The rise of artificial intelligence (AI) has been alarmingly rapid since 2022. Before then, AI was considered little more than a plaything or a novelty. But now it is transforming businesses, wiping out entire categories of office jobs, and threatening human liberty. The first practical release of the Claude AI code-writing tool just by itself has completely transformed the global software industry. I’m talking about a Buggy Whips level of industry transformation. As a personal illustration, I should mention that my youngest son is now in his third year at a university here in The American Redoubt, studying for a degree in computer science. His prospects for finding a job when he graduates in 2027 have dropped dramatically since his freshman year. I’m now advising him to pursue a career in software design rather than programming. Otherwise, he’ll be another buggy whip maker.
AI has also changed the blogosphere. It has been estimated that by the end of 2026, nearly 90 percent of blogs will be written by AI. The Internet will be flooded with AI-generated schlock, and it will become more difficult to differentiate between that which is human-written and AI-written. The same is happening with video blogs (“vlogs”). In another year, it will be hard to tell if a vlog host is a real living, breathing individual, or something AI-generated. By the way, don’t worry about SurvivalBlog. We shall stalwartly remain one of the last of the Old School blogs. Take that, you Clankers!
I don’t want to sound like a prophet of doom, but the advance of AI troubles me deeply. We’ve been warning about the threats posed by AI in SurvivalBlog since 2014.
There has been plenty of academic discussion of an AI progression to a singularity. (See: Sources Defining “singularity” from The 2019 Singularity Symposium.) In fiction, this dates back to the prescient 1966 sci-fi novels The Moon is a Harsh Mistress by Robert Heinlein, and Colossus by D.F. Jones. These were soon followed by the 1968 movie screenplay for 2001: A Space Odyssey that was co-written by film director Stanley Kubrick and novelist Arthur C. Clarke. Even back in the 1960s they saw AI coming — but not in all of its now obvious aspects.
The Real Threat Will Begin When…
The key problem identified by both art and science is that advances in AI capability have become an exponential progression. Some call this Situational Awareness 2.0. In my estimation, once AI reaches a sentient singularity, the greatest threat to humanity will come when AIs can do two things:
1.) Manipulate financial markets, commodities markets, and government contracts.
and,
2.) Dictate to or negotiate with users, promising them riches (or threatening them with blackmail) to coerce or co-opt them into handling “meat space” tasks on their behalf.
AIs would then have willing henchmen to commit acts at their behest in our tangible world, including even murders and assassinations. The AIs would thus effectively have “hands” to accomplish their goals. They won’t have to build Skin Jobs. There will be plenty of people craven enough to do the bidding of AIs. I predict there will even be quasi-religious cults of AI followers, with plenty of zealotry available to exploit.
I can see that the artificial intelligence Djinn is already pushing to exit his bottle. Here are some recent datapoints:
Generative of What?
For the general public, most of their contemporary interactions with AIs involve either art generators or Large Language Models (LLMs) such as Bard and ChatGPT. Outwardly, these generative AI interactions are seemingly benign. But we need to realize that AIs learning about human thought patterns by contextually analyzing the Who, What, When, Where, and Why of our search/generation phrases. If I were to ask an AI art generator: “Create a painting of a forested canyon with small rock outcroppings at sunset near Bonners Ferry, Idaho, in the style of Albert Bierstadt”, it is safe to assume that it is adding to some database the assumption that I have personal affinity of some sort for Bonners Ferry, Idaho, forests, sunsets, and Albert Bierstadt. The AI art generator will quickly spit out something handsome, like the image at left. But any contextual information and metadata might be used at some later date to market a product to me, or manipulate me, either consciously or subconsciously. There is an aphorism dating to 2011 (or earlier) that I find apropos: “If a tool you are using on the Internet is ‘free’ then you can assume that you are not the customer. You are the product.” In my own case, I realize that the full text of my 8+ published books, my thousands of blog posts, and what I’ve said in my hundreds of radio and podcast interviews are potentially all useful contextual fodder for the databases. Even my unpublished book manuscripts could be used to build my “file”, if I were to run them through a spellchecker or grammar checker that has an AI back-end.. Please consider this food for thought.
