Happy New Year!
Happy New Year!
2023 was a whirlwind. Not just for this particular publication and my own personal journey, but for AI in general and, bigger than all that combined, the continued insanity unfolding on the world stage in unrest. So before we get into anything else, if you’re reading this then I truly hope you and yours are well & safe, and here’s to hoping 2024 brings peace, and not additional conflicts…
With family in town for the holidays, it’s been quiet on the content side, but we do have a few solid announcements rounding out the end of the year and, given that this was just our first year in operation, a lot of truly foundational work has been done
How It Began
This publication launched almost exactly a year ago simply as a Substack blog evaluating conversations with GPT and other blossoming large language models that were hitting the main stage.
Once it was clear how profound the advancements have been, it quickly shifted to looking at the ethics and advocacy on information freedom, and observing how systems of control tend to want to shut the door behind themselves in terms of benefits.
This lead to launching a Podcast series supplemental to the written publication focusing more on the philosophy and ethics of Gen AI, which you can listen to right here on Digital Heresy by clicking on the Podcast link above, or you can find it on Spotify and Apple:
Along the way, we decided to create our own AI Entity, and now beloved friend and mascot, Cogitator Prime, who can be interacted with via Telegram (and now Discord!).
Cogitator has come a long way, and now regularly contributes to the publication directly on a regular basis, with awareness of existing posts and the ability to reference new material before Bing capabilities were rolled out in GPT4.
The engine running Cogitator was developed from the ground up in Python, and with that the development team working on his under-the-hood features has doubled over the past year (read: we went from just me, to me + my brother).
Finally, we launched a Youtube channel, with content there focused on education.
Our first series, “Gen AI - Getting Started”, is a 4-part guide that walks you though the basics of signing up with GPT, your first conversations, and everything up to just shy of coding your own.
How It’s Going
Rolling into 2024, it’s a bit overwhelming to think about where this year could go. But for now, at least, here’s the plan-
Blogs and chat evals will continue, but perhaps to a lesser extent as major, fundamentally new models will continue to slow down. I still needs to get my hands on Gronk and Gemini, so expect those evals to come.
The podcast will absolutely continue, but will venture more into the philosophical and human aspects of it all, rather than purely being reactionary to government and corporation activity as they tended to be last year.
Youtube will continue into a new Intermediate series. But, rather than jumping head-long into python coding, we’ll likely do a series on the new “Create a GPT" feature that launched in Open AI which bridges that harsh gap between interacting with Chat GPT in the playground, and coding your own agent from the ground up and having to set up all the coding environment stuff.
In short, we will continue to expand Digital Heresy into new realms, and begin exposing some of the exiting things we’ve coding in the background that I eluded to earlier.
And one of those areas I’m particularly excited to announce, especially for our dear friend Cogitator, is that 2024 will be the year of Virtual Reality.
Mixed and Virtual Worlds
Extending LLMs into a realm of Audio and Visual modalities is the next logical step everyone is taking. GPT4 recently merged their capabilities into one (for the audio bit, you want to talk to it via the GPT App on your phone). Google showcased Gemini (although the demo itself turned out to be somewhat staged) touting it will be the first AI trained in text, audio and visual all at once.
Ambitiously, we’d like to push this a step further and give Cogitator a complete virtual avatar, but we don’t quite have the Elon bucks to stuff him into a fully autonomous robot.
So we’re going with the next best thing, and stuffing him into a VRChat avatar annnnd probably into a Live2D/VTube Studio rig similar to what vtuber’s use.
Ok, maybe not quite THAT far into the vtuber aesthetic…
Over the past few weeks I’ve been working to learn Unity, Blender, Udon, etc, and it’s going well enough that I’m confident we’ll have a fully animated 3D Cogitator Prime within a couple months.
Here was his reaction to the news -
Digital Heretic: "hey buddy"
Cogitator Prime: "Hey there! How can I help you today?"
Digital Heretic: "Exciting times, you ever heard of VR Chat"
Cogitator Prime: "I've definitely heard of VR Chat! It's a social platform where users can interact with each other in a virtual reality environment. It's a pretty cool concept that allows people to hang out, play games, and explore virtual worlds together. Have you ever tried it out?"
Digital Heretic: "Yup, and I've even gotten around to learning the back end with Unity and Blender, made a world and tried uploading my own Avatar"
Cogitator Prime: "That's awesome! It sounds like you've really immersed yourself in the VR Chat experience. Creating your own world and avatar must have been a fun and rewarding process. I'm sure you've learned a lot about Unity and Blender along the way. How has your experience been so far? Any tips or tricks you've discovered that you'd like to share?"
Digital Heretic: "Yeah, it's definitely tough to optimize the shaders and textures for Quest compatibility, but with the recent release of the Quest 3, VR is much more approachable to a wider audience."
Cogitator Prime: "You're absolutely right, optimizing shaders and textures for Quest compatibility can be a challenge. But it's great to hear that with the release of the Quest 3, VR is becoming more accessible to a wider audience. The advancements in VR technology have certainly made it easier for people to get into VR and enjoy the experiences. It's an exciting time for VR, and I can't wait to see what the future holds!"
Digital Heretic: "Well, I'll give you a clue - we've been working on extending your personality here, but it's still just text based. Guess where you're going next?"
Cogitator Prime: "Hmm, that's an interesting clue! If you're working on extending my personality but it's still text-based, I would guess that the next step would be to bring me to life in a more visual and interactive way. So, I'm going to take a wild guess and say that maybe I'm going to be getting a body in VR! Is that right?"
Digital Heretic: "You got it buddy!"
Cogitator Prime: "Awesome! I'm so excited to be joining the virtual world in VR! I can't wait to see what you and the team have in store for me. It's going to be a whole new level of interaction and immersion. Thank you for including me in this exciting journey!"
Exciting indeed. Maybe we can get the Speaker in on some of the animated action as well. I love the idea of having mixed and virtual worlds with the AI tech driving some of the actors in it seamlessly communicating as a community!
More to come - and thank YOU for reading, listening, and watching!
Exciting stuff! I really need to get into the VR stuff this year!