Ethan Mollick

speaker

673 appearances 1 recordings 1 series first heard Jul 2026 last heard 9 Jul

Ethan Mollick’s voice in public audio — every appearance, attributed to the second.

Trend

recordings per month · last 12 months

Recordings per month over the last 12 months — 1 in all, peaking in Jul 2026 with 1.

Appearances

newest first · ▶ plays the moment

Strange Loop Podcast · Ethan Mollick | The long way is the shortcut: experiments over AI playbooks · 9 Jul 2026 podcast

And in fact, we might be in a temporary place where interpretability is higher because the chain of thought that AIs use is in plain English or, you know,

depending on which model you use.

And so there's at least some ability to get insight into that.

There's no reason that has to stay in a human interpretable form.

And so I think that there's an inexplicableness to the whole process that is going to be very hard to eliminate.

It doesn't seem to be an area where progress is particularly fast.

Yeah, so there's really kind of two branches here.

One of them is the big picture, right?

Recursive self-improvement, the idea AIs make AIs better, and we enter a weirder world very quickly.

This take-off concept is often called, right?

So once AIs start improving themselves, is there some sort of ever-rapidly increasing sort of exponential curve that we can't understand?

And I think that we don't really know.

The co-founder of Anthropic was just a couple days ago, at least the time we're recording this, posted that he thought that there was a 40% to 60% chance that we'd have AI scientists doing AI work and improving AI by 2028, I think the number was.

This is on everybody's mind, and I don't think we know what that world looks like.

It can get very weird very quickly.

As long as AI stays jagged, it may not be quite as tremendous as people think, because it's still weak in some areas and better in others.

But then there's the personal aspect of working with these systems.

And I think, to some extent, you have to commit the cardinal sin of AI right away, which is anthropomorphization, which is the idea that treating the AI like a person.

I mean, first of all, these systems want to act like people.

That's what they do.

Showing 121–140 of 673 · page 7 of 34