Artificial Intelligence
Current trends and investments suggest that modern AI is frighteningly capable.
So much so it threatens to upset industries, replace jobs, and revolutionize industries.
Thousands of new startups have received billions of dollars in funding based on the groundbreaking nature of these new AI models power.
Why then is modern AI so breathtakingly stupid, incapable of even the simplest of tasks?
My Humble Opinion: Hype.
This emperor not only has no clothes — he’s also not an emperor.
He cannot stand and wave and say hello as well as a five-year-old child.
Dislike my opinion? Fine, test it yourself:
- Ask DeepAI’s image generator to draw you 18 crows or ducks.
Last I checked, in 10 trials, it failed 9 times to generate exactly 18 of anything. - Ask ChatGTP to draw you a clock showing 12:05 a.m.
It literally cannot and will not—even if you, like I did, give it corrections more than a dozen times.
AI Is Powerful—But Mostly Useless
Let’s be clear: AI models are insanely powerful.
Data science, collection, modeling, aggregation, and prediction are incredibly useful tools with real social and business value. We’ve gained significant ground in the last decade.
But in most cases?
AI is still dumber than your average second grader.
Where AI Excels (So Far)
Modern AI has shown exceptional performance in a small handful of very specific use cases:
- Health care diagnostics
- Sustainable farming techniques
- Energy management systems
In these areas, AI tools have not just matched humans—they’ve blown them out of the water.
And Yet… It Can’t Draw Ducks
It can't draw a specific number of ducks.
It can't show a specific time on a clock—unless that time happens to match what’s shown in a century of clock advertisements (like the ever-famous “10:10”).
Final Thought: Incredible, but Dumb
AI is an incredible tool.
It is also dumber than a distracted 3 year old child who's had too much sugar.
Current industry investments in "AI" startups are unsustainable, as 90% of these companies will inevitably either:
- Fail to find a business use case, or
- Worse—find a use case people will pay money for and then FAIL those uses.
