12 Comments
User's avatar
Karo (Product with Attitude)'s avatar

Thank you for the post!

You're right, GenAI excels in isolated tasks (like image synthesis / text generation), but integrating it into end-to-end systems requires hybrid architectures. Traditional ML (e.g., decision trees, regression, or rule-based NLP) is often more stable, especially for edge cases. In other words, it's not that GenAI isn’t powerful, it’s that we’re still learning how to partner with it wisely.

Expand full comment
Geoff Livingston's avatar

Thanks! I bet a few of these AI companies can use a product developer like you!

Expand full comment
Paul Chaney's avatar

Read her Substack. They sure could!

Expand full comment
Paul Chaney's avatar

Karo, I didn't understand most of what you said, but I'm sure you're right! :-)

Expand full comment
Bette A. Ludwig, PhD 🌱's avatar

Unfortunately, I think some people see this as the ultimate solution.

It’s an amazing tool, but if you just implement parts of it without constantly

assessing, reassessing, updating, and adapting, that’s where the problem lies.

Even with writing, you have to edit, refine, and make sure it makes sense and that what you're saying is actually accurate.

Case in point: I asked ChatGPT yesterday when April Fools’ Day was,

and it confidently told me Monday… when April 1 is actually Tuesday.

Expand full comment
Geoff Livingston's avatar

Yeah, unfortunately, you are right! And of course these are the same people who accept a summarized search as verbatim truth. Or what they read on social media as verbatim truth, too!

Expand full comment
Bette A. Ludwig, PhD 🌱's avatar

Critical thinking skills have been on the decline for years now. I think we are def in the era of idiocracy Geoff.

Expand full comment
Paul Chaney's avatar

Bette, it was pulling an April Fool's joke on you! :-)

Expand full comment
Bette A. Ludwig, PhD 🌱's avatar

I was like what. I looked at my calendar and no way was April 1 on Monday. Just bizarre, right?!?

Expand full comment
Paul Chaney's avatar

Geoff, thank you for this investigative-style reporting on the state of AI. I'm sure you're right—a bubble burst is coming. Isn't that always the case? (I well recall the dot-com bubble burst in 2000.)

The Ponzi scheme analogy is so on point. It reminds me of the Wizard of Oz scene where Toto pulls back the curtain, revealing the "wizard," the snake-oil-selling con man.

But to your point, there is value in using generative AI. While I don't play in enterprise sandboxes like Salesforce, I use ChatGPT, Claude, and Perplexity daily to help me produce content and run my business.

Perhaps AI is a blessing and a curse. All I know is that I am a tech Kool-Aid drinker, and have been since blogging's early days and always will be.

Expand full comment
Geoff Livingston's avatar

Lots of bubbles between then and now, too, and it seems like several were recent, ala the Metaverse, crypto, and the Big Data boom, though in the case of the latter, I feel like AI is actually the big data thing rebranded.

It's sad that the market has left it to the buyer to figure out the solutions, just sloppy.

Expand full comment
Paul Chaney's avatar

That's why you and Greg are there -- and me to a certain extent. Problems we can help the market solve.

Expand full comment