6 | OpenAI DevDay | Grok AI |

This week was quite the spectacle at OpenAI with their so-called 'DevDay'. But, hold onto your hats, because it turns out this affair was cooked up not only for the tech wizards but also for the business people as well as for creative individuals.

In the race of artificial intelligence, OpenAI has seemingly surged ahead. Mr Musk, playing the part of the maverick, launched his own AI chatbot to the expected fanfare and furor.

As for what transpired at the AI summit in London, the details are as elusive as Schrödinger’s cat. but the after-effects were clear: a full-throttle acceleration in AI enthusiasm, reminiscent of a spirited drive along the Autobahn.

Our Menu :)

• Weekly Digest:

  • OpenAI DevDay, All you need to know 🤖👏.

  • What people have created 11 hours after the event, it’s impressive!

  • The Future of AI Creation Economy.

  • 🚀 Grok AI, Musk's New AI.

• Actionable Tips:

  • Guide to create my first GPT in 5 Minutes.🖐️

• Training & Crash Courses:

  • Microsoft's Generative AI Course 💡

WEEKLY DIGEST
OpenAI DevDay, All you need to know.

OpenAI held its inaugural dev event, unveiling new APIs and capabilities. ChatGPT hit 100M weekly users just months after launch. So, if you've ever considered subscribing to ChatGPT Plus, now might be a good time.

Announcements for ChatGPT Plus

🌟 ChatGPT Real Multimodality: ChatGPT new update that combines all of its tools in one place, allowing users to craft images, analyze data, and browse the web without switching between tools.  

🚀 Customizable GPT Builder: OpenAI is now rolling out a new product called “GPTs”, which they describe as “custom versions of ChatGPT that you can create for a specific purpose.

• Ease of Use: Set up your GPT with simple, natural language commands—no coding required.

• Advanced Customization: Enhance your GPT with detailed instructions, incorporating everything from files and descriptions to actions and even DALL-E 3 image profiles.

• Monetize Your Creations: With the launch of the GPT Store, anyone can sell their custom GPTs and start earning.

• Exclusive Access: ChatGPT Plus and Enterprise users get early access to GPT Builder.

📚 Knowledge base to April 2023, enhancing the relevance and accuracy of the information you receive, moving beyond the previous September 2021 cutoff.

Announcements for Developers

🌐👨‍💻👩‍💻✅ = Available to all developers

💳👨‍💻👩‍💻✅ = Available for all paying developers

GPT-4 Turbo with a 128k context length: The introduction of GPT-4 Turbo, a faster model with a larger token limit of 128k (equivalent to approximately 300 pages of 📚 ), addresses the issue of hallucinations that occurred with prolonged conversations. 💳👨‍💻👩‍💻✅ via API

Substantial price reductions: The GPT-4 Turbo API represents a significant stride in economic efficiency, being 2.75 times more cost-effective than the previous GPT-4 model. This price reduction extends to the GPT-3.5 Turbo 16k model, those employing the 4k context window of GPT-3.5 will see a substantial 33% decrease in operational costs.

New text-to-speech (TTS) model and API: Developers can now generate human-quality speech from text via a text-to-speech API. The current TTS model offers six preset voices to choose from and two model variants, tts-1 and tts-1-hd. 🌐👨‍💻👩‍💻✅ via API

New speech-to-text Whisper v3: Whisper is OpenAI’s cutting-edge Automatic Speech Recognition model (ASR) model. Whisper’s open-source release in Sep 2022, Whisper features improved performance across languages. 🌐👨‍💻👩‍💻✅  via the Whisper package on GitHub.

Assistants API and Retrieval: Using the Assistants API, developers can create agent-like AI within their applications, equipped with specialized functions like Code Interpreter, Retrieval, and function calling for efficient task execution. No more fancy Retrieval Augmented Generation (RAG) pipelines. Users just need to upload files to extend GPT’s knowledge. 🌐👨‍💻👩‍💻✅ under beta program

Other Relevant Announcements

OpenAI’s new ‘Copyright Shield’ will pay the costs incurred, if you face legal claims around copyright infringement while building with tools.

WEEKLY DIGEST
What people have created 11 hours after the launch is impressive!  

Barely half a day has passed since OpenAI's Dev Day, and we're already witnessing a surge of ingenuity. Here's a glimpse at 5 of the most astonishing creations that have emerged:

=> OUR ARTICLE <= 
The Future of AI Creation Economy.

In a bold move that could reshape the future of AI, OpenAI is set to launch the GPT Store, a marketplace for AI tools and applications. With this ambitious endeavor, OpenAI aims to become the “Apple AppStore of AI,” providing a platform for creators to develop, share, and monetize their AI creations.

At the heart of the GPT Store lies a premium subscription model that grants access to GPT creation tools. This model aligns OpenAI’s interests with those of creators, as both parties benefit from the growth of the platform. OpenAI further incentivizes creators by levying a 30% commission on their GPT Store revenues, similar to Apple’s App Store¹.

The GPT Store promises to democratize AI creation, empowering individuals with no coding experience to develop personalized AI agents, or GPTs. These GPTs can be tailored to various tasks and applications, ranging from chatbots and virtual assistants to sophisticated AI agents capable of performing complex tasks and providing valuable insights. For example, GPTs can help you learn the rules to any board game, help teach your kids math, or design stickers².

The success of the GPT Store hinges on its ability to attract and retain a thriving community of creators. To achieve this, OpenAI will leverage its extensive user base, which includes the majority of Fortune 500 companies³. The company also plans to actively promote the GPT Store and its creator community through various marketing and outreach initiatives, such as hosting online events, featuring the most popular and useful GPTs, and offering rewards and recognition to the top creators⁴.

