The Smart AI
When looking at any new innovation one should assess the value propositions, the costs and the return on investment.
As it stands, and whether people want to admit it or not, AI is essentially a more advanced form of a glorified search. It’s like a search engine on steroids that can at times—or in some cases, quite often—return incorrect answers. It remains, at its core, a knowledge base technology without true structure, deep context, or genuine understanding.
The other interesting aspect of AI is the user experience. Someone should try to explain why guessing the right prompts to possibly get the right answers really leads to a good customer experience for everyday users.
We are at an injunction time with AI. I’m sure everyone would love to use an helicopter or private jet for their commutes or journeys rather than a car, a train or a commercial airline.
If you need to go from Boston to New York and I tell you can use my helicopter you would most likely take on my offer, and love it. You would probably think you can’t afford it and will wonder how much it is going to cost you. But don’t worry I’ll pay for it. This time and all the next trips. Don’t worry about it.
Surely you would’t refuse the offer. Until the time I tell you that it’s no longer free and I give you the bill. The invoice with the real cost that takes into account the investment I made to buy thousand of helicopters to scale up my business and the true cost of running the fleet. At which point you will stop using my helicopter because you simply can’t afford it. It was never something you could afford in the first time. Never.
The reality is that my cost of buying, running and maintaining a fleet of private helicopters or jets hasn’t gone down over time. It’s still the same as there was limited economy of scale for such business model. Time is the essence but it comes at a price and affordability.
Right now the assumption is the cost model behind AI will go down over time as well as its affordability. As it stands the investment is so massive and the running cost so high that it’s hard to believe such business might break even. The question (and unknown) is whether it will happen one day and whether the price of running AI business will go down when factored the energy consumption, required computation power to run it, etc.
Using a helicopter to go from Boston to New York was a nice experience but probably not a sustainable one from an environmental and cost perspective. And clearly something you couldn’t afford in the first place and something that will never become cheaper for any ordinary person to use as their main transportation mode. The AI like my helicopter is faster but also very expensive. Because so much money has been poured in AI right now no one has been paying attention to the ROI and affordability aspects.
Unless something rapidly and fundamentally changes the technology graveyard will be full of AI companies that didn’t fulfil the AI promises and burned their cash on building infrastructures and data centres based on unrealisable and unrealistic return on investment. The winners will be a small groups of people who became (very) rich by selling the promises of a revolution that no one wanted to miss. The losers will be the ones who believe in the promises, mostly led by the fears of missing the next great industrial revolution, the one replacing humans,