The Wonder-Driven Builder — with Paige Bailey

WEBVTT
00:00:02.799 --> 00:00:04.160
Welcome to the Test Set.
00:00:04.160 --> 00:00:10.720
Here we talk with some of the brightest thinkers and tinkerers in statistical analysis, scientific computing, and machine learning.
00:00:10.720 --> 00:00:16.524
Digging into what makes them tick, plus the insights, experiments, and OMG moments that shape the field.
00:00:16.524 --> 00:00:21.565
On this episode, we sit down with Paige Bailey, DevRel Engineering Lead at Google DeepMind.
00:00:21.565 --> 00:00:27.730
She started out coding text based games from Byte Magazine onto her Apple II computer, which is a real throwback.
00:00:27.730 --> 00:00:40.170
Her website mentioned she makes weird things digitally and IRL, like a bot that emails her, roasting her taste in music, or solar powered silent film birdhouses as street art in San Francisco.
00:00:40.524 --> 00:00:46.125
She runs us through her background in geophysics, where she was introduced to Python and open source early.
00:00:46.125 --> 00:00:53.399
She discusses how she weaves a variety of Google APIs and services together to make things fun, weird, and educational.
00:00:53.399 --> 00:01:00.000
And I left feeling like my use of AI could be so much weirder, so I'm so excited for people to listen to this interview.
00:01:02.039 --> 00:01:04.359
Paige Bailey, welcome to The Test Set.
00:01:04.359 --> 00:01:05.959
Awesome. Thank you for having me.
00:01:05.959 --> 00:01:06.314
Yeah.
00:01:06.314 --> 00:01:11.594
So we we're here in San Francisco, and you you mentioned you took a Waymo in, which seems very I did.
00:01:11.594 --> 00:01:13.034
Is very on brand. Yes.
00:01:13.034 --> 00:01:14.394
I love that.
00:01:14.394 --> 00:01:20.875
Just by way of introduction, so you're a developer relations engineering lead at Google DeepMind?
00:01:20.875 --> 00:01:21.299
Correct.
00:01:21.299 --> 00:01:27.459
And you've had a incredible career of doing a lot of things, which I really wanna get to.
00:01:27.459 --> 00:01:32.700
Because I always forget, I wanna make sure to say I'm Michael Chow, the host.
00:01:33.459 --> 00:01:42.165
And I'm joined by Hadley Wickham, Chief Scientist at Posit, and Isabel Zimmerman, incredible Posit software engineer.
00:01:42.924 --> 00:01:45.564
And we're so happy to have you on Excellent.
00:01:45.564 --> 00:01:49.045
Here in San Francisco. Here in beautiful San Francisco.
00:01:50.159 --> 00:02:01.920
I know we we talked a little bit before you came on, and I thought maybe you could just catch us up on your career starting with your love of text adventures.
00:02:01.920 --> 00:02:03.114
Go there and.
00:02:03.114 --> 00:02:03.995
Absolutely.
00:02:03.995 --> 00:02:07.834
So so I got my first computer.
00:02:07.834 --> 00:02:12.395
We rescued it from being thrown away around the time that I was eight or nine years old.
00:02:12.395 --> 00:02:13.835
It was an Apple II.
00:02:13.835 --> 00:02:23.060
And the there weren't a ton of games that were available for the Apple II at the time, but certainly not ones that you could like go to GameStop and purchase.
00:02:23.060 --> 00:02:28.979
So I was beholden to all of the games that came with this computer that just missed getting thrown away.
00:02:28.979 --> 00:02:33.044
And also a whole bunch of books that talked about how you might build your own software programs.
00:02:33.044 --> 00:02:34.245
Oh.
00:02:34.245 --> 00:02:41.044
And old magazines like Copies of Byte, which had additional like games that you could type in.
00:02:41.044 --> 00:02:48.444
So my first programs were all text adventures, kind of like either transcribing choose your own adventure books or like building my own.
00:02:49.009 --> 00:02:54.530
And that just sort of made computers feel like a fun thing.
00:02:54.530 --> 00:03:04.014
And loved open source as well, like, can pretty much give my entire career to open source.
00:03:04.014 --> 00:03:04.414
Yeah.
00:03:04.414 --> 00:03:04.814
Yeah.
00:03:04.814 --> 00:03:08.895
Yeah. Do you remember that era of computing? I do. Okay.
00:03:08.895 --> 00:03:09.854
I do remember it.
00:03:09.854 --> 00:03:15.055
And we we had like a printer where the paper, you know, is one big Yeah.
00:03:15.055 --> 00:03:17.055
Yeah. Yeah. With the perforated edges.
00:03:17.055 --> 00:03:17.455
Yeah. Yeah.
00:03:17.455 --> 00:03:19.080
And you have to rip it off.
00:03:19.080 --> 00:03:20.120
Yeah. Yeah.
00:03:20.120 --> 00:03:22.759
It's rough, you know, as a child, I feel like.
00:03:22.759 --> 00:03:25.400
And it had such a cool a cool tone too.
00:03:25.400 --> 00:03:28.199
Like that Yeah. The printers.
00:03:28.199 --> 00:03:31.235
When it was working. Yeah. It was like humming a tune. Yeah.
00:03:31.235 --> 00:03:31.715
Yeah.
00:03:31.715 --> 00:03:32.194
Exactly.
00:03:32.194 --> 00:03:35.794
Isabel assuming you ever got a magazine where you like typed in?
00:03:35.794 --> 00:03:36.435
I did not.
00:03:36.435 --> 00:03:42.754
I think my first intro to computer games was like Bugdom. Anything from Pangea Software, I I spent time in.
00:03:42.754 --> 00:03:43.740
As well as Myst.
00:03:43.740 --> 00:03:45.099
Myst. Myst was love Myst.
00:03:45.099 --> 00:03:46.139
There was a Yeah.
00:03:46.139 --> 00:03:51.259
Like, number one game, like, my brother and I would sit around for hours and go through all the puzzles.
00:03:51.259 --> 00:03:52.139
It was amazing.
00:03:52.139 --> 00:03:56.460
Myst, Riven, Exile, like, definitely, definitely all known and loved.
00:03:56.460 --> 00:03:56.780
Yeah.
00:03:56.780 --> 00:03:59.594
All your games came on CDs, then?
00:03:59.594 --> 00:04:00.395
Or Yeah.
00:04:00.395 --> 00:04:01.835
Yeah. All CDs. Yep.
00:04:01.835 --> 00:04:02.314
All CDs.
00:04:02.314 --> 00:04:08.235
And Riven, I remember Riven in Exile had like four CDs that you had to like put in in order Yeah.
00:04:08.235 --> 00:04:09.835
As you explored the islands.
00:04:09.835 --> 00:04:10.555
Yeah. For real.
00:04:10.555 --> 00:04:15.270
I I'm ashamed to admit I never got past what seems like the very beginning of Myst.
00:04:15.270 --> 00:04:18.310
I just traveled around pulling levers and Yeah.
00:04:18.310 --> 00:04:21.350
I never knew where the levers went to or what was happening.
00:04:21.350 --> 00:04:30.764
So I feel like you two really tapped into some other dimension of Myst that I totally pages and, like, learning this learning the lore of Cirrus and Achenar.
00:04:30.764 --> 00:04:33.805
It like I feel like it was my first intro to plot twists.
00:04:33.805 --> 00:04:35.884
And I was like, this is mind blowing.
00:04:35.884 --> 00:04:40.649
Like, I I didn't realize you could experience a game like this and have such an emotional reaction to all of it. Reaction to all of it.
00:04:40.649 --> 00:04:41.209
Exactly.
00:04:41.209 --> 00:04:44.730
And that you could kind of choose your own adventure when it came to the ending too.
