but

play22:53

again that does require some coding

play22:55

knowledge if you want to kind of just

play22:56

play around and have a good time then

play22:58

let me show you some other use cases

play23:00

before we build something ourselves

play23:01

because you could not just ideate but

play23:03

you could also build with CLA 3.5 Sonet

play23:05

when it comes to games in this Reddit

play23:06

example Sir W tafu here made a simple 3D

play23:10

firstperson shooter touchcreen game

play23:12

right in the chat interface okay so all

play23:14

of this is created within the artifact

play23:16

feature in the game you shoot happy

play23:18

emojis at sad monsters to make them

play23:20

happy kind of a neat concept right by

play23:22

the way the ridiculous idea for this

play23:23

game is

play23:26

claws so you could first generate the

play23:28

idea for the game and then create it all

play23:30

you need to be aware of is the various

play23:32

things that you need which Chris's video

play23:33

here perfectly outlines you need

play23:35

opponents you need Sprites unique music

play23:37

sounds maybe a menu to navigate it Etc

play23:40

and as we're going through and exploring

play23:41

some of these ideas let me show you

play23:42

another one here because this one is all

play23:44

about adding AI players and this is

play23:45

something that got a lot of people

play23:46

excited because you can create single

play23:48

player games and add AI players to it

play23:50

really simply with CLA 3.5 Sonet this

play23:53

was something that was a little tricky

play23:54

with GPT 40 and let me tell you from

play23:55

personal experience it just did not work

play23:57

reliably if you want want all the

play23:58

prompts they're also here in this Reddit

play24:00

fret look he posted every single prompt

play24:02

and as you can see this is a 15 prompt

play24:04

sequence so don't expect to build some

play24:05

of these more impressive use cases and

play24:08

one or two prompts going to have to

play24:09

iterate a little bit but I hope that

play24:11

this video is enough inspiration and

play24:12

pointed towards enough resources that

play24:14

you'll going to be able to figure it out

play24:15

yourself or you join the advantage

play24:17

community and you can do it together

play24:18

with others Shameless plug okay here's

play24:20

another use case that you can build as I

play24:22

mentioned these basic games are easily

play24:23

buildable so here it combines a basic

play24:25

game and an AI enemy to build a checker

play24:28

game where you're playing against the AI

play24:29

and as you can see this is running in

play24:31

the browser on your phone okay so you

play24:34

don't need to host it externally you can

play24:35

build these things while you sit at

play24:37

dinner with your family and show them

play24:38

hey Mom I built a checkers game where

play24:40

you can play versus AI myself while you

play24:42

ate your

play24:45

appetizer actually thinking about it

play24:47

don't do that most people are going to

play24:48

consider that weird not me though I

play24:50

think that's kind of a badass move or

play24:52

here we have one more example from Ethan

play24:53

mullik before I show you the last two

play24:55

prompts I have here and he actually

play24:56

turned the academic research paper that

play24:58

he wrote himself you could do this any

play24:59

academic research paper into an

play25:01

interactive game and this is a 3D

play25:03

shooter that references Concepts from

play25:05

the paper to be used in the game again I

play25:07

can't say everything is possible in here

play25:09

but a lot is possible and my hope is

play25:11

that for examples like this that becomes

play25:13

clear so there we go finally my cloud

play25:15

usage actually reset and we're going to

play25:16

be able to do the thing that I've been

play25:19

waiting for all day and that is generate

play25:20

an image of a cat with a hat in the 8bit

play25:23

style okay if you're not familiar with

play25:25

8bit in a second you'll see exactly what

play25:27

that is and you could see it writing the

play25:29

code to generate a little cat with a red

play25:31

hat is this a cat in a hat isn't that

play25:35

beautiful and then what I want to do

play25:36

here is I want to follow up with a

play25:37

prompt and I just want to show you this

play25:39

combo this is very basic right this is

play25:40

something you could figure out yourself

play25:41

but it's this ability to combine that

play25:43

you really need to become aware of

play25:45

because as I showed you some of the best

play25:46

things in here are built with 15 prompts

play25:48

or more so if I now say create a snake

play25:50

game with the cat with a hat as the

play25:52

player character you can create a snake

play25:54

game but you can use the little 8bit

play25:56

style Cat in SV format that we created

play25:59

before in the same conversation as your

play26:02

player character so let's see how it

play26:04

handles that it should be able to pick

play26:05

up the SVG write the code for snake that

play26:08

should be pretty straightforward and in

play26:10

just two prompts we should have a

play26:11

playable game if that's not the case

play26:13

I'll follow up further but let's see use

play26:15

the arrow keys to move the cat this is

play26:17

really great look and my little cat is

play26:18

getting longer and longer but like I'm

play26:20

missing one more thing right is it just

play26:22

me or do we want the cat to actually

play26:25

hunt hats so let me see if I can do this

play26:27

in one prompt create an image of a hat

play26:29

and 8bit style and use it as the food

play26:32

and then why am I saying food because I

play26:34

looked at what it wrote and I saw that

play26:35

it references to the thing that we need

play26:37

to eat up as food this is a very old

play26:40

trick but it's one of the most reliable

play26:41

ways to reduce ambiguity just reference

play26:43

the words that it uses itself can't go

play26:46

wrong with that and there you go the

play26:47

very first time on this channel we

play26:49

created a little snake game where a cat

play26:50

with a hat is hunting more hats and

play26:52

extending its Empire extending its body

play26:55

by finding more hats it might look basic

play26:57

from the outside but consider

play26:58

considering the fact that we just made

play26:59

this while I was talking in under a

play27:01

minute I think that's seriously

play27:03

impressive and this is the worst these

play27:05

tools are ever going to be I can't wait

play27:07

for the large model Claude 3.5 Opus or

play27:10

open eyes response to this move I hope

play27:13

this video brought you some inspiration

play27:14

and some knowledge and if you want to

play27:16

get more into this if this sparked your

play27:18

interest I would again recommend this

play27:19

little playlist that is still relevant

play27:21

on how to build a custom chatbot with

play27:23

code without any coding knowledge it's

play27:25

the most popular long form tutorial on

play27:27

this channel for a reason and using

play27:28

something like vs code and installing

play27:30

various coding languages yourself on

play27:32

your own machine is The Logical next

play27:33

step to Bringing games like this to life

play27:35

yourself all right that's all I got for

play27:37

today see you soon

Rate This

5.0 / 5 (0 votes)

الوسوم ذات الصلة
AI DevelopmentCode GenerationWeb CreationGame DesignData VisualizationInteractive DashboardAPI IntegrationUser ExperienceTechnical TutorialInnovation Trends
هل تحتاج إلى تلخيص باللغة الإنجليزية؟