François Chollet on Deep Learning and the Meaning of Intelligence

**Sean Carroll** (0:00)
We are all driven by the search for better, but when it comes to hiring, the best way to search for a candidate isn't to search at all. Don't search match with Indeed. If you need to hire, you need Indeed. Indeed is your matching and hiring platform with over 350 million global monthly visitors and a matching engine that helps you find quality candidates fast. You can ditch the busy work. Use Indeed for scheduling, screening, and messaging so you can connect with candidates faster. And Indeed doesn't just help you hire faster.
93% of employers agree that Indeed delivers the highest quality matches compared to other job sites. And Mindscape listeners will get a $75 sponsored job credit to get their jobs more visibility at indeed.com/mindscape. Just go to indeed.com/mindscape right now and support the show by saying you heard about Indeed on this podcast. indeed.com/mindscape, terms and conditions apply. Need to hire? You need Indeed.

**SPEAKER_2** (0:59)
This summer, let your imagination soar by listening on Audible. Whether you listen to stories, motivation, expert advice, any genre you love, you can be inspired to imagine new worlds, new possibilities, new ways of thinking. With Audible, there's more to imagine when you listen.
And speaking of listening, you can listen to the best-selling fantasy romance, A Court of Thorns and Roses by Sarah J. Maas.
Right now, on the Audible app, transport yourself to a realm of magic and purses, all from the comfort of your living room. As an Audible member, you choose one title a month to keep from their entire catalog. New members can try Audible free for 30 days. Visit audible.com/wonderepod or text wonderepod to 500-500. That's audible.com/wonderepod or text wonderepod to 500-500.

**SPEAKER_3** (1:52)
Get in the car, kids. We're going to T-Mobile.

**SPEAKER_4** (1:55)
I thought we were going to Grandma's.

**SPEAKER_3** (1:56)
Nope, we're switching to T-Mobile. We're getting four 5G smartphones, and T-Mobile is covering the cost.

**SPEAKER_4** (2:02)
Can I get one with a bigger screen for video games?

**SPEAKER_2** (2:05)
Can I get a pink one?

**SPEAKER_5** (2:06)
Can we get the deal where we get four lines for 25 bucks per line a month?

**SPEAKER_4** (2:10)
Wait, does that mean I get a bigger allowance?

**SPEAKER_3** (2:14)
Whoa, that's a lot of questions, but yes, yes, absolutely, and probably not.

**SPEAKER_6** (2:23)
Get 4 5G phones on us, and 4 lines for $25 a line per month when you switch with eligible trade-ins at T-Mobile, all on America's largest 5G network.

**SPEAKER_7** (2:33)
Minimum 4 lines for $25 each per month with auto pay using debit or bank account, $5 more per line without auto pay. Up to $830 off each phone via 24 monthly bill credits plus taxes and fees for well-qualified customers. Contact us before canceling account to continue bill credits or credit stop and balance on a required finance agreement due. $35 per line connection charge applies. ctmobile.com.

**Sean Carroll** (2:52)
Hello, everyone, and welcome to the Mindscape Podcast. I'm your host, Sean Carroll.
You know that artificial intelligence is in the news. We've talked about AI in various different ways here on the podcast, especially over the last couple years where ChatGPT and other large language models have really become an enormous study of interest to many people for financial reasons, for intellectual reasons. They're becoming everywhere, right? Google has put them on the first page of its search results. Lots of people are using large language models to write texts.
You can write programs using large language models. You can write the syllabus for your college course, et cetera. It's clear that this technology is going to have an enormous impact on how humans behave and live going forward. But there are subtleties. One of the things that I've talked about is the idea that large language models are amazing because they are able to mimic human speech and behavior, right? They are able to sound enormously human without actually thinking in the same way that human beings do. Large language models, in some sense, memorize lots of things. They know a lot of facts about the world, and they're super good at interpolating between things that they know.
That includes interpolating different kinds of things that have never been interpolated before, so they can seem creative. They can do things that have never been done based on the training data of things that have been done before.
They're less good at going outside of the range of that training data, and one can argue that the processes by which they come up with their outputs are very different than what a human being does in actually thinking, reasoning about the problem presented to it.
And many people, especially people who are experts in AI, understand this attitude perfectly well. It's certainly not new with me. It's well known to many people, but it is denied by other people who are much more impressed with the progress in large language models and think that we're close to AGI, artificial general intelligence. So I thought it would be fun to talk to someone who is in the front lines of developing deep learning models and AI more generally. So today's guest is François Chollet. He's a relatively young guy, but just to give you a sense of his accomplishments, he's a deep learning researcher at Google.