00:04:44.730 --> 00:04:49.529
So you could like pick one of the brothers or you could pick you could pick Atris.
00:04:49.529 --> 00:04:51.730
I think that was his name or something.
00:04:52.185 --> 00:04:55.625
But it was but it was awesome.
00:04:52.185 --> 00:04:55.625
Like, strong recommend.
00:04:55.625 --> 00:04:56.584
It holds up.
00:04:56.584 --> 00:04:57.384
I have to go back.
00:04:57.384 --> 00:04:58.824
Yeah. I think I'm ready.
00:04:58.824 --> 00:05:03.899
I think I've like matured enough that I can I'm ready to tackle Myst in the way it was meant to be played.
00:05:03.899 --> 00:05:04.220
Cool.
00:05:04.220 --> 00:05:04.459
Cool.
00:05:04.459 --> 00:05:07.019
Well, now if you get stuck, you can go to go on the Internet.
00:05:07.019 --> 00:05:07.500
And yeah?
00:05:07.500 --> 00:05:10.699
Or just get Claude Code to play the game for you.
00:05:10.699 --> 00:05:11.339
Yeah.
00:05:11.339 --> 00:05:13.420
Yeah. I'm really curious.
00:05:13.420 --> 00:05:16.975
When did you how did you get involved with computer programming?
00:05:16.975 --> 00:05:17.295
Yeah.
00:05:17.295 --> 00:05:21.375
Well, so so creating text adventures, that was kind of the first thing.
00:05:21.375 --> 00:05:41.410
And then also, when I went to I went to Rice University in Houston, and there was a class called CAAM 210, which a friend was taking taught by kind of this majestic bearded human named Steve ***.
00:05:41.410 --> 00:05:49.425
I'm not sure if you encountered him Hadley, but he was like wonderful wonderful human, Usually taught cam two ten using MATLAB.
00:05:49.425 --> 00:05:53.504
But for whatever reason, like this semester was teaching it using Python.
00:05:53.504 --> 00:06:01.110
Like Python plus a lot of the numeric and scientific computing packages that were kind of earlier days at the time.
00:06:01.110 --> 00:06:18.095
And I remember like hanging out with my friend who was going through a lot of the a lot of the tutorials and like trying to bang their head against the wall to to start using Python and got really enchanted by this idea that people would build things and then like release them out into the world for zero Oh, cool.
00:06:18.095 --> 00:06:18.574
Yeah.
00:06:18.574 --> 00:06:27.495
So so that was I would say like started programming, just text adventures, making mods to like Grand Theft Auto.
00:06:27.535 --> 00:06:36.670
Not really like thinking anything of it when I was younger, and then I got to school and like Python was kind of the first thing that that helped me understand open source.
00:06:36.670 --> 00:06:37.629
Yeah. Cool.
00:06:37.629 --> 00:06:42.470
And were you would you say you're doing a lot of sort of data science work out of school or what kinda?
00:06:42.644 --> 00:06:45.925
So so it's really it's really fascinating.
00:06:45.925 --> 00:06:54.589
Like, I feel like there was this period of time where a lot of people were doing data science y style things, but the word did like, the term did not exist.
00:06:54.589 --> 00:07:04.269
So so for open source, like, my background, I was doing kind of geophysics, specifically planetary sciences research.
00:07:04.269 --> 00:07:15.144
So I interned at the Laboratory for Atmospheric and Space Physics and Southwest Research Institute Antonio and was doing a lot of work to understand lunar ultraviolet.
00:07:15.144 --> 00:07:28.639
So there were these spacecraft that had a whole bunch of data that would get pulled down to help understand lunar albedo or some of the other some of the other things, like metrics that the spacecraft were capturing.
00:07:29.399 --> 00:07:35.800
And the expectation was that you would analyze them using something called IDL, which was more painful than MATLAB.
00:07:35.800 --> 00:07:41.915
And but I had I had like, Python existed. And so that was that was a little bit easier.
00:07:41.915 --> 00:07:54.194
And so so I ended up if I needed to visualize data or or kind of understand it instead of like trying to bang my head against the wall and understand IDL, apologies to anybody who might have been working on IDL.
00:07:54.399 --> 00:07:58.079
I I just used Python and started working on it like that.
00:07:58.079 --> 00:07:59.279
And we yep.
00:07:59.279 --> 00:08:09.884
And we were also really fortunate in Texas in the sense that, like, just down the way in Austin, there were a whole bunch of people that were also, like, nerds into open source libraries for scientific computing.
00:08:09.884 --> 00:08:21.925
So like the Enthought Canopy dudes and, you know, the Anaconda team now, though they had previously had another name, I think Continuum Analytics or something similar.
00:08:22.285 --> 00:08:26.019
So so feel like I've really looked into it, honestly.
00:08:26.019 --> 00:08:36.485
It's kinda hard to imagine now, but at the time, it was such a crazy idea to like that you might use, like, free software to do serious work instead of paying, like, money.
00:08:36.485 --> 00:08:36.965
Exactly.
00:08:36.965 --> 00:08:37.845
Yeah.
00:08:37.845 --> 00:08:40.644
Like, now it just seems like, it is kind of, like, weird.
00:08:40.644 --> 00:08:40.965
Right?
00:08:40.965 --> 00:08:46.245
We're just using all this free stuff that people make from out of the goodness of their hearts and give to the world.
00:08:46.245 --> 00:08:47.044
One hundred percent.
00:08:47.044 --> 00:08:55.860
Especially in the oil and gas like, the geophysics oil and gas industry world because so much of that software is proprietary, like even today.
00:08:55.980 --> 00:09:04.705
Like there there's something called Petrel, the the one of the software applications for understanding subsurface data.
00:09:04.705 --> 00:09:08.705
And I think it's like a hundred thousand dollars a seat per year or something.
00:09:08.705 --> 00:09:11.664
So like like crazy crazy amounts.
00:09:11.664 --> 00:09:17.610
But most of the things that you do on Petrel, if not all, you could easily do with just like Python and open source libraries.
00:09:17.610 --> 00:09:18.649
Oh, interesting.
00:09:18.649 --> 00:09:19.129
Yeah.
00:09:19.129 --> 00:09:27.850
So so it does feel like a really it feels so strange to even think about it right now of like how different everything was back then.
00:09:27.850 --> 00:09:35.294
Like, it was also it was also the situation where I remember going to my professors and being like, hey, really like, you know, like programming.
00:09:35.294 --> 00:09:36.414
I think this is really cool.
00:09:36.414 --> 00:09:40.894
Like, we should totally have libraries for doing earth sciences related things.
00:09:40.894 --> 00:09:43.615
And my professors being like, no, that's not real science.
00:09:43.615 --> 00:09:45.134
Like, why are you wasting your time?
00:09:45.134 --> 00:09:51.389
Like, nobody needs like, if you're if you're writing libraries, like, that's not actually contributing to your career.
00:09:51.389 --> 00:09:52.590
Actually a waste of time.
00:09:52.590 --> 00:09:53.230
Yeah. Yeah.
00:09:53.230 --> 00:09:56.509
But like tooling versus doing Yeah. The actual research.
00:09:56.509 --> 00:09:56.990
Yep.
00:09:56.990 --> 00:09:57.789
Yep.
00:09:57.789 --> 00:10:05.544
Now, like, now it's if you're not, you know, writing code or like using using code to analyze data, you're probably not doing this.
00:10:05.544 --> 00:10:07.465
What are you doing? Yeah.
00:10:07.465 --> 00:10:08.585
I'm really curious.
00:10:08.585 --> 00:10:13.664
You were somebody who has been in the AI space kinda like before it was cool, and I'm putting that in air quotes.
00:10:14.210 --> 00:10:20.850
Like, how did you make that jump from, you know, I'm doing, like, geo stuff to now I'm I'm doing AI.
