More AI Ramblings

I spent some time writing a blog on AI and it's ended deciding to post as three separate blogs, the first being Are We Insane? This one is a follow-on to Are We Insane. The third one is just me having some aMUSEment.

In my own world it is nearly impossible to imagine the kinds of resources being used to develop AI to remedy environmental problems, and create a sustainable future. Like most technology, AI will contribute to technical problem solving, but will it provide solutions to humanities foundational problems? In general, technology is good at solving problems, but far less capable at creating solutions.  As someone whose concerns focus on the effects of climate change, the  numbers being spent on AI, trillions of dollars, are nearly unimaginable.

To cite one example, an area of particular interest to me is determining if we can supplement anemic parts of the world’s ocean, conservatively estimated at around a third, with iron to increase phytoplankton productivity. Because phytoplankton use sunlight to capture carbon dioxide and turn it into biomass, increasing phytoplantkon productivity could pull carbon dioxide from the atmosphere and transfer it to the deep ocean for hundreds to thousands of years, thus reducing the likelihood of catastrophic climate change. There’s geological evidence that this iron fertilization has worked in the past to exert significant, but by no means full control, on climate variability that  To fully research this approach would cost around 500 - 750 million dollars, and, if it is shown to work, scale-up would be on the order of 10 - 50 billion dollars over, let’s say 15 years. So, conservatively, this approach would cost way less than a penny on the dollar compared to what will be spent in the next 15 years on AI based on current trajectories, i.e. trillions. Furthermore, that fraction of a penny won’t be spent, because in the grand scheme of things, with AI being the grandest of grand themes, developing carbon capture methods is deemed unimportant and irrelevant. For another perspective on this, according to Google AI, Apple’s net daily profit in 2025 was around 250 million dollars, so if Apple donated 3 days of profit to this cause, they could fund it entirely. Google (or Alphabet) could probably fund it in two days, and Nvidia in a day.

Of course, the AI evangelists have technical answers for all the problems I would deem important and worthy of putting at least a few fractions of a penny of each dollar we will spend on AI in the next decade. The energy problem is simple, super AGI will quickly figure out how to successfully do fusion power, giving us an unlimited supply of clean energy. Global warming will be solved by designing atmospheric injections of fine particulates to fine tune Earth’s energy balance by scattering the amount of sunlight reaching Earth’s surface in ways so subtle no one will notice. AGI will fine tune our economic inputs and outputs to produce global equity, balanced toward wealth for the masses. Warfare will disappear, because AGI will simply make it so. Of course, another fascinating outcome for Homo sapiens, is if super-intelligent AGI decides its prime directive is to achieve ecosystem balance, and the most efficient course would to remove 90% of humans from the planet, starting with the one's who use the most resources per capita!!

So it’s an odd world we live in where these days, where it feels like one of the least likely scenarios is that we will simply evolve human consciousness to develop human solutions that will harness technology to the betterment of humankind and all life we share this blue sphere with, but not be driven by technology. Now that’s seriously insane.