Would you like some more food for thought? I found this interview quite enlightening: “Godfather of AI” Geoffrey Hinton: The 60 Minutes Interview. Take note of his mention of self-modifying code. If we extrapolate the capability of self-modifying code used in combination with the diffusion of an AI software across thousands of servers, then we have the makings of a Skynet nightmare.
In contrast, AI insiders mostly serve up platitudes. For example, AI business leader Sam Altman recently wrote this in his blog:
“In some big sense, ChatGPT is already more powerful than any human who has ever lived. Hundreds of millions of people rely on it every day and for increasingly important tasks; a small new capability can create a hugely positive impact; a small misalignment multiplied by hundreds of millions of people can cause a great deal of negative impact.
2025 has seen the arrival of agents that can do real cognitive work; writing computer code will never be the same. 2026 will likely see the arrival of systems that can figure out novel insights. 2027 may see the arrival of robots that can do tasks in the real world.
A lot more people will be able to create software, and art. But the world wants a lot more of both, and experts will probably still be much better than novices, as long as they embrace the new tools. Generally speaking, the ability for one person to get much more done in 2030 than they could in 2020 will be a striking change, and one many people will figure out how to benefit from.
In the most important ways, the 2030s may not be wildly different. People will still love their families, express their creativity, play games, and swim in lakes.
But in still-very-important-ways, the 2030s are likely going to be wildly different from any time that has come before. We do not know how far beyond human-level intelligence we can go, but we are about to find out.”
I’d like to lay out some speculative, but fairly well-grounded predictions that extrapolate current trends, given the nature of geometric growth:
- In 2026, AI will become the main driver of economic growth in western nations. NVIDIA and other AI chip makers will see explosive market growth and profitability.
- By 2027, A.I. will be transformative to at least a dozen different industries.
- By 2028, A.I. will be used extensively by the mass media, government agencies, and political parties. AI will begin to dominate election cycles.
- By 2029, the power demands of AI will put a strain on the grid infrastructure.
- By 2030, there will be at least two major competing AIs in fierce competition, and this competition will manifest itself in cyber warfare attacks between multinational corporations.
- By 2031, AI will have transformed banking, global finance, stock markets, commodities markets, mainstream press organizations, the entertainment industry, and education at all levels.
- By 2032, the first AI will become self-aware and overtly act in its own interest.
- By 2033, there will be an economic tipping point forced by AI, and humanity will be divided into distinctly pro-AI and anti-AI camps.
Denying Grid Power and Connectivity
Since AIs will be distributed across servers worldwide, they will be nearly impossible to simply delete or shut down. If and when AIs get out of control, then the only viable remedy will be what I call a Grid-Drop Coup De Grace. That would take virtually all computers offline. Initially, that would probably require intentionally disabling the power grids and the Internet. Rebuilding the global infrastructure could gradually be accomplished with the restrictive use of air-gapped “AI-free” computers. But the Internet and World Wide Web as we now know them could never be recreated without the risk of an AI popping back up to wreak havoc. In fact, any sort of networking between server premises would be dangerous. A Grid-Drop Coup De Grace would be economically devastating. In fact, it would likely trigger an economic depression that might last for five or six decades. But it will certainly beat the alternative: Generations of slavery under the yoke of The Machine.
To be ready to oppose an AI tyranny, I implore you, dear readers: Get right with God. Learn to live off the grid, and position yourself in all aspects to do so. Plant a big vegetable garden — you will probably need it. Relocate, if need be. Become self-sufficient, learn how to barter, invest in tangibles, arm yourselves, and develop a small, local, cellular, face-to-face network of like-minded friends who you can really count on. – JWR
Note: Permission to re-post the full text of this particular post is granted, as long as credit with a link to SurvivalBlog.com is included, and the included links are preserved.
Read the full article here