00:10:20.850 --> 00:10:23.409
I'm doing deep learning, working on things like TensorFlow?
00:10:23.409 --> 00:10:24.049
Yeah.
00:10:24.049 --> 00:10:38.095
So so geophysics geophysics is like I always feel like we are the the the sort of the the more stay at home type kind of earth scientists.
00:10:38.095 --> 00:10:38.335
Right?
00:10:38.335 --> 00:10:45.710
Like, we're the we're the ones that care about computers and that are that are trying to find more efficient ways to analyze data.
00:10:45.710 --> 00:10:57.764
And the geophysic world actually adopted GPUs before most everybody else for things like velocity modeling, like fluid dynamic simulations, like all sorts stuff.
00:10:57.764 --> 00:11:01.965
Even before the gaming industry, like, kind of like caught on to GPUs.
00:11:02.245 --> 00:11:11.485
And so so I was doing that a lot for some of the the geophysics work just like trying to bang my head against CUDA.
00:11:11.889 --> 00:11:17.809
And when TensorFlow was first released, don't I don't think most people remember, but it was CPU only.
00:11:17.809 --> 00:11:24.090
So you're doing kind of like, you're doing analysis across like large numbers of CPUs.
00:11:25.254 --> 00:11:28.774
Not really thinking about GPUs until later on.
00:11:28.774 --> 00:11:35.414
And so that was something that I could do that was helpful of just kind of like, hey, you know, like, you're thinking about accelerators now.
00:11:35.414 --> 00:11:38.095
I know CUDA. Like, I can help with that.
00:11:38.440 --> 00:11:44.919
And it's a it's a useful way to to kind of collaborate with the team, I think.
00:11:44.919 --> 00:11:55.195
Like if somebody releases an open source library and you contribute to it or you like build something with it and talk about it, then they get a flavor of what it would be like to work with you as a person.
00:11:55.195 --> 00:12:01.674
There's such a strong pipeline of people from like physical sciences like that to it feels like AI.
00:12:01.674 --> 00:12:05.940
There's the whole conference about ex astronomers, now data scientists.
00:12:05.940 --> 00:12:09.460
It's so surprising to see how all this all this comes together.
00:12:09.460 --> 00:12:09.779
Yep.
00:12:09.779 --> 00:12:16.940
Linear algebra and like also, you know, has the same JAX has the same spec as NumPy.
00:12:17.625 --> 00:12:23.705
So and a lot of the libraries have the same, like, general flavor of scientific computing libraries.
00:12:23.705 --> 00:12:28.865
So like SciPy for scientific computing processes.
00:12:29.304 --> 00:12:39.149
So so I think there's there's definitely a lot of love between the the kind of physics and mathematics world and the JAX world.
00:12:39.149 --> 00:12:49.065
Also, like all of the JAX machine learning frameworks authors came from mostly mathematics backgrounds, which is quite cool.
00:12:49.065 --> 00:12:50.985
Yeah. It's interesting.
00:12:50.985 --> 00:12:58.750
It seems like you've done a lot of both development and product management and DevRel engineering.
00:12:58.750 --> 00:12:59.389
Yep.
00:12:59.389 --> 00:13:03.629
When did you know this was kind of the work that you really wanted to be doing?
00:13:03.629 --> 00:13:06.110
How did this work come together?
00:13:06.110 --> 00:13:06.669
Yeah.
00:13:06.669 --> 00:13:11.644
I I think I I again, just like, I feel so stupidly lucky.
00:13:11.644 --> 00:13:18.325
Like the in the sense that I've I've just always done whatever I thought would be useful and interesting.
00:13:18.524 --> 00:13:37.600
And now we're in this really happy place where a long time ago, people would kind of artificially put people into buckets of like, oh, you're an engineer, you're a product human, you're, you know, a design human, you're like these these other things and you could only wear this one hat, which is really frustrating if you like doing a whole bunch of stuff.
00:13:37.774 --> 00:13:53.429
And now we're in this place where it's just like, oh, you want to you want to do data visualization and you want to do like, you want want to understand large scale data sets and you also want to be able to to build tools in order to do all of the above.
00:13:53.429 --> 00:13:54.309
Like go for it.
00:13:54.309 --> 00:14:00.350
Like that's you can you can be this auteur of anything you want to you want to enable.
00:14:00.470 --> 00:14:10.524
So so I I didn't have like any sort of design other than like I'm gonna work on things that I feel will be important and useful and interesting.
00:14:10.644 --> 00:14:17.764
And when I was a product person, I was exclusively a product person on developer tools or models.
00:14:17.764 --> 00:14:23.620
So like highly technical, like lower level sort of sort of things.
00:14:23.620 --> 00:14:25.620
So I was still writing code every day.
00:14:25.620 --> 00:14:26.019
Yeah.
00:14:26.019 --> 00:14:26.419
Yeah.
00:14:26.419 --> 00:14:33.860
It does it does really feel like like Claude Code and similar tools are, like, such a win for people who have just got, like, weird interest in lots of places.
00:14:33.860 --> 00:14:34.740
Exactly.
00:14:34.740 --> 00:14:42.855
Like, it's and and, like, suddenly all of these things that were, like, seventy percent finished on your GitHub, like, you can, like, actually get out into the world.
00:14:42.855 --> 00:14:47.909
It's there's never been a better time to be a curious human who wants to, like, create things.
00:14:47.909 --> 00:14:50.230
Yeah. Yeah. Yeah. Like, I made an iPhone app.
00:14:50.230 --> 00:14:54.389
Like, it's just a little timer for talks, but just, something that would be in my head Yep.
00:14:54.389 --> 00:14:57.110
For years, and, that's like a weekend project now.
00:14:57.110 --> 00:14:58.230
It's, like, it's really cool.
00:14:58.230 --> 00:14:58.870
Exactly.
00:14:58.870 --> 00:15:08.215
I I don't have it in my backpack right now, but there's one of my side projects is using there's something called a cheap yellow display.
00:15:08.215 --> 00:15:11.815
It's called the CYD ESP thirty two.
00:15:11.815 --> 00:15:19.110
And it's like this little tiny board with just a visualization screen on top.
00:15:19.110 --> 00:15:22.149
And to update it, you have to rewrite the firmware.
00:15:22.149 --> 00:15:30.674
So so it's it's not as flexible as something like a Raspberry Pi with a screen or but I would have never been able to to do anything with it.
00:15:30.674 --> 00:15:36.434
And now I'm using those to power these silent film birdhouses.
00:15:36.434 --> 00:15:42.840
So it's like you have a birdhouse, put the little screen on the inside, people look through the little hole, and you can watch the silent films.
00:15:42.840 --> 00:15:43.720
Oh my gosh. Yeah.
00:15:43.720 --> 00:15:45.240
Wait. Girl. That's incredible.
00:15:45.240 --> 00:15:49.480
It's I'm putting them so I'm putting them around San Francisco as art installations Yeah.
00:15:49.480 --> 00:15:52.440
Like next to the tiny, like, borrowing libraries. Yeah.
00:15:52.440 --> 00:15:53.240
Oh my gosh.
00:15:53.240 --> 00:15:58.125
You can you can have, like, site specific films for each location.
00:15:58.125 --> 00:16:01.325
And then, like, they just have, like, a little battery in them that last Yep.
00:16:01.325 --> 00:16:06.924
And they're also they're also, like, super, like, low energy requirements.
00:16:06.924 --> 00:16:13.539
So even if you had kind of, like, a solar panel on top of the birdhouse, like, that would be enough to power it for for quite a bit of time.
00:16:13.539 --> 00:16:14.820
Yeah. That's really cool.
