A lot of people are championing ChatGPT as the be-all-end-all for everything. And, to be sure, it’s an amazing tool.

But it’s important to remember that it’s only a tool.

Specifically, it’s something called a large language model, which works by taking a massive amount of data – most of what’s publicly available on the web – and stringing together what it thinks is a plausible output. It doesn’t know anything. It can’t reason, or think, or empathize.

In my early dealings with AI, I have had success thinking of it like a 3-year-old child:

• It can’t keep secrets
• It has a vivid imagination
• It often comes up with things that make absolutely no sense
• It’s surprisingly clever sometimes
• It needs a lot of close supervision
• You need to check (repeatedly) to make sure it’s done what you’ve asked
• You should not entrust it with anything important or fragile

Never miss a post! Click here to subscribe to blog updates.