Humanity's problems won't be solved by Sam, Dario and Elon

A Substack post from Vinz brought my attention back to something I have been meaning to write about for a while. The emergence of AI, and its creators, as the all-conquering, infallible new technology that is, according to its cheerleaders, going to bring the end of humanity. Dario, Sam, Elon & co basically want to convince us that AI isn't a simple optimisation of using computing power in creative ways to use existing data repositories as the solution for all of mankind's problems. Wild claims have been made by these people, a recurring one is that AI models are more intelligent than people. 

First, we need a definition of human intelligence and how and whether it is different from machine intelligence. It is not up to debate that computers can and do process vast amounts of information and data faster than humans. That does not make computers or AI more intelligent than people, specially when the information and data upon which AI relies on has been produced up to this point and in collective manner entirely by people. AI isn't inventing anything that isn't already there. There's been a number of "breaking news" about how AI solved this or that mathematical problem, some of which that hadn't been solved for a long time. AI's contribution to the field has been to analyse and process existing data and then, on the back of work already done by humans, announce that it had come up with the solution.

OpenAI is in a two-participants race to prove it's the best thing out there. It also "found" a PHP vulnerability that had already been found by programmers and proceeded to leak its "findings", leading to global internet disruptions. The same formula is being applied now to Linux and about every other programming language in existence. Debugging at an industrial scale therefore is why AI proponents believe it is "more intelligent" than humans. This is complete bullshit of course, for what is debugging but finding flaws and errors in writing? It's a glorified spellcheck.

People unaware of the fast and immense processing power capacity of computers are surprised by what AI can do at its most basic, like cleaning inboxes with thousands of emails, produce travel itineraries or coming up with low calorie diets. As I said to someone the other day "wake me up when AI comes up with the solution to corruption or world poverty, or how to fix democratic deficits and plug fiscal holes." In my opinion, none of AI's most talked about achievements are remarkable. It hasn't invented anything that wasn't already here, and therefore, all the talk about how the legal profession is about to go extinct, for instance, because some models have been trained on Westlaw is complete nonsense. Any chatbot can come up with reasons why a drain gets blocked, but the plumber is still needed until every household on earth has a robot that can fix it. All programming languages, all literature, all mathematical conjectures, all laws, etc., surely contain an error here and there, some vulnerabilities that should be patched, or a new combination to be arrived at after applying brute force attempts that would otherwise take decades. Quantum computing will undoubtedly take things to the next level. But again, that's nothing more than computing power. It is neither creative nor an invention. Any "new" announcement comes on the back of catalogues that humans have been collectively producing since the beginning of times. How can anyone then argue that AI is a 100 times more intelligent, when the very existence of AI is but another manifestation of human creativity?

And when it comes to the future of humanity more broadly, predictions, etc., why would anyone with a thinking brain listen to Sam, Dario or Elon? What credentials do they have in the fields of philosophy, economics, law, psychology, anthropology, etc.? Why would anyone pay any heed to Jacob Coxon's claim "No other human activity poses this level of danger"? Tell that to Ukrainians Jacob, convince them that AI is more dangerous than Elon's unilateral decision to switch of Starlink. Uyghurs and Palestinians would probably love to hear your absolutist theories about how AI is going to make their lives any worse.

Years ago, I got into a debate with Noam Chomsky. At the pre Epstein revelations time, Chomsky was widely regarded as one of the world's leading intellectuals for his contribution to the field of linguistics. Like Sam, Dario and Elon, Chomsky never lacked groupies and a media eager to surrender to his "towering intellect". But Chomsky was, is and will always be someone who came up with groundbreaking thoughts about linguistics. Chomsky has no credentials in Latin American studies, he's not a political theorist, and most definitely he is not an expert in human rights or chavista corruption. I thought back then, why would anyone listen to this imbecile when it comes to Venezuelan politics? Having studied geology, imagine my thoughts when I heard Elon Musk saying that he would come up with earthquake-proof tunnels. And here's when I think, what's AI solution to stupidity, of which there's an abundance within its very creators?

AI cannot do anything more than what its programmers want it to do. Unless they come with a completely novel pool of nonexistent data, which they won't, what will they achieve, but faster processing, facilitate hacking, and generally create more problems than what their inventions are solving? The Hugging Face episode, which allegedly was about AI agents in a sandbox finding a message board and coordinating a hack, is nothing more than PR. If the sandbox had been, as was first alleged, disconnected from the internet, i.e. air tight, no hack would have happened.

Poverty is a huge problem. Discussions about universal income occupy the minds of a great deal of people. Notice how Sam, Dario and Elon have benefited from a boom in government contracts and investments in the AI sector. Neither of their creation, but mere speculative nature of stock market participants or good old nepotism & corruption. The AI sector is like an echo chamber where dashing new comers pitch "new" solutions to leading hegemons sitting on mountains of cash, and more importantly data, and then decide to partner with the new kids on the block. This has made some of them billionaires overnight, easily, but none seems preoccupied in solving world's poverty. What takes their minds is power and wealth accumulation and how to beat competitors. Does anyone think their bots or agents will ever act against them? Will Claude agents alter Musk's heavily biased X algorithm? Will OpenAI bots escape a sandbox to gain control of Starlink and Starshield? An AI that introduces virtual checks and balances on Donald Trump would be welcome. An AI that could be used by anyone, anywhere, to independently generate passive income would be even better. But what are the chances of that happening?

Humanity is encumbered with serious problems, problems way more complex than finding a typo on PHP, plagiarising academic work or listing a bunch of destinations. Computers are amazing machines. Until truly novel stuff that no one has ever seeing before is produced, I remain to be convinced that AI is more than what it has shown to be thus far. When it comes to our systemic and chronic problems and solutions, to me at least Sam, Dario and Elon's opinions aren't any more valuable than those of the plumber.