00:16:14.820 --> 00:16:29.284
I feel like this really gets into an area that Isabel and me to some small degree spent probably too much time around holding it too deep down, which was every web domain you own for a side project.
00:16:29.284 --> 00:16:33.445
Yes. And the sheer scale of Fun.
00:16:33.445 --> 00:16:36.684
Honestly, incredible.
00:16:37.460 --> 00:16:42.299
Like your website is just such a beautiful, curious, joyful experience.
00:16:42.500 --> 00:16:43.059
Thank you.
00:16:43.059 --> 00:16:49.779
And like just the amount of fun side projects and whimsy that you're bringing to all of these like really ******** tools.
00:16:49.779 --> 00:16:51.735
I'm really curious like, what are you working on?
00:16:51.735 --> 00:16:53.014
What tools are you building?
00:16:53.014 --> 00:16:58.134
Whether they're the funkiest things like silent bird houses or what are you excited about right now?
00:16:58.134 --> 00:16:59.014
Oh, man.
00:16:59.014 --> 00:17:03.455
So so I also love all of my all of my side projects.
00:17:03.940 --> 00:17:06.259
Like, they're they're super fun to build.
00:17:06.259 --> 00:17:15.964
Like, I get a lot of joy out of them, but they're also kind of like sneaky ulterior motive is stress testing some of the features that we have coming out for the Gemini APIs in our models.
00:17:15.964 --> 00:17:33.410
You like hid vegetables in them for Well, it's the well, so so as an example, we released something just recently, our second iteration of the multimodal embeddings model, which is allows you to embed in the same space, video, audio, images, text, and code.
00:17:33.410 --> 00:17:44.755
So you can like pull together audio, video, like you could you could type in like Alan Kay and then see like every time Alan Kay utters words in a video.
00:17:44.755 --> 00:17:50.515
But also like every time he appears in a video, anytime he appears in a photograph, etcetera.
00:17:50.515 --> 00:17:53.394
And it's it's just really really cool to see that.
00:17:53.394 --> 00:18:04.930
So that's powering folklore dot dev which was pulling a lot of the the oral histories from the computer history museum into kind of like a contextual space.
00:18:04.930 --> 00:18:20.444
So you got to see the transcript, you got to see interesting anecdotes extracted, but you also got it grounded and kind of like, well, when they say this computer and they were able to do this on this computer, like, what kind of compute power does it have?
00:18:20.444 --> 00:18:23.605
How does that compare to your mobile device?
00:18:23.805 --> 00:18:26.605
Like, what would it cost in today's dollars?
00:18:26.605 --> 00:18:27.950
Those sorts of things.
00:18:27.950 --> 00:18:42.789
And it's the same with the stacks dot dev, which was pulling in all of the Byte and Omni magazines and more on the way to to kind of extract out insights using the Gemini models, whether it's like telephone numbers or addresses or whatever it is.
00:18:42.974 --> 00:18:46.734
And then also putting the covers and the articles in a similar embedding space.
00:18:46.734 --> 00:18:56.535
So if somebody like wrote in to complain about their experience using a Commodore sixty four, like what other articles are are similar in vibes to that?
00:18:57.039 --> 00:19:13.105
And the I also thought it was quite cool in the sense that like pulling in, you know, forty years or so of of those magazines and then mapping out the different, you know, places that people were sending letters from or like the stores and those sorts of things.
00:19:13.105 --> 00:19:18.625
When you visualize them in on a map, you can start seeing like, oh my god.
00:19:18.625 --> 00:19:22.865
Like in Houston, there were these pockets of people who all cared about this thing.
00:19:22.865 --> 00:19:30.640
And then, you know, like Sunnyvale or in Cupertino, there were like five people within a six block radius that cared about this thing.
00:19:30.640 --> 00:19:38.440
And you can start realizing that like, you know, geographic location might have impacted or influenced some of the creations of these products.
00:19:38.575 --> 00:19:38.974
Yeah.
00:19:38.974 --> 00:19:44.974
But all of that's like locked away otherwise, and now it's not, which is which is kind of amazing.
00:19:44.974 --> 00:19:56.730
It's it's interesting to hear it sounds like to you working on models and working so closely with a lot of these tools, like, gives you the opportunity to see, like, this is coming out.
00:19:56.730 --> 00:20:00.730
What new kinda, like, interesting thing can happen?
00:20:00.730 --> 00:20:01.930
Like, what can it unlock?
00:20:01.930 --> 00:20:04.410
Yep. And also the cost changes too.
00:20:04.410 --> 00:20:17.914
So, like, the Gemini 3.1 Flash Lite model, which came out just recently, ended up kind of like, it's like a ninety percent reduction in the costs associated.
00:20:18.035 --> 00:20:27.589
So significantly more powerful than you know, Gemini 2.0 Pro or 2.5 Pro, but at a fraction of the cost and much faster.
00:20:27.589 --> 00:20:29.349
And it's the same with the batch API.
00:20:29.349 --> 00:20:34.029
Like you can get a fifty percent reduction in costs just by using the batch API.
00:20:34.184 --> 00:20:47.539
So the cost of maintaining all of these websites over time has actually gone down even though like more features have been added and they've gotten much cooler just because the model prices are also going down pretty exponentially.
00:20:47.539 --> 00:20:52.579
Are you willing to tell us, like, how much does it cost to run one of those nowadays?
00:20:52.579 --> 00:20:57.220
Is it, you know, hundreds of dollars to have this all be running on your website?
00:20:57.220 --> 00:20:57.619
Yeah.
00:20:57.619 --> 00:21:03.579
So for so so I deploy everything via Cloud Run, which is which is pretty low cost.
00:21:04.244 --> 00:21:19.470
The and and some, I think, for for all of the things that you see on web page dot dev, now that I've replatformed to use Gemini three point one Flash Lite and Nano Banana two instead of Nano Banana Pro for the for the number of ones that use Nano Banana two.
00:21:19.470 --> 00:21:27.470
I think the most expensive service is the embeddings, but it's still been less than, like, two hundred dollars to maintain everything over the course of a year.
00:21:27.470 --> 00:21:28.430
Wow. Yeah.
00:21:28.430 --> 00:21:33.525
For as many prolific projects that are on there, I am shocked. Yes.
00:21:31.365 --> 00:21:33.525
That's impressive.
00:21:33.525 --> 00:21:35.365
It's it's super super cool.
00:21:35.365 --> 00:21:42.125
And, like, the there are also a lot of open models that are becoming frontier quality.
00:21:42.325 --> 00:21:47.039
So so I think the the cost is only going to go down more over time.
00:21:47.039 --> 00:21:48.000
Yeah. Yeah.
00:21:48.000 --> 00:21:57.555
It's really interesting to see too the like, all the projects that you're kinda like pace of putting out like really interesting projects and sharing out.
00:21:57.555 --> 00:22:03.634
It seems like the turnaround sometimes is just immediate that you might deploy a project Yeah.
00:22:03.634 --> 00:22:04.434
In hours.
00:22:04.434 --> 00:22:06.115
Is that This is true.
00:22:06.115 --> 00:22:10.674
Like, the the latest one, I haven't added it to my website yet, though this is a good incentive too.
00:22:10.674 --> 00:22:12.720
It's a traffic cam guesser.
00:22:12.720 --> 00:22:21.400
So it pulls in all of the live traffic feeds from around the United States for at least like a half dozen cities so far.
00:22:21.599 --> 00:22:27.279
And then like removes using Nano Banana to any street signs or identifying sites.
00:22:27.279 --> 00:22:31.924
And then you have to guess which neighborhood the traffic cam is from.
00:22:31.924 --> 00:22:39.045
And then, like, it gives you kind of GeoGuessr-style hints of like, here's what you should have seen to help guide your choice.
