# We’re using AI all wrong Date: 2026-09-10 --- --- ### We’re using AI all wrong ![Image](https://disrael.github.io/blogs/assets/we-re-using-ai-all-wrong-69832aa976.png) So at least 75% of all AI usage is a complete waste of resources. Actually that’s being charitable because the corporate applications might be doing more harm than good for reasons well covered elsewhere — employment, slop, and expertise drain. The misconception is that AI is a tool for taking over humanity’s intellectual drudge work. The promise of AI is actually that it would do somewhat the opposite; perform tasks that even very large numbers of humans cannot. A team of humans, no matter how large or organized, runs into communication issues when they take on complex problems in biology and other fields. Similarly robotics has the promise of robots doing work too delicate or involved for humans. The problem with AI taking on drudge work is corporations become embroiled in a red queen race of software and marketing that benefits no one. Software engineering and marketing was already mostly in this state before AI and now the output of useless features and advertising campaigns will only increase. That corporations are not making people’s lives better in a few graphs: ![By RCraig09 — Own work, CC BY-SA 4.0](https://disrael.github.io/blogs/assets/we-re-using-ai-all-wrong-d4e2a314ec.png) *By RCraig09 — Own work, CC BY-SA 4.0* ![Ani Turner, George Miller, and Elise Lowry, (Commonwealth Fund, Oct. 2023).](https://disrael.github.io/blogs/assets/we-re-using-ai-all-wrong-ea9c72113b.png) *Ani Turner, George Miller, and Elise Lowry, (Commonwealth Fund, Oct. 2023).* ![Investopedia / Sabrina KarlSource: U.S. Bureau of Labor Statistics](https://disrael.github.io/blogs/assets/we-re-using-ai-all-wrong-95ed6b060b.png) *Investopedia / Sabrina KarlSource: U.S. Bureau of Labor Statistics* Okay life expectancy slightly up recently but even that one is complicated by health expectancy. ![Image](https://disrael.github.io/blogs/assets/we-re-using-ai-all-wrong-e0e7a8bc0a.png) Anyway the doldrums of wealth inequality and inability to reward a broader base of Americans has been covered in detail elsewhere. What’s important to understand is that LLMs are the wrong kind of AI for non-corporate usage. Scientific research and robotics need a different kind of machine learning that for research doesn’t specialize in language manipulation or for robotics require a very large memory footprint to run. We’ve basically invented the wheel and used it as a coffee table. However we are being teased with Navier-Stokes: > OpenAI claimed it had solved the Navier-Stokes problem. Thanks to an 88-hour assault by some 10,000 artificial intelligence agents — AI systems that can act autonomously — the company found a case in which the equations falter. One researcher estimates that represents one of the [largest amounts of computing power ever thrown at a single scientific problem](https://karthik-duraisamy.blogspot.com/2026/09/navier-stokes-regularity-what-does.html). ![Image](https://disrael.github.io/blogs/assets/we-re-using-ai-all-wrong-7d40f3c043.png) Suppose for a minute that LLMs turn out to be great at science and or robotics. Someone still has to fund those advances and we certainly can’t rely on OpenAI to do that on a regular basis. And just how many times have we seen big tech go down this road? * Computers * Internet * Search (original flavor) * Mobile / wearable * Social media * Cloud computing * Bitcoin * Virtual reality * Machine learning (original flavor) * Large language models Anyone of these technologies could have been a game changer for humanity with the right follow through, but that’s not what happened. If anyone is moving the goalposts it’s big tech. Here is an excerpt from “The Computer for the 21st Century” written by Mark Weiser (Chief Technologist at Xerox PARC) and published in Scientific American in September 1991. > Even today, people holed up in windowless offices before glowing computer screens may not see their fellows for the better part of each day. And in virtual reality, the outside world and all of its inhabitants effectively cease to exist. Ubiquitous computers, in contrast, reside in the human world and pose no barrier to personal interactions. If anything the transparent connections that they offer between different locations and times may tend to bring communities closer together. Even 35 years ago it was understood that eventually technology had to do something for humans instead of just humans doing things for technology. Big tech goalposts now move so quickly that OpenAI wrote this in 2015: > As a non-profit, our aim is to build value for everyone rather than shareholders. Researchers will be strongly encouraged to publish their work, whether as papers, blog posts, or code, and our [patents⁠](https://openai.com/approach-to-patents/) (if any) will be shared with the world. We’ll freely collaborate with others across many institutions and expect to work with companies to research and deploy new technologies. That’s just 11 years to go from super idealistic to scum of the earth — a truly groundbreaking achievement in goalpost movement. Still we can’t say that people are fools to believe in this incarnation of big tech. As Charles Schulz said in 1967: > Charlie Brown will always keep hoping.