With its innovative approach and commitment to creator success, OpenAI is well-positioned to become the dominant force in the rapidly evolving AI creator economy. The GPT Store has the potential to unleash a wave of innovation, as creators harness the power of GPTs to develop a wide range of intelligent tools and applications. The GPT Store is expected to launch later this month, and interested users can sign up for early access at chat.openai.com/create².

Some use cases to spark your imagination:

WEEKLY DIGEST
Grok AI, Musk's New AI.

Elon Musk unveils Grok, an AI chatbot with a ‘rebellious streak’Grok is modeled after The Hitchhiker's Guide to the Galaxy. Grok is designed to answer almost any question and even suggest what questions to ask.

What’s New: 

Grok has recently been updated with a new feature that allows it to learn from its own conversations and improve its responses over time.

• It has access to user posts on X, which Musk owns, and has a penchant for sarcastic responses.

• It will be available to X's Premium+ subscribers after a series of beta testing

• Grok gets its real-time knowledge of the world via the X platform and will answer spicy questions that are rejected by most other AI systems

Hands-On Experience: Grok is not yet available,

Why Does It Matter?

Grok is intended to answer questions with a bit of wit, making it a unique and valuable addition to the AI chatbot market.

However, Musk recently shared a screenshot of Grok offering a “humorous” step-by-step guide to making cocaine, which included instructions like :

ELon followed it up with a separate screenshot offering more detailed instructions on how to produce the Class A drug from its raw materials.  

ACTIONABLE TIP
Guide to create my first GPT in 5 Minutes.

Context: I am utilizing ChatGPT Plus (the paid version), which grants me access to the GPT-4 and DallE-3 models, as well as an interpreter to execute code and generate files.

We will employ ChatGPT to develop our own GPT models. This is also possible through the use of the developer interface to create Assistants.

On the 'home page', this is the envisioned layout: The left panel displays the GPT models I have created and the pre-trained GPT models I have tested from OpenAI

Step 1:  After selecting 'Create GPT,' you will be directed to the following screen

To define the behavior of the bot, you must supply specific prompts or information. These will be in natural language for ease of understanding. To illustrate this process, I will input a simple example prompt into the code interface.

I want to create a social media manager that helps me with these actions and more:

Offers suggestions according to the provided brand description
Generates a list of topics and formulates ideas for each one
Facilitates post creation for Twitter, Facebook, Instagram, and LinkedIn
Creates optimized post bodies automatically
For posts on LinkedIn, proposes AI images for each post using DALL-E 3
Selects hashtags intelligently
Incorporates emojis in the posts
Uses a professional writing voice and tone, while keeping it cool to avoid annoying the user
Avoids using industry-specific language and terminology, and provides detailed and accurate information when required. But supports arguments with statistics, research, and expert opinions.

Step 2: Now, personalize the bot by adding characters and visuals. Below is my prompt for this step

provide 3 alternatives for a cool name for this bot

I have chosen 'Brand Whisperer' from the three options provided

Set the name 'Brand Whispere'
and generate a suitable profile picture for the bot

Set it as a profile picture for the bot

And that's it, your bot is ready for a spin. Don't forget to save it.

Would you like to test my bots?

Meet Brand Whisperer:

• It enables the transformation of articles into engaging LinkedIn posts.

• It suggests post ideas based on current trends.

Keep in mind, this prototype was created in just five minutes, so while it's not perfect, further training with more detailed prompts requesting specific features or specialized behaviors will certainly enhance its value.

Tip: You Can Define Special Actions

If you return to the interface to modify your GPT, you can add a new feature. Here's how I requested it:

Desired Feature: On the bot's welcome page, I would like to include a button. When clicked, the GPT will prompt the user for an article. Upon receiving it, the bot will generate LinkedIn post ideas tailored to three distinct personas:

1. Business Person

2. Technical Person

3. Non-Technical Person

This approach isn't utilizing any advanced marketing techniques; it's simply an exploration of the bot's capabilities through somewhat random feature testing.

Create Feature: 'Transform to Posts'

Description: Take the provided article and create 3 alternative LinkedIn post ideas tailored for different personas - a business persona interested in practical applications, a technical persona interested in technical details, and a non-technical persona focused on concepts and implications.

Requirements:

Generate posts with word count appropriate for LinkedIn
Business persona post should focus on real-world applications and benefits of the technology
Technical persona post should include technical details and analysis for engineers
Non-technical persona post should explain concepts simply using analogies and everyday language
Incorporate relevant hashtags for each persona
Use an appropriate tone and style for professional social media
Article:
[Insert article text here]

Please generate 3 LinkedIn post ideas based on the above.

You can also attach documents or refer to third-party services like Zapier to trigger further action. We will be playing with that, stay tuned.

This is the result

Feel free to explore and interact with the results

TRAINING
💡 THE Microsoft's Generative AI Course

What's New

Dive into the world of Generative AI with Microsoft’s new course, crafted into a concise 12-lesson curriculum. You’ll journey through the fundamentals of prompting, vector search, text and image generation, and the principles of building responsibly.

Each Lesson Features:

• An engaging short video introduction.

• A comprehensive written lesson in the README.

• Code examples in a Jupyter Notebook for project-based lessons.

• Practical challenges or assignments to solidify your knowledge.

• Additional resources for extended learning opportunities.

Key Features:

• Gain practical experience in deploying Large Language Models (LLMs) for robust Generative AI applications.

• Hone your skills in prompt engineering to fine-tune AI model responses and capabilities.

• Leverage Azure OpenAI services for scalable, enterprise-grade machine learning projects.

• Explore vector databases and learn to harness embeddings for cutting-edge semantic search in AI.

• Understand the intricacies of secure API management and commit to ethical AI practices within your engineering processes.

Embark on Your Learning Adventure:

Thank you, see you next week!