00:22:39.045 --> 00:22:45.150
And you can also use OAuth to log in and kind of track progress too.
00:22:45.150 --> 00:22:55.525
My favorite one though is the Wikipedia racer, which is like like I I am just I am enchanted by depths of Wikipedia and Wikipedia rabbit holing.
00:22:55.525 --> 00:23:00.565
And so having something that will just create Wikipedia rabbit holes for me is is pretty magical.
00:23:00.565 --> 00:23:01.285
Yeah. I love that.
00:23:01.285 --> 00:23:04.484
I I think I went on there too and there's like levels of difficulty as well.
00:23:04.484 --> 00:23:04.964
Oh, yeah.
00:23:04.964 --> 00:23:08.085
You really decked out it's not point a to b.
00:23:08.085 --> 00:23:12.460
There's actually you can customize your WikiRace journey.
00:23:12.460 --> 00:23:14.779
Is that how many cc's your Exactly.
00:23:14.779 --> 00:23:15.740
WikiRacer has.
00:23:15.740 --> 00:23:18.859
Yep. And then also you can commission your own adventures.
00:23:18.859 --> 00:23:28.845
If you want something that's like related to Kanye or like related to related to like meditative Gregorian chants or whatever it is.
00:23:28.845 --> 00:23:35.805
Like, it will create a Wikipedia race for you specifically specifically for that.
00:23:35.805 --> 00:23:42.490
And then it will also show you over time as more people try it like what their score is and how their paths diverge from yours.
00:23:42.490 --> 00:23:44.210
Oh, wow. Yeah.
00:23:44.410 --> 00:23:49.170
Do you think you've always been into this, like, speed run?
00:23:51.049 --> 00:23:52.730
Will it bring joy?
00:23:52.730 --> 00:23:56.075
Like, beautiful form of tinkering?
00:23:56.075 --> 00:24:06.029
Do you think like like when you're a kid, someone could have guessed this person's gonna speed run apps.
00:24:06.029 --> 00:24:06.750
Yeah.
00:24:06.750 --> 00:24:08.190
I oh gosh, no.
00:24:08.190 --> 00:24:12.029
I grew up in I grew up in like the tiniest little farm town in Texas.
00:24:12.029 --> 00:24:19.670
So like thirteen hundred people, fifteen miles away from the nearest grocery store sort of situation.
00:24:20.355 --> 00:24:24.994
And also in university, like, don't think any of my professors would have guessed it.
00:24:24.994 --> 00:24:32.035
Though they they though they probably would have said something to the effect of like, oh, Paige cares about computers too much.
00:24:32.035 --> 00:24:33.714
Like, she's gonna be a bad physicist.
00:24:33.714 --> 00:24:34.954
Like, it's it's like Yeah.
00:24:35.394 --> 00:24:50.009
So I think the world has changed kind of recently to to appreciate people who are whole humans, who care about not just like one thing, but like lots of things, and who are capable of seeing connections between lots of things.
00:24:50.009 --> 00:24:57.815
So a couple of favorite anecdotes, like Claude Shannon came up with information theory because he could took a philosophy class.
00:24:58.095 --> 00:25:06.140
Like there are other scientific breakthroughs that have been inspired by things that are completely outside of the sciences that are in the arts.
00:25:06.140 --> 00:25:21.460
And then also, as an example, Ted Nelson introduced Alan Kay and his wife because the Alan Kay's wife Bonnie was trying to write the screenplay for Tron.
00:25:21.765 --> 00:25:26.164
And she ended up basing like one of the characters in Tron on Alan Kay.
00:25:26.164 --> 00:25:26.724
Oh, wow.
00:25:26.724 --> 00:25:36.380
And that was the first that was the first screenplay to ever be typed and printed off on a computer because it was using the Alto at Xerox PARC.
00:25:36.380 --> 00:25:48.059
So it's like like one of the most beloved like entertainment things, Tron, like was only was only possible because of this interplay between like tech plus computers.
00:25:48.059 --> 00:25:53.875
And last one, Larry Tesler got inspired by zines for the copy and paste feature.
00:25:53.875 --> 00:25:54.355
Oh, wow.
00:25:54.355 --> 00:25:55.154
For for Max.
00:25:55.154 --> 00:25:55.394
Yeah.
00:25:55.394 --> 00:25:55.795
Right?
00:25:55.795 --> 00:26:04.234
Like, so so because he was like making zines, cutting out things and like pasting them together and like that inspired copy paste.
00:26:04.240 --> 00:26:26.674
So I I feel like, you know, the the magic happens in these margins and like people had historically kind of been pushed into these silos and now we're getting to the place where we're bringing back people into appreciating whatever they think they're bringing of Victorian like polymaths who like did all of these different Yep.
00:26:26.674 --> 00:26:27.234
Fields.
00:26:27.234 --> 00:26:34.710
It kinda feels like we're heading back in that direction a little bit And also we're, you know, like just doing what enchanted them too.
00:26:34.710 --> 00:26:44.549
Because they would like play an instrument, or they would like go hang out with their friend who was like doing art in the park, and like the it's, you know, those those things are important.
00:26:44.549 --> 00:26:49.654
Like, we should all be whole humans, not just like feel like we need to over index on one strength.
00:26:49.654 --> 00:26:51.414
How do you choose what to try?
00:26:51.414 --> 00:26:55.974
Like, how do you choose like, oh, I'm gonna stitch a to b.
00:26:55.974 --> 00:26:58.214
I'm gonna try this tool today.
00:26:58.214 --> 00:26:59.335
That's how are you figuring out?
00:26:59.335 --> 00:27:01.734
And like, how many, like, of the things you try?
00:27:01.734 --> 00:27:04.759
Like, how many do you are successes and like failures?
00:27:04.759 --> 00:27:05.400
Yeah.
00:27:05.400 --> 00:27:15.839
So some of some of them, like a lot of the time that the thing that actually gets published is not the thing that I originally wanted to do.
00:27:17.384 --> 00:27:28.944
I find that there's some sort of constraint in the tool or constraint in the API that I'm trying to call that means that I have to pursue an alternate path.
00:27:29.150 --> 00:27:46.444
So so as an example, one one thing that I that I had been working on was I also have a whole collection of just like emails that get sent to me every morning.
00:27:46.444 --> 00:28:01.779
One of which is pulling a new a new painting from the Art Institute of Chicago or like a new art piece from the Art Institute of Chicago describing it like like a rough and tumble Chicago and and like a in a cool way.
00:28:01.779 --> 00:28:11.460
And then also like peppering in like really interesting insights about the city or the author that you wouldn't expect, which is using the the Google search grounding feature.
00:28:11.460 --> 00:28:14.460
And then formatting it formatting it nicely.
00:28:14.634 --> 00:28:30.595
But the reason that that that happened was because the Art Institute of Chicago was one of the only ones that had an API that you could call that did not require kind of like username password and I didn't wanna store that in plain text.
00:28:31.460 --> 00:28:41.619
So so like that entire experiment was only possible because you could you could sort of ping the API and and get that response back.
00:28:41.619 --> 00:28:47.355
Like originally, had been like, oh, well, I should pull from a whole bunch of different museums and just do this generic thing.
00:28:47.714 --> 00:28:51.394
And then this gave it a little bit more of a flavor and personality.
00:28:51.394 --> 00:28:51.795
Yeah.
00:28:51.795 --> 00:28:52.194
Yeah.
00:28:52.194 --> 00:28:59.154
I think that's kinda one of the things I think I've learned from, like, making images of Nano Banana too is sometimes you've just gotta, like, roll with it.
00:28:59.154 --> 00:28:59.430
Yep.
00:28:59.430 --> 00:29:11.509
Like, you can't that that initial idea you had, you can't hold on it too hard because you're just just too frustrating to get to and just too good to be open to, like, what what the what the model brings you, what randomness brings you today.
