If you thought 2024 or even 2023 were big for AI, then you’ve got a shock coming.
AI, or at least the talk about AI, has been around forever. Most Google searches over the last twenty years have used AI – but we all called in Machine Learning (ML) back then.
And then ChatGPT landed and we had to recalibrate our definitions and our expectations of what AI means to society. It was Autumn 2022 and AI woke up through the creation of Generative Pre-Trained Transformers (GPT).
It was the first real glimpse of man’s greatest invention, and the start of the next stage of societal development. The “end of the beginning” if you will.
Less than a year later, all the world’s technology rapidly started to pivot towards AI. And the big money came rolling in as the blueprint for a brave new world came sharply into focus.
The major AI players are rushing headlong in a race to build-out millions of TPUs … which will then be handed over to a new generation of pre-training.
It will differ in subtle but very significant ways …
- in its scale. Several orders larger and faster than GPT4.
- in unsupervised learning. With only rudimentary guardrails in place, GPT will self-learn true multimodal comprehension of our world.
- due to iterative learning. Rather than pre-training from a single data sweep, GPTx will operate in a continuous learning cycle
- it will not stop, it will speed up
- it will fully reason based not just on the world’s existing knowledge – but create new knowledge at a rate that is difficult to comprehend.
- the arrival of agentic-AI will see AI not just responding to human input with digital out – agentic AI will carry out actions in the real world. And then learn from those real-world interactions to “improve” itself.
- it will make GPT4 seem as antiquated as steam trains as the current narrow intelligence model evolve in months to be truly super-intelligent.
Buckle up folks, it’s coming real fast and it’s going to be one hell of ride.
