Image Source: Generated in Bing’s Image Creator AI, then modified with Sketch.io. Based on Craig Alanson’s Expeditionary Force series featuring the best AI of all, Skippy “Trust the Awesomeness.”

Is AI dumb? If you read Seth Godin, "ChatGPT is dumber than it looks,” maybe it is. Of course, he explains why that is so:

The reason AI language models are dumb is that they don’t actually know anything, the model is simply calculating probabilities. Not about the unknown, but about everything. Each word, each sentence, is a statistical guess.

He then goes on to point out that he uses Claude.ai with specific instructions as to how to best do so. He’s not alone in doing so.

AI Can Boost Productivity 20-40%

The authors of The Neuron newsletter write:

Despite the hyperbolic forecasts we often hear from tech CEOs about AI during earnings calls, here’s a reality check: AI in 2024 isn’t all that impressive. Yeah, we said it, don’t @ us. That’s not to say it’s worthless—if leveraged properly, today’s AI can boost productivity by 20-40% in areas like coding, writing, and HR.

Really? Dumb?

You know when your organization or team says, “Oh, they don’t know anything or they’re not really all that good,” about a competitor? It makes me wonder, “Am I suffering from superiority bias?”

I suspect the real reason we’re saying AI is so dumb is that we’re afraid of what that really means for us. If AI is smart, intelligent, or at least, good at mixing words together in ways that make sense to our monkey brains (you know, break it down Barney style), then it’s smart enough to not be called dumb.

Having relied on AI for data analysis, I’m inclined to think it’s pretty smart. Whether I’m using Perplexity or Claude, I can see that it’s able to do things that would take me hours to do myself. What’s even more incredible, ensuring the right prompt with sufficient context, interactive dialogue and explanation can be quite helpful to the AI. It’s the kind of clarifying conversation you would have with a human being.

What am I arguing for here? Nothing really. Only that we should stop calling AI dumb. Instead, let’s be realistic about its capabilities without ascribing godlike intelligence to language learning models running on huge server farms.