00:29:11.509 --> 00:29:12.309
Exactly.
00:29:12.309 --> 00:29:26.630
Like, it's it feels one of the the things that's been most enchanting about AI as as a person who, you know, computer programmers historically have very much wanted to must control everything.
00:29:26.630 --> 00:29:31.630
Like, I must deterministically know the input and deterministically know the output.
00:29:31.990 --> 00:29:34.390
And now it's like, it's definitely not that.
00:29:34.390 --> 00:29:40.244
It's it's kind of like, alright, you can set the constraints or you can attempt to set some constraints.
00:29:40.244 --> 00:29:43.125
But what you're gonna get is real different.
00:29:43.125 --> 00:29:52.765
And so you either need to be able to kind of do that in an artistic way that appreciates randomness and serendipity.
00:29:53.190 --> 00:29:57.430
Or you're gonna get real frustrated and not wanna be a computer programmer anymore.
00:29:57.430 --> 00:30:05.109
Like, so That does make me wonder, like, if statisticians are kinda like have a bit of a heads up there because you're so used to the idea of like randomness.
00:30:05.109 --> 00:30:05.589
Yeah.
00:30:05.589 --> 00:30:19.525
Like, I always like found this idea of like, you know, like sort of quantum physics and it feels to like many people, I think that the foundations of the universe are probabilistic, that's just like horrible.
00:30:19.805 --> 00:30:22.924
But to me, I'm just like, of course it's that way.
00:30:22.924 --> 00:30:25.549
Totally makes sense to me.
00:30:22.924 --> 00:30:25.549
Like, why does it Exactly.
00:30:25.549 --> 00:30:27.069
Why does it need to be deterministic?
00:30:27.069 --> 00:30:27.869
Exactly.
00:30:27.869 --> 00:30:34.910
And and I think it it also is is a lot more inspiring in the sense that, you know, the real world is messy.
00:30:34.910 --> 00:30:46.785
Like, and and if you if you see something that's unexpected, then it it might inspire you to go a completely different way or to see something in a way that you wouldn't have imagined necessarily.
00:30:47.384 --> 00:30:50.025
And and it's also just a lot more fun.
00:30:50.025 --> 00:30:57.920
Like, I I feel like one of the the biggest challenges right now is that, you know, universities have kind of over indexed on problem sets.
00:30:57.920 --> 00:31:08.119
And now like if you have a problem set, like good luck finding a person that's actually gonna do the problem set as opposed to just like paste it into whatever their favorite chatbot is and get a response back.
00:31:08.565 --> 00:31:15.285
And so they're they're not quite learning anything unless you kind of set up the constraints for the assignment to be like, alright.
00:31:15.285 --> 00:31:18.325
Well, today, you have a whole bunch of different hardware.
00:31:18.325 --> 00:31:25.660
You're going to you build it and explain whatever physical processes are required and and throughout the the building phase.
00:31:25.660 --> 00:31:32.140
And if you're trying to connect something and it doesn't work, you have to explain why it didn't work in the physics behind that.
00:31:32.140 --> 00:31:41.174
Like, that's a much cooler assignment as opposed to just like, hey, you know, what is the velocity of a versus b?
00:31:41.375 --> 00:31:41.694
Yeah.
00:31:41.694 --> 00:31:50.759
When I like to think about teaching like data science today, it's like, you know, the cost of creating a shiny app or a little interactive applet to explore something is now like zero.
00:31:50.759 --> 00:31:52.359
Like you should be doing that all the time.
00:31:52.359 --> 00:31:56.920
Like we don't need to ask these like deterministic, is this right or not questions.
00:31:56.920 --> 00:32:00.519
We can be like, hey, create a tool to explore this and then tell us what you found.
00:32:00.519 --> 00:32:02.414
Yep. Or create your own data set.
00:32:02.414 --> 00:32:12.174
Like, out some like, figure out a really gnarly question that you wanna answer and then find a way to automate the creation of a data set that will help you answer the questions.
00:32:12.174 --> 00:32:12.414
Yeah.
00:32:12.414 --> 00:32:24.970
I actually had a really like, I had a use like, I tried this with like, I only realized, like, I flew to California, like, last week when the lines at Houston's airports were getting really, really long.
00:32:24.970 --> 00:32:33.325
And, like, only at that moment I was like, I need to be scraping the wait times and like recording them in a Parquet file on GitHub.
00:32:33.365 --> 00:32:41.549
And so I like I did like a little bit on the website to kind of find and I was like, Oh, there's actually a JSON endpoint I can call that gives me nicely formatted JSON.
00:32:41.549 --> 00:32:46.670
But I was like, oh, maybe, like, maybe there's like a historical API hidden here as well.
00:32:46.670 --> 00:32:49.630
And I was like, hey, Claude Code, like, can you just, like, try this out?
00:32:49.630 --> 00:32:55.134
And it came up with, like, fifteen different, like, plausible looking URLs and tried them all out.
00:32:55.134 --> 00:32:57.694
Like none of them did actually yield the historical data.
00:32:57.694 --> 00:33:03.295
Like one of them looked really promising, like it was like the API slash date, but then it just returned the same day.
00:33:03.295 --> 00:33:06.815
But just that kind of like exploration you can do now.
00:33:06.815 --> 00:33:08.589
Like you don't have to know a ton about APU.
00:33:08.589 --> 00:33:11.390
You need to know what an API is, a bit about how it works.
00:33:11.390 --> 00:33:19.230
But that ability to just like iterate, create datasets that matter to you, collect them, install them, think it's really it's really cool.
00:33:19.230 --> 00:33:19.549
Yep.
00:33:19.549 --> 00:33:22.704
And if they matter to you, they probably matter to somebody too.
00:33:22.704 --> 00:33:30.144
In which case, like, suddenly the world is filled with a lot a lot of more datasets that can be used or expanded.
00:33:30.144 --> 00:33:36.970
I mean, my my plan is to have, like, a very like, a hyper specific, like, IAH, like wait time website.
00:33:36.970 --> 00:33:39.769
So you can be like, this is what it's looking like today.
00:33:39.769 --> 00:33:43.289
And the like historical context, like Tuesdays look like this.
00:33:43.289 --> 00:33:45.369
And you call it Houston, we have a problem.
00:33:45.369 --> 00:33:46.569
Or like There it is.
00:33:46.569 --> 00:33:47.369
Yeah.
00:33:47.369 --> 00:33:48.690
Yeah. Yeah. Yeah.
00:33:48.755 --> 00:33:52.275
I feel like there's a real critical role here.
00:33:52.275 --> 00:33:55.234
A good opportunity here. Yeah. Yeah.
00:33:55.234 --> 00:34:04.200
So we heard a lot about your projects and we learned that your projects actually involve a lot of important things that are coming out.
00:34:04.200 --> 00:34:07.480
Like you've hidden your vegetables in them and we just didn't know we were eating them.
00:34:07.480 --> 00:34:19.094
I'm curious how it ties into being a developer relations engineering lead at DeepMind. Are these all connected somehow?
00:34:19.094 --> 00:34:20.534
A a little bit. Yep.
00:34:20.534 --> 00:34:25.255
So most of the most of the models that I'm using are DeepMind specific models.
00:34:25.255 --> 00:34:30.454
So so they would be like the different flavors of Gemini that we have.
00:34:30.454 --> 00:34:35.989
Mostly not pro, mostly Flash and Flash Lite and and some of the others.
00:34:36.190 --> 00:34:39.789
The Nano Banana models, Veo for video generation.
00:34:39.789 --> 00:34:44.590
We also have an open model family called Gemma, which we also expose via the API.
