OpenAI's recent announcement, allowing the creation and sale of custom ChatGPT versions in their store, mirrors Apple's approach with the iPhone App Store with one key difference, you don’t need to be a developer to cash in.

This is a big deal as it widens the entry point for developing products with AI.

In the Open AI Dev Day Keynote by CEO Sam Altman, he demonstrated how easy it is to create a Custom GPT as they are calling them.

Custom GPT creation is done in 3 Steps:

1. Instruction

2. Input your own IP

3. Name it and post in the Shop

Pretty simple.

So simple that we will likely see millions created for every possible use case haha

The GPT Shop is expected to open at the end of November.

Grok is a term coined by American writer Robert A. Heinlein for his 1961 science fiction novel Stranger in a Strange Land.

Wikipedia defines it as meaning"to understand intuitively or by empathy, to establish rapport with" and "to empathize or communicate sympathetically (with); also, to experience enjoyment"

So what is this new Chatbot?

If you don’t use X/Twitter you’ll probably never know haha. Its a new Chatbot that was created in super fast time by the team at X with input from Elon Musk. Musk’s contribution was to give Grok all the humour they could input, primarily from his favourite book ‘A Hitchhiker’s Guide to the Galaxy’.

Grok has a serious mode too and apparently very capable, enough to be a challenger to ChatGPT. This is yet to be seen as most of the commentary around the Chatbot has focused on its humorous responses that Its critics & competition have labelled as ‘Dad Jokes’ & ‘Boomer Humour’. However, it is nice for a change to have a Chatbot that is allowed some intellectual freedom unlike ChatGPT which has been highly hand-railed.

All things considered though, Grok adds more to the landscape and proves that LLMs can be produced relatively quickly.

Finally, the last key difference is in Grok’s constant evolution. The model is drawing upon the whole X database in ‘Real Time’. Unlike other LLM’s it will respond to current events without doing a search.

I wonder, given this feature will Grok become a conglomerated personality made up of all those on X?

A scary but intriguing thought 🤔. With some of the commentary around ‘walled gardens’ of data we may see this with other platforms.

I asked ChatGPT to give me a few Platforms Personality Profile as an AI chatbot:

Reddit - A chatbot programmed with every conspiracy theory, cat photo, and unsolicited advice known to humanity, always ready to argue why pineapple does belong on pizza.

Facebook - A chatbot that constantly reminds you of people's birthdays you don't care about, shows you ads for things you mentioned once, and keeps suggesting you befriend your ex's new partner.

Instagram - A chatbot obsessed with filters and hashtags, always asking if you want to see another perfectly staged photo of avocado toast or a sunset yoga pose.

Mmm I should have put some effort into this myself 🙄, but you get the idea.

As I mentioned in the opening to this newsletter there have been significant enough advancements in AI that mean it can be applied to just about everything we do.

What is the current Base?

Language - Everything we do can be explained in a shared Language. A powerful tool that goes deep. From technical instructions to writing our Laws to expressing our emotions, we have used language to build everything we do. Language is knowledge recorded or shared.

Image - A way to share ideas, explain things, record events and attract attention.

Funny enough, both of these also form the basis for storytelling. What some would argue is the deepest layer of our civilisation.

The applications of the two has amazing reach and in somewhat unexpected ways too. For example, Language Models are capable of writing Code. Because it is predicting the next word in a sentence that best relates to all the other words that precede it, it turns out this simple logic works for code as well. Just think, all of our Digital Infrastructure (the Matrix haha) is based on a Language(s) that Chatbots can write too.

So with Language, Image and even Code Generation there is nothing that Machines cannot help us with. We now have a powerful and effective non-human true collaborator.

From what I’ve seen talked about in the Research space for AI, this base layer is just the beginning.

One area closer to being a realised is in Systematic Generalisation Models. They can process information like human cognition, i.e. ‘understanding’ the concepts it is writing and responding to rather than predictive text. Check out this story to give you an idea.

What did you think about this weeks Newsletter?

Did we nail it? Can we do better? Want to know more? Hit reply and let us know!

If you like our content and think it might be worth a shout out, please do!

Right click the below button and share the link.