🐙 Lunch & Learn: Let's Build An AI Assistant With GPT-4o (w/ Joe & Winston of @Posit)

----------------------------------------------------
@Posit Channel - / @positpbc
🤖 Sign up here to get accompanying workbooks, summaries, and notifs for future lunch & learns: www.lonelyocto...
✉️ NEWSLETTER: tinahuang.subs...
It's about learning, coding, and generally how to get your sh*t together c:
🐙 Lonely Octopus: www.lonelyocto...
Check it out if you're interested in learning AI & data skill, then applying them to real company projects!
🔗Affiliates
========================
My SQL for data science interviews course (10 full interviews):
365datascience...
365datascience... (link for 57% discount for their complete data science training)
Check out StrataScratch for data science interview prep:
stratascratch....
🎥 My filming setup
========================
📷 camera: amzn.to/3LHbi7N
🎤 mic: amzn.to/3LqoFJb
🔭 tripod: amzn.to/3DkjGHe
💡 lights: amzn.to/3LmOhqk
📲Socials
========================
instagram: / hellotinah
linkedin: / tinaw-h
discord: / discord
🤯Study with Tina
========================
Study with Tina channel:
/ @studywithtina7726
How to make a studying scoreboard:
• The scoreboard I use t...
Scoreboard website:
scoreboardswithtina.com
livestreaming google calendar:
bit.ly/3wvPzHB
🎥Other videos you might be interested in
========================
How I consistently study with a full time job:
• How I consistently stu...
How I would learn to code (if I could start over):
• How I would learn to c...
🐈‍⬛🐈‍⬛About me
========================
Hi, my name is Tina and I'm an ex-Meta data scientist turned internet person!
📧Contact
========================
youtube: youtube comments are by far the best way to get a response from me!
linkedin: / tinaw-h
email for business inquiries only: [e-mail]
========================
Some links are affiliate links and I may receive a small portion of sales price at no cost to you. I really appreciate your support in helping improve this channel! :)

[@harrychou123]

Thank you for the video. I am an old-time software engineer and this is a perfect demo of the power of shiny and chatGPT integration.

[@tikkivolta2854]

00:00 Introduction 00:45 GPT-4 Demo and Project Goal 02:18 Guest Introductions (Joe & Winston from Posit) 04:55 OpenAI GPT-4 Demo Walkthrough 09:03 Project Overview and Architecture 12:51 Clarification on OpenAI App Capabilities 13:28 Open Source Alternatives to OpenAI Components 13:51 Jupyter Notebook Walkthrough 17:52 Discussion on Graphics Card Requirements 18:32 Transitioning to Shiny for Web App Development 19:07 Introduction to Shiny for Python 22:39 Discussion on Shiny Deployment Options 24:18 Comparison of Shiny and Streamlit 26:45 Starting with Shiny for Python 27:32 Demonstration of Basic Shiny App Features 28:12 Demonstration of Interactive 3D Visualization 29:18 Live Coding a Shiny App 30:00 Converting Jupyter Notebook to Python Script 31:58 Creating a Basic Shiny App Structure 34:07 Handling File Uploads and Displaying Information 36:07 Integrating OpenAI Query Function 38:10 Adding Progress Indication 39:37 Implementing Live Webcam Input 41:58 Styling the Web App Interface 42:56 Testing the App with Live Questions 44:43 Discussion on App Features and Improvements 45:26 Deploying the App and Opening for Audience Testing 47:13 Testing the App with OpenAI Demo Scenarios 51:38 Troubleshooting App Issues 55:24 Discussion on Ethical Considerations and AI Safety 57:21 Testing the App with Code Interpretation 57:57 Testing Language Capabilities (French, Japanese, Mandarin) 1:00:01 Demonstration of Multi-Model Comparison (GPT-4, GPT-4 Vision, LLaMA) 1:06:26 Testing Math Problem Solving Abilities 1:08:43 Discussion on Posit's Values and Approach to AI 1:11:27 Demonstration of Language Capabilities (Taiwanese, Klingon, Hungarian) 1:12:05 Testing Image Recognition and Recipe Generation 1:17:06 More Language Tests (Hindi, Esperanto) 1:20:04 Wrap-up and Call to Action 1:20:34 Shiny App Showcase and Q&A 1:21:14 Discussion on Shiny Use Cases and Templates 1:26:08 Exploring the Shiny App Playground 1:27:09 Shiny and Database Integration 1:30:11 Limitations and Future Possibilities 1:34:45 Closing Remarks and Next Week's Live Stream Announcement

[@IdPreferNot1]

Great demo. Excited to check out shiny.

[@hsinRamen]

I speak Mandarin, Taiwanese, Japanese, and English, and holy crap that translation was incredible. Interestingly, it kept the Japanese name, but read in Chinese rather than trying to translate it. I would've expected a "dumb" translator to try to translate the name. I was really hoping to hear it try Taiwanese lol. Hoping to give this a try myself.

[@taoxu1798]

Where can i find the code for the demo?

[@viperjay1]

Thanks for the demo!

[@shivamjaiswal-t7f]

Please add screen sharing feature

[@ZeeroCool1979]

Im Hungarian and it was impressive. Definitely not good translate at all but impressive progress. The video is really cool! Thanks!

[@emddom]

Surprisingly good Hungarian, including the pronunciation! Nekem is tetszett a magyar történet, bár van még mit fejlődni :)

[@csguru8211]

Yes the hindi Translation was absolutly correct but the voice tone seemed a little robotic, I mean not as good as english tone. Thanks for doing this I actually wanted to make something like this just to see the result but you guys showed the final result.

[@louayeltabech11]

The code that I executed below, shows an ATTRIBUTE ERROR: with datauri.as_tempfile(audio_uri) as audio_file: transcription = client.audio.transcriptions.create( model="whisper-1", file=Path(audio_file) ) user_prompt = transcription.text user_prompt It says that openai doesn't have the attribute 'audio'. I already have both the newest versions of python and opeanai libraries. Is anyone facing the same problem? Also, 'media_extractor' library doesn't exist, so I had to use a different library. I am using Jupyter by the way.

[@ABMyoutube]

Isn't ChatGpt 4o Multimodal? So why do you use txt2speed and speech2txt APIs? ChatGpt 4o is capable of handling Audio2Audio!

[@isarapp6262]

Because 4o API version is not released yet at the time of the stream?!

[@jcheng_]

At this time, the only API available for GPT-4o is chat completion (text/images in, text out). The multimodal capabilities OpenAI demonstrated aren't available to the public yet.

[@dactoanho6026]

Thank you.

[@Discovery_Nuggets]

Nice video!

[@roberttrajkovski9288]

Will you post this live session later

[@Learn4Satoshi]

Thanks ;)

[@TinaHuang1]

Tysm for supporting me 🤗

[@peterxyz3541]

Hong Kong❤️❤️❤️

[@MiroKrotky]

Omg i love TL so much happier

[@eurekadocumentary]

lol

[@MiroKrotky]

Yt sux

[@jurgenblick5491]

So how can I trust you

[@TheGuillotineKing]

Don't trust them Trust the AI

[@isarapp6262]

Well, the hungarian part was okey, but definitely not perfect! Does anybody have any idea, why speech synthetising doing so bad? I mean singing works perfect, if you look at AI systems like Suno AI, the vocals are just perfect (!!!), i cannot tell any difference to natural voice, not even in hungarian. Actually the AI vocals are way better, than an average human could sing. So why is speaking so much harder than singing for AI?