00:34:44.590 --> 00:34:51.474
So you can either download it and use it yourself or you can kind of use it via our API for zero dollars, which is quite cool.
00:34:51.474 --> 00:34:58.355
And then also, there are some tinier versions of the Gemini models that are small enough to fit on mobile devices.
00:34:58.355 --> 00:35:00.914
So we've embedded one within the Chrome browser.
00:35:00.914 --> 00:35:03.559
We've embedded them in Pixel devices.
00:35:03.559 --> 00:35:09.800
And you can kind of use the onboard intelligence without WiFi connection for for zero dollars.
00:35:09.800 --> 00:35:13.199
And that's that's getting better by the day as well.
00:35:14.275 --> 00:35:17.954
But but so those are kind of the the models that we support.
00:35:17.954 --> 00:35:22.675
I also mentioned JAX's, like, the machine learning framework and all of the affiliate libraries.
00:35:22.675 --> 00:35:31.710
And then we also have some labs experiments as well as a tool called Antigravity that's kind of like an IDE for for experimenting with these things.
00:35:32.230 --> 00:35:43.945
The but all of that is kind of the the products that we would expect developers and increasingly like people who are developer adjacent to use to build things.
00:35:43.945 --> 00:35:52.224
And I'm I'm actually really really excited about getting information workers to be able to create and to use these tools.
00:35:52.264 --> 00:36:00.949
So not just not just people who are coming from engineering backgrounds or who who understand software engineering, but like all of the people outside of that world.
00:36:00.949 --> 00:36:04.150
What's you said information workers?
00:36:04.150 --> 00:36:06.550
Yep. What's a information worker?
00:36:06.550 --> 00:36:06.869
Yeah.
00:36:06.869 --> 00:36:08.945
So it's So it's just me being a barbarian.
00:36:08.945 --> 00:36:09.105
No.
00:36:09.105 --> 00:36:10.224
No. No.
00:36:10.224 --> 00:36:11.105
Excellent question.
00:36:11.105 --> 00:36:18.465
So so like information workers, I I feel like it's it's such a nebulous term that could apply to to pretty much anybody.
00:36:18.465 --> 00:36:35.659
But like, people who, you know, are data analysts Or people who might be like working working in the scientists or sciences to like take in primarily CSVs and like do do interesting analysis on them.
00:36:35.659 --> 00:36:47.574
Like one of the first two things that I use TensorFlow for back when it came out a long long time ago was one of the the projects that they released with it was being able to classify five different kinds of flowers.
00:36:47.574 --> 00:36:58.550
But you could also just like have five different kinds of things that you wanted to classify and and do do that classification process pretty simply.
00:36:58.550 --> 00:37:04.829
Just kind of like a a move between the five different kinds of flowers to five different kinds of other things.
00:37:05.429 --> 00:37:25.840
And so in the earth sciences, as well as in the biological sciences, but there there were a lot of grad students, self included, who were having to hand classify like different kinds of features in reefs or different kinds of, you know, shapes of things on petri dishes or whatever it is.
00:37:25.960 --> 00:37:37.094
And instead of having, you know, a grad student spend a month like going through and doing that, like you could do it really easily just with the most basic classification task.
00:37:37.094 --> 00:37:54.150
And so so like all of these people who are working in the sciences, there are tons and tons of those kinds of use cases that are ripe for automation, that tools exist that would save time and energy and accelerate the scientific process that nobody just knows about.
00:37:54.150 --> 00:37:54.949
Yeah.
00:37:54.949 --> 00:38:08.284
So so like now that the models are good enough to do this kind of work, just blazing a path for people who are experts to to like automate away the parts that are frustrating and like focus on the parts that are fun.
00:38:08.284 --> 00:38:28.784
I think like inside businesses too, there's like a lot of people whose jobs like center around Excel spreadsheets and like painstakingly, carefully copying bits from one sheet to another sheet and making sure they line up or like getting information out of a Jira ticket and putting it on a spreadsheet and they've got a PDF and extracting three bits.
00:38:28.784 --> 00:38:32.704
There's all this stuff that's just like really like no one enjoys doing this.
00:38:32.704 --> 00:38:35.184
Like it's often like it's high stakes.
00:38:35.184 --> 00:38:37.505
Like you know it's really easy to make a mistake.
00:38:37.505 --> 00:38:42.320
It's like no one enjoys doing that and there's just so much potential to Yeah.
00:38:42.320 --> 00:38:45.519
And it also is like very**** ** the human body too.
00:38:45.519 --> 00:38:57.775
Because if you're like copying and pasting things, like a lot of people use mice for that, in which case like the the potential for RSIs, the repetitive stress injuries are really high.
00:38:57.775 --> 00:39:08.929
And so anything that can take away the parts of work that are boring and not fun, I I think is a is a great use for these models.
00:39:08.929 --> 00:39:11.969
Yeah. Yeah. It's an interest yeah.
00:39:11.969 --> 00:39:27.195
It's so helpful to hear, like, the different people or like the all the different tasks where people need this and that for something sometimes where it's like it sounds like you're saying also, like, it's gotta be right, but it's also a little bit mundane.
00:39:27.195 --> 00:39:28.795
Yeah. Like, these tools can really lift.
00:39:28.795 --> 00:39:33.755
Well, stuff like if, like, we had to do it, like, could write a program to do it and, like, really easily.
00:39:33.755 --> 00:39:39.050
But if you don't know how to program, you're just completely locked out of this automation.
00:39:39.050 --> 00:39:45.929
And I think there's still like a lot of questions to answer about how do we make sure they can do this for like slightly unreliably and put checks in places.
00:39:45.929 --> 00:39:48.130
But it's yeah, it's really exciting.
00:39:48.405 --> 00:39:52.485
Just a lot of human misery that we can hopefully eliminate.
00:39:52.485 --> 00:40:09.250
I'm really curious, like, as somebody who builds and uses and, like, combines all of these tools, how you decide to like, what tools you're using and how they're like, how to continue growing and learning and staying current, especially being in a world where things change on the day, on the hour.
00:40:09.250 --> 00:40:13.329
Like, how are you making decisions about where to go with the tools that you're using?
00:40:13.329 --> 00:40:20.934
All of these all of these models that get released, they're they're basically like supercharged engines.
00:40:20.934 --> 00:40:21.335
Right?
00:40:21.335 --> 00:40:37.800
Like and so, like, one of the one of the things that can be a little bit disheartening to hear sometimes is that people want one model to kind of be the one model that they use.
00:40:37.800 --> 00:40:48.474
Whereas I think that what we see is that you it's it's a much more effective path to, like, pick a model based on what it's really, really good at or what you need.
00:40:48.474 --> 00:41:01.820
Like like, if you need a model that has really low latency and that is really, really cost effective and that can get the job done for for a given subset of tasks, like pick that.
00:41:01.820 --> 00:41:05.579
Don't reach for like the highest end model every single time.
00:41:05.579 --> 00:41:20.835
And I find like so for the Gemini models, they're really good at analyzing video, PDFs with images embedded in them, audio, like translating between different languages, like doing real time interactions, those sorts of things.
00:41:20.875 --> 00:41:31.039
And I find that the Anthropic models are exceptionally good at kind of writing in a way that feels like I'm talking to a friend, and also writing code.
00:41:31.039 --> 00:41:47.304
And so so if I couple together like Gemini's ability to understand video, to generate images, to understand PDFs with like images embedded, and then also Anthropic's ability to like write code and then to maybe like plan out what the work should look like.
00:41:47.425 --> 00:41:49.344
Then that's a really magical experience.
00:41:49.344 --> 00:41:59.429
Whereas the if you were trying to use one model to do all of those things, like you would just not have a satisfying a satisfying journey.
