“At the intersection of innovation and transformation, AI isn’t just changing the game — it is the game. We’re leaning into a paradigm-shifting, synergy-driven ecosystem where machine intelligence amplifies human potential at scale. Our AI-first strategy unlocks unprecedented value across the entire customer journey, enabling data-driven decision-making that moves the needle on every key metric. This isn’t disruption for the sake of disruption — it’s purposeful, agile disruption. The organizations that fail to embrace this transformational moment will find themselves on the wrong side of history, while those who boldly invest in the AI-powered future will redefine what’s possible. We’re not just future-proofing the business — we’re building tomorrow, today.”
Yes, I had “AI” write that because my brain does not have the capacity to be that vapid and deluded. My employer has been pushing it, and has seemed upset that they’re not seeing the same gains that other companies have seen. Well, I’m sorry that there aren’t lots of good open-source projects using COBOL for companies like OpenAI and Anthropic to build their models with. Maybe if all of our code was written with JavaScript and Python like the apples you’re comparing yourself to, we would see similar gains.
That’s part of why I hate it, but let’s get into why it really sucks.
Predicting a Confirmed Solution
When I was wrapping up my undergraduate degree at the University of Nebraska in biochemistry, I had to take a capstone course as part of my coursework. My thesis for that class was using computers to determine possible antibiotic resistance. Note the usage of the word “possible” as it’s going to be a keyword here.
I wrote a program to take a genetic sequence from bacteria (which is can be represented as a string of characters) and compare that against known sequences that impart antibiotic resistance. The closer the match, the likelier a given strain of bacteria had resistance to a given class of antibiotics. Note how it didn’t confirm that there would be resistance.
The current technology of AI known as large-language models operate similarly. They ingest a crapton of data and based on a given input give the most statistically likely answer. It’s also why these models hallucinate and always will: because they operate on statistics, there’s always a chance for error.
My computer program wasn’t thinking, and neither is AI. It’s just doing some black-box educational guessing as to what someone wants based on the inputs. Yet we’re being asked to give up our cognitive processes, economic stability, and our planet for this?
Much like my computer program for my undergraduate thesis, which cost me quite a bit of cash when I ran it on AWS back then, AI has a high economic cost. This is important later, because a lot of it is being obscured by debt and venture capital.
Computers are Really Dumb and Need to Specifically be Told What to Do
There’s a reason why programming languages are very verbose with things: computers are really dumb and need to specifically be told what to do. This is good because it means that any process we automate with computers will have deterministic outcomes. This is good because lots of business operations need to have predictable, deterministic results when they execute.
In the previous section, we got into what AI currently is: a non-deterministic statistical answer generator. When I have an automated process at work run, I want solid guarantees that it will work 100%. Sometimes things do break, and the process should have predictable outputs when that’s the case. I shouldn’t have to worry about if the process wants to change the output just because.
Sure, it can write code for me but that’s not why my employer keeps me around. My employer keeps me around because I tell them what they need to hear, and to fix things when they do break. I don’t see a statistical answer generator quickly and properly fixing a business process. In fact, I have several examples of those statistical answer generators making things worse.
I really don’t feel like replacing deterministic outcomes with potentially random outcomes because I like my weekends to be drama-free, thank you.
Sam Altman is a Con Man
Every time I hear Sam Altman speak, I feel like a small part of my brain dies. This is the same dude who “can’t imagine raising a child without ChatGPT,” apparently not getting that we’ve done pretty good for millions of years beforehand. Raising children can definitely be hard, but it’s not so hard that it was impossible before the advent of his latest scam.
This is a man that created a social media company that ended up failing, but not before he cashed out and went off to Y Combinator. He was eventually kicked out of his leadership role at Y Combinator, at which point he went full throttle at OpenAI.
His style of speech reminds me of another notable Stanford dropout and convicted con person: Elizabeth Holmes of Theranos. She created a device that “does a chemistry” to blood samples to determine diagnostic criteria. That phrase would’ve had my high school chemistry teacher eviscerate me if I used it in high school, let alone in a setting for a company that was once worth almost a billion dollars.
The world I hope will eventually see him for what he is: an amoral grifter who will eventually have an Icarian fate.
Conclusion
AI is being sold in a way where consumers are being told they don’t need to worry about the fundamentals, but I only see it as a super-powerful search engine. You still need the fundamentals to judge whether the results are correct or not if it is to speed up your workflows. Much like how search engines in the 1990s sped up exploring the web and finding what you needed.
I don’t think betting our entire planet and economy on a search engine that is impervious to SEO slop is worth it. I also believe that all we’re seeing right now is techbros way too high on their own supply thinking they know more than us. It’s a consequence of them being stuck in a bubble and venture capital ensuring that bubble stayed protected so they can prey on it for their next big payday.
I can’t wait until this bubble pops. Maybe then Silicon Valley will wake up and smell the roses, or at least learn to listen to people outside San Jose. Sam Altman will likely go down as a conman greater than Bernie Madoff, Elizabeth Holmes, and Billy McFarland. I don’t envy him when this all falls. There will be some serious economic damage when this does happen, so definitely be prepared folks as the debt financing all of this will hit us all.
Leave a Reply