Dispatches from the AI Frontier — An Expedition into Generative Intelligence

Most people will have heard of some version of generative AI. Likely that will be ChatGPT — even my parents had heard of ChatGPT.

Generative AI has been open to the broader public since OpenAI launched ChatGPT in November 2022 (according to ChatGPT). Others have joined in as well.?

Bing Chat is another generative AI. It opened up to the public on February 7, 2023 according to Wikipedia.?

The other generative AI player — with household name status — is Google. Bard is its generative AI. According to Wikipedia Bard was opened up to a limited audience in March of 2023.?

• It’s not available in Canada, but using a VPN I’ve been able to benefit from using Bard

These three — ChatGPT, Bard, Bing Chat/Copilot — are the ones that have come up in conversation when I’ve talked about generative AI with clients.?

It’s maybe been 6-7 conversations, and that's been with clients that are either IT, or Marketing types of roles, and in Higher Education and Not-for-profit verticals. I wanted to give this context, and recognize that the topic of generative AI is enormous? — I’ve only started my exploration!

Generative AI has exploded with a really big bang! It’s already forming galaxies of code & applications, and blossoming into uncharted constellations. It is transformative in nature. Generative AI has vast potential. Some people think it possesses the ability to redefine our understanding of the world.

I encourage you to explore for yourself.

Here’s what I’ve discovered.

There are many other noteworthy generative AI models. Each with their own strengths and areas of focus. It’s not just about digital marketing, SEO and content strategy — though this is where I spend most of my time. Here are a few other examples of generative intelligence:

Text Generation:

Jurassic-1 language models: the largest and most sophisticated language model ever released for general use by developers. Apparently it excels at generating different creative text formats, like poems, code, scripts, and even musical pieces.

Megatron-Turing NLG: This model from Google AI boasts impressive fluency and factual accuracy, making it ideal for tasks like generating product descriptions or news articles.

WuDao 2.0: Developed by the Beijing Academy of Artificial Intelligence, WuDao 2.0 focuses on Chinese language tasks and showcases proficiency in poetry generation, question answering, and dialogue.

Image Generation:

DALL-E 2: You must play around with this. It’s interesting. It’s fun. It can be frustrating — more on this later. DALL-E 2 is an impressive model from OpenAI. It can generate highly realistic and creative images based on textual descriptions, pushing the boundaries of text-to-image generation.

Midjourney: Maybe the image generation algorithm that most people have heard of and experimented with? Have you experimented with it yet?

Stable Diffusion: An open-source diffusion model, Stable Diffusion offers incredible image generation capabilities and has led to numerous creative projects and communities.

• Imagen: Developed by Google AI, Imagen excels at generating photorealistic images with specific artistic styles and intricate details, allowing for high levels of customization. Looks like it’s only “enterprise-ready” and not available to the general public.

Other Domains:

MuseNet: This model generates impressive musical pieces based on text descriptions or existing samples, showcasing the potential of AI in music composition.

Jukebox: Another music-focused model from OpenAI, Jukebox specializes in recreating musical styles of specific artists, offering a glimpse into AI-powered music generation.

AlphaStar: This AI agent developed by DeepMind has achieved superhuman performance in the StarCraft II real-time strategy game, demonstrating the potential of generative AI in complex simulations.

AI Test Kitchen: There’s the TextFX tool which is one that I want to experiment with for creative writing, and there’s the MusicFX tool.

Here’s what I’ve done with generative AI

• Research & learning
• One of the areas that has been very useful for me has been in coding — I’ve successfully used ChatGPT and Bard to write javascript, regular expressions, and python
• Topic ideation and keyword research AND in clustering topics and keywords
• Generating outlines for presentation topics?
• Creating executive summaries
• Writing SEO titles and meta descriptions — this was a test only. I wasn’t really that impressed with the results. I feel I’m more effective and produce better results when I sit down and focus on the task.

In Conclusion (for this stage)

So, using the metaphor of a journey of exploration, today marks the setting up of my first basecamp.

Here’s what I’m certain about:?

• Generative AI is a tool, albeit a powerful one with untapped potential, but it's only a tool
• If you prefer to think of it another way, then try ‘assistant’ (see NN/g The 4 Degrees of Anthropomorphism of Generative AI)
• To become proficient with this tool, I’ll need to learn and practice — I need to think of it like it’s a craft, like I’m a apprenticing to be a blacksmith?
• I feel no hint of threat or uncertainty about my career and job stability because of AI — actually excitement, and optimism

There’s loads of terrain to cover — it would be like coming to the edge of our solar system and looking at all the planets and moons and setting up basecamp — so much to explore!

Here’s what I’m focussing on for the next leg of the journey:

• Becoming more effective at prompting — becoming a certified prompt engineer
• Working out the best way to use generative AI for my team and my clients

Side note:

At the start of this article I mentioned something about the frustration I’ve felt using DALL-E. I think this is a little tiny bit about me learning how to be more effective at prompt engineering, and more about DALL-E learning how to listen.

In full transparency, here’s the conversation. Note: I still ended up with a globe! The desk was less cluttered and more organized, but that globe!!?

Until my next dispatch — here’s looking at an amazing 2024!