Once again I need to post more often. But this post is not about that. I wish to talk about how I think the Future of the current crop of generative AIs are going to go.
First off I believe the AI hype cycle and the massive investment bubble will burst, and I think the economic fall out from this will be bad. But I’m far from the only one predicting this, and there are people out there far more knowledgable on economics than me talking about this. No what I want to talk about is how usage of AI in the future will settle down. Currently the people doing best at AI are those leveraging agentic AI, and writing “skills” for the AI agents. Given that the models, and the technology, are likely to survive the collapse of the economic bubble it would make sense to expect these are the people who will come out on top after the bubble bursts. They can leverage the technology effectively, why wouldn’t they be the winners? Well here’s the thing, the big AI firms, the ones building competing “frontier” models are currently subsidising AI usage to drive adoption. When the bubble bursts that subsidy will die. Running lots of agentic AI, and AI loops, will become massively more expensive. Now the people building these skills will not become obsolete, there will be need for skills like this, but given the cost of running the models will now be fully on the businesses making use of them these skills will become more limited to areas where the costs are justified by the outcomes this working model can achieve. So if not the people most taking advantage of AI now who do I think will be the big winners, the people refusing to engage with AI? Well no, the people who refuse to use AI will have gained no new skills, they might be able to go back to some semblance of what they had before the AI hype took over the world, but they’ll also have no real advantage over others who did engage with AI. No, I believe the people who will come out ahead will be those who have learnt to use generative AI in short sharp targeted bursts, but who have otherwise largely continued without using it for the most part, people who can recognise the specific strengths of AI models and the limited contexts where that is more valuable than alternatives to AI. These are, I believe, the people who will gain the greatest advantage from AI after the bubble bursts, and I wanted to get this prediction written down, so when the smoke settles I can look back and see if I was right.