The curious case of lacking talent

If one is in the space of AI startups, one is mostly busy understanding the latest algorithmic approach to improve model training throughput, or connecting developer centric workflow tools into an AI enabled machinery to increase some other performance characteristic of deploying inference in cloud-native apps (insert cool technology here), or of course - if you are consumer-facing - it's about the latest application of AI to our everyday lives: generation of images, answering queries, doing work on your behalf, answering phone calls and overall improving our human-machine interaction (blast from the past!) beyond the issuance of computer programming commands.

However, recent experiences have profoundly altered my opinion regarding the relevance of any of these factors. No doubt all these start up efforts are important in their own right; and as an entrepreneur, I suppose it doesn't really matter how you make your money. But I can't help but see blissful ignorance in the Silicon Valley, self-enamored with their own technologies and buzzwords. One could write it up to aura-farming in the eyes of other technologies, but ... 67 (this is me trying hard).

Like any technology, unless it is useful to humans, it is not turning into a product. However, if the product is sufficiently complicated to use, either as a result of its inventors overly optimistic estimation of its' customers' ability to apply it, or because it simply is too hard, the product will be "ab-used". It will not be used in the way its inventor imagined, or according to how it was designed, or for the purpose it was intended. Assuming it survives at all. Accidental duct-tape, to give it a positive spin, SPAM, if you are less gracious. The point being that unless you have the ability to use the product, it will take weird turns in the course of its lifespan, in many ways forge it's own path into the consumer's conscience.

AI absolutely has that potential. ChatGPT can be a better Google Search for 90% of the people, leaving 90% of its capabilities on the table, without anyone the wiser. I suppose this bifurcation between those who are interested and capable and those who are either not interested or not capable (or both) has always existed when it comes to computers: nerds vs non-nerds. However, this is not the case with electricity; few know how to make it, some do, but most of us understand you can't exceed the amperage of a breaker without tripping it, and beyond that we use it all the time, even as lay persons. Non-geeks. No geekery necessary to use electricity.

Well, the issue is AI unfortunately is like electricity in the sense that it will, truly, be everywhere. It may not be prominently visible, but we'll all get used to it to see things "AI powered" then those labels will disappear, until it's just - of course it's AI powered, is there any other way? Lacking the talent, sociologically, to not abuse AI it all its shapes and forms, therefore becomes a much more central issue for us all. We are, literally, at risk of designing a world experience that is incomprehensible to most of us, and the reasons for why it will happen anyway is because it is just like electricity. Companies, countries, states - none will be able to afford to leave out an AI strategy, and apply it. It will happen.

States see this very clearly, actually, and realize that unless comprehensive workforce development programs are put in place, the intended effect of adding AI will not improve ones productivity (GDP) - it will reduce it. Addressing this, and creating solutions that prevent the negative impact of adding AI to an ever-more complicated world, seems like a huge opportunity.

I have yet to meet a startup that understands this, and plays off that angle. I guess, it's more sexy to be "agentic". But if VCs really are looking for sustainable investments, they should look for comprehensive playbooks that do more than just explain the new MCP protocol addition. They should look closer, and understand that unless the solutions are impactful at scale across society, it won't last: "Of course it doesn't work, but look how fast it is!"