As I researched and wrote my OpenAI and LLMs blogs (see Introduction to OpenAI and LLMs, Introduction to OpenAI and LLMs – Part 2 and Introduction to OpenAI and LLMs – Part 3, along with a presentation on that topic that I did for the Toronto Data Professional Community which you can view and download the slides), I found and played with many fascinating AI products and features. I am continually amazed at the progress we are making with AI, especially those in the GenAI world, and feel like we are just getting started. Here are my favorites:

HeyGen – Speak to an avatar live. Check out their demos where you can interact with avatars such as a therapist, fitness coach, and doctor. They keep adding new ones. Amazing stuff!

OpenAI Sora – Create video from text. My favorite is Historical footage of California during the gold rush.

GPT Store – Discover and create custom versions of ChatGPT that combine instructions, extra knowledge, and any combination of skills. My favorites are Books, Movies, and Therapist/Psychologist.

ChatGPT advanced voice mode – Have spoken interactive conversations with ChatGPT, where you can screen share and also share live video. I had a long conversation with ChatGPT while driving alone to help keep me from getting sleepy – we discussed the best Yankee teams of all time. The ChatGPT voice is animated so it’s like talking to real person (choose from nine lifelike output voices for ChatGPT, each with its own distinct tone and character). Screen share can be used for things like help guiding you through settings on your computer in order to help troubleshoot a problem, while live video can be used in ways such as helping you recognize objects or tell you the color of a shirt (great for colorblind people like me).

ChatGPT DALL-E 3 – Create images from text. This used to be a separate product but is now built into ChatGPT. I recently asked it to draw me a picture of a crawfish blowing out a birthday cake for a crawfish boil I did for someone’s birthday and sent that picture with the birthday invitations.

Azure AI Speech Studio – Create a custom text to speech avatar using the Azure AI Custom Neural Voice and the Custom Avatar Self-Service capabilities in Azure AI Speech Studio. You can use the video and voice of anyone you wish. I hope to one day create an avatar that looks and sounds like me and then talk to it. Freaky!

Clonos – Create your own virtual avatar from existing video and sound, and make it say anything. See examples of Clonos in the sports world that is very funny at memerunngergpt on Instagram, where they modify video from sports figures to say hilarious things in their own voice (warning: foul language).

Microsoft Copilot – An AI chatbot that is now in many Microsoft products. Check out a few of my favorites: Copilot in Teams, Copilot in Microsoft Fabric, Copilot on Windows, Copilot in Word, Microsoft 365 Copilot Chat.

If you have not used ChatGPT, you need to do so immediately! One way to use it that you may not be aware of is via roleplay, where OpenAI takes on a persona. This is a great way to learn about a subject matter. For example, I am reading about the crusades, so I prompted OpenAI: “Pretend you are a knight from the first crusade and fought from the very beginning to the very end of the crusade. I will ask you questions, and I want you to answer in the first person. Draw upon historical knowledge and accounts of the first crusade to immerse yourself in the mindset, beliefs, and experiences of such a knight, responding to me as though you were truly from that era”. Then I was able to ask questions directly to a “knight” and get all sorts of great info. Other amazing persona’s you can ask it to emulate are such things as “pretend you are SQL Server and I’m using SSMS” and “pretend you are the game Zork” (for you old-timers out there like me).

I also used ChatGPT on my iPhone and told it to pretend it was Santa Clause and to talk to my 6-year old grandson. I turned on voice mode, which has a seasonal Santa voice, and my grandson had a long and animated conversation with Santa (ChatGPT)!

To see some of the capabilities of OpenAI and ChatGPT, check out OpenAI YouTube that has demo videos. My favorites are Interview Prep with GPT-4o, Live demo of GPT-4o’s vision capabilities, and Interview roleplay with GPT-4o voice and vision.

A helpful tip on asking questions with ChatGPT: you will get wrong answers sometimes, especially if your question (“prompt”) does not have much information. If you get a wrong answer, tell ChatGPT via a follow-up prompt that the answer is wrong and tell it what the correct answer is. If it then responds with the correct answer, prompt it with “change my original prompt to prevent the incorrect answer if I were to ask the question again”. You will then get an improved prompt that you can use to ask the question again, or to build upon it.

More info:

AI Companions Will Change Our Lives