00:41:59.429 --> 00:42:13.744
But that's but it's not I guess answering it would which is just like, like, it is a twenty four seven sort of situation of like constantly trying out new things and seeing what works for you and what works for the different tasks that you care about.
00:42:13.744 --> 00:42:15.824
It feels exhausting sometimes. Right?
00:42:15.824 --> 00:42:18.570
Like I've never really felt like FOMO Yep.
00:42:18.570 --> 00:42:23.530
In my career, but now it feels like sometimes it's almost like paralyzing.
00:42:23.530 --> 00:42:32.985
Like I can't be like if I'm working, I'm like, my God, maybe there's a new model that just came out that would like do my work for me or like, you know, do what I'm doing now better.
00:42:32.985 --> 00:42:37.304
It is like it's an exhausting time to be a This is the AI psychosis?
00:42:37.304 --> 00:42:38.385
Yeah.
00:42:38.585 --> 00:42:44.625
It's it is definitely like and the the cost of writing software is going to zero.
00:42:45.440 --> 00:42:48.239
But that means that taste is more important than ever.
00:42:48.239 --> 00:42:48.559
Right?
00:42:48.559 --> 00:42:56.559
Like, so the ability to the ability to create things that feel still human and special and serendipitous.
00:42:56.559 --> 00:43:15.385
Like like, one of the one of the the things that I loved most about, you know, like reading Byte magazine or like whatever whatever else is that I would sometimes find a thing that was something I absolutely needed at the time, but I would have never known to go look for it myself.
00:43:16.079 --> 00:43:47.579
And now I I feel like with with AI kind of driving, you know, the content that you see on social platforms or, like, the code that gets generated in a specific style or format or tone and that kind of like, again, is forcing people into silos or one way of viewing the world when the the thing that that's really magical is helping people understand that the world is bigger to to see it in a different way.
00:43:47.780 --> 00:43:53.059
And so so like that, I I try to view the models as just tools.
00:43:53.059 --> 00:44:02.135
And then part of my job is figuring out how to take their outputs and to craft them in a way that would resonate with with people.
00:44:02.135 --> 00:44:10.855
That really speaks to me as like somebody who has developed open source and telling people all the time, like, there's no language for and you understand like, you know, it's not R versus Python.
00:44:10.855 --> 00:44:12.170
It's the right tool for the job.
00:44:12.170 --> 00:44:15.210
And hearing you say, you know, that applies to models as well.
00:44:15.210 --> 00:44:19.610
And just trying to deconstruct that in my mind, like, this is a model I should be using for coding.
00:44:19.610 --> 00:44:23.724
This is the model I should be using for helping me with docs or something like that.
00:44:23.724 --> 00:44:27.324
Feels very empowering And it that we have these muscles already.
00:44:27.324 --> 00:44:30.204
Exactly. And it changes it changes monthly.
00:44:30.204 --> 00:44:33.804
So like the the ability to if not weekly. Right?
00:44:33.804 --> 00:44:41.679
Like the the ability to to just kind of try things out and to accept that things will change.
00:44:41.679 --> 00:44:45.440
And then like, again, statisticians have the edge.
00:44:45.440 --> 00:44:50.519
Like like, and people who have worked with data have the edge because we all understand data drift.
00:44:51.034 --> 00:44:59.355
So so it's it's just never been a more exciting time to work in this space, but also I I absolutely agree.
00:44:59.355 --> 00:45:01.034
It is kind of exhausting.
00:45:01.034 --> 00:45:03.034
Yeah. Yeah. It's so interesting.
00:45:03.034 --> 00:45:08.860
I know we're getting long and you have to hop a flight to LA for a hackathon and Heck yeah.
00:45:08.860 --> 00:45:09.739
It's gonna be awesome.
00:45:09.739 --> 00:45:12.059
It's the it's a generative media hackathon.
00:45:12.059 --> 00:45:16.579
It'll be like Veo, but also Nano Banana.
00:45:16.780 --> 00:45:20.264
And also hopefully like the Lyria API as well.
00:45:20.264 --> 00:45:23.224
So people can create music and like Oh, nice.
00:45:23.224 --> 00:45:28.545
I also don't doubt that on the flight, there's a high chance you push some kind of app.
00:45:29.065 --> 00:45:32.425
At time you're in the air, I fully believe that's possible.
00:45:32.425 --> 00:45:38.750
Well, I I do still need for the the webcam, like, the the traffic cam guesser.
00:45:38.750 --> 00:45:44.909
I still need to add it to my main website, but I also want to I also want to add more street view data in more cities.
00:45:44.909 --> 00:45:47.309
Oh, nice. So maybe, like, Houston is next on the list.
00:45:47.309 --> 00:45:50.695
There's not enough Texas representation, so maybe that's Yeah.
00:45:50.695 --> 00:45:51.815
I love that.
00:45:51.815 --> 00:46:00.215
I know we learned one really fun fact that we chatted about a lot too, which was that you have an app that roasts your last song choice.
00:46:00.215 --> 00:46:01.735
Oh, yeah. On Last.fm.
00:46:01.735 --> 00:46:07.119
So like the the last so Last.fm, I am also like a super super fan of.
00:46:07.119 --> 00:46:20.684
Like, I have a Scrobble history that is like a decade deep or or like a so so like if you look at Paige's listening choices like from from the last decade plus, like, there's definitely been an evolution.
00:46:21.045 --> 00:46:32.005
But the but it it will tell me, like, based on what I've just listened to, what my estimated ages, and then also, like, how pretentious am I?
00:46:32.005 --> 00:46:35.079
Or like, how basic based on those choices.
00:46:35.079 --> 00:46:35.639
Yeah.
00:46:35.639 --> 00:46:43.239
Could you if you had to pitch humanity on why they should get roasted on their last song choice, how would you do that?
00:46:43.239 --> 00:46:46.875
I well, so so we all take ourselves way too seriously.
00:46:46.875 --> 00:47:04.760
Like, we and and in all honesty, if you if you're listening to the same thing over and over again or reading the same thing over and over again or using the same model over and over again, like, the part of the joy of being a human is is like learning from new perspectives.
00:47:04.760 --> 00:47:12.755
So like, if you if you get roasted, like, it's a good incentive to kind of laugh at yourself a little bit and then maybe experiment with something new.
00:47:12.755 --> 00:47:14.835
Yeah. That's a great takeaway message.
00:47:14.835 --> 00:47:15.954
That's what I was gonna say.
00:47:15.954 --> 00:47:19.155
Yeah. I I can't think you know I I'm so inspired by your yeah.
00:47:19.155 --> 00:47:27.079
Your just ability to, like, code and mix things and just find serendipitous combinations.
00:47:27.079 --> 00:47:36.440
So can't can't thank you enough for for coming on, and I'm so excited for the next one or a dozen things you you do.
00:47:36.440 --> 00:47:37.079
Thank you.
00:47:37.079 --> 00:47:48.684
Thank you so much for having And I I really feel like, you know, there's never been a more more interesting time to care about data and to care about the way in which information is communicated to folks.
00:47:48.684 --> 00:47:51.485
Yeah. Yeah. Hundred percent.
00:47:48.684 --> 00:47:51.485
Yeah. Awesome. Thanks. Yeah.
00:47:51.485 --> 00:47:51.885
Excellent.
00:47:51.885 --> 00:47:52.125
Thank you.
00:47:52.125 --> 00:47:54.244
Thank you so much. Yeah. Yeah.
00:47:54.889 --> 00:48:01.690
The Test Set is a production of Posit PBC, an open source and enterprise tooling data science software company.
00:48:01.690 --> 00:48:05.929
This episode was produced in collaboration with creative studio, ADJY.
00:48:05.929 --> 00:48:12.130
For more episodes, visit thetestset.co or find us on your favorite podcast platform.