Behind the Scenes of Nano Banana Pro

00:00:00:10 - 00:00:20:12
Unknown
Coming up, Jeff Jarvis and I take a close look at Google's new Nano Banana Pro model, and we even try to get it to create an infographic detailing John Le Kuhn's latest research. Let me tell you, it didn't go so well. Also, Nvidia didn't pop the bubble, but Google's encroaching on its turf. Is there something going on there?

00:00:20:12 - 00:00:39:24
Unknown
We talk about that. And Warner Music Group strikes a deal with Suno, which we all saw coming. Right. That's next on the AI inside podcast.

00:00:39:26 - 00:00:59:12
Unknown
What's going on, everybody? Welcome to another episode of the I Inside podcast, the show where we take a look at the AI that is layered like, I was going to say lasagna, but that's not very Thanksgiving. I don't know what's layered in a Thanksgiving meal. I'm one of your hosts, Jason Howell. Maybe Jeff Jarvis knows what might be layered.

00:00:59:15 - 00:01:02:00
Unknown
Would it be a bean casserole layered like a bean

00:01:02:05 - 00:01:23:26
Unknown
And that that works. That works? Yeah. It could be like a Napoleon if you're a really fancy and, and bougie. So my my sister and I were. She's coming tomorrow. We were comparing things from our heritage, and, the worst thing that our mother made for us on holidays was what was called a salad with scare quotes around it.

00:01:23:28 - 00:02:15:18
Unknown
It was made with lime Jell-O, Cool Whip, cottage cheese, pineapple and maraschino cherries. It was just awful. I had to dread it every year. Had to wear the salt. My mother oh no, you don't like the salad. I never have salad. Yeah, there's no lettuce. Oh, God. Yeah. God knows what magazine that was ripped out of.

00:02:15:20 - 00:02:30:13
Unknown
Yeah.

00:02:30:16 - 00:02:54:13
Unknown
Yeah. America. Yeah. It was. It was the view of mass culture in America, right? It was influenced by TV and magazines and. And the brands like that. My, my my brother's, it was, it was to sell Cool Whip and Jell-O. So somebody sat somewhere and said, how could we possibly put these things together? Who would think that? Cool Whip, Cool Whip?

00:02:54:13 - 00:03:16:03
Unknown
But it was awful. Non whipped cream. Whipped cream and and jello which is pure sugar. Yeah. No.

00:03:16:05 - 00:03:21:22
Unknown
Yeah yeah yeah yeah yeah.

00:03:21:24 - 00:03:41:11
Unknown
Oh right right. Mandarin oranges. Yeah. Oh yeah. Those were hot for a while. Yeah. Yeah. Yeah.

00:03:41:13 - 00:03:44:25
Unknown
Sweet and colorful. Yeah.

00:03:44:28 - 00:03:58:16
Unknown
No no. Sorry. For those of you listening in five days this is obviously we're recording the day before Thanksgiving.

00:03:58:18 - 00:04:04:03
Unknown
Yes.

00:04:04:05 - 00:04:15:19
Unknown
And if you're in Canada it's a different time. Time of the of the calendar. Yeah.

00:04:15:22 - 00:04:23:03
Unknown
Yeah. Wow. Yeah.

00:04:23:05 - 00:04:32:14
Unknown
It's pretty amazing. Yeah. Who to funk.

00:04:32:17 - 00:06:19:04
Unknown
Yeah.

00:06:19:06 - 00:06:41:06
Unknown
Yeah. That's the great thing.

00:06:41:09 - 00:07:01:14
Unknown
Cool.

00:07:01:16 - 00:07:05:29
Unknown
Oh.

00:07:06:01 - 00:08:22:15
Unknown
Yeah. Google is really, If anyone thought they had I cooties, they're expunged. Now Google is kind of the leader and heat right now on I.

00:08:22:17 - 00:08:50:19
Unknown
Yeah. It's it's pretty good. We're we're both fans of Nate Jones. He does he does amazing stuff with I. So for those of you who are watching you're seeing a Lego diorama go on behind. He wanted to look at the at the life of one token, and he instructed that whole banana with one, one, prompt to create a Lego, storyboard for what happens.

00:08:50:19 - 00:09:01:03
Unknown
And it's intricate and exacting, and he says it's really good. I can't really judge it because it's moving behind him. But.

00:09:01:05 - 00:09:24:22
Unknown
Yeah, probably dead. And that's a that's a tough critic. And and he's just wowed by it. He also put up his resumé or his or some of his resume, in the style of, you know, some middle medieval game. And that was pretty amazing, too. So, so the ability to create with this is, is quite something.

00:09:24:25 - 00:09:33:22
Unknown
And I'm putting Jason right now into the rundown. The Google announcement.

00:09:33:25 - 00:10:33:17
Unknown
There line 16.

00:10:33:19 - 00:12:15:15
Unknown
Oh.

00:12:15:17 - 00:12:42:03
Unknown
Yeah. I just pulled up to, What's happening to is these these things are coming as distinct products, but then they're being added into other products. So notebook Elm, put out a whole bunch of releases this week. They can now do slide decks. It's probably the first, natural banana. Fully dependent product. People are exploring fonts, styles, trips and text, but most of, but it's cool.

00:12:42:03 - 00:13:16:05
Unknown
Infographics, as you just showed, you can now do a notebook. Elm. And and so on and so on. So, so this this advance that Google is doing for once, they're also, doing a good job of coordinating at least the beginnings of it, with other products.

00:13:16:07 - 00:14:01:25
Unknown
Did you have to bring a pie?

00:14:01:27 - 00:14:10:25
Unknown
Wow.

00:14:10:28 - 00:14:29:29
Unknown
Product. Right.

00:14:30:01 - 00:14:37:15
Unknown
Yeah.

00:14:37:17 - 00:15:00:04
Unknown
They got great press. Killer press, as they should have.

00:15:00:07 - 00:16:12:17
Unknown
Yeah, yeah, yeah. And and we saw a lot about, open. I got sobered up fast from all this. Google did a great a great job with this. And the nano banana stuff is fun. I haven't played with it enough yet. But, I want to see what it can do.

00:16:12:19 - 00:16:31:05
Unknown
The history of mass media.

00:16:31:07 - 00:17:02:03
Unknown
Right.

00:17:02:05 - 00:17:07:11
Unknown
I presume these are separate teams, too. And that's part of the reason.

00:17:07:13 - 00:17:24:15
Unknown
Yeah.

00:17:24:17 - 00:17:53:04
Unknown
It's it's pretty good. Yeah. The preindustrial era from, you know, cave paintings up through, printing press, my stuff then, then the industrial electronic era, pushes a lot together from telegraph to television. That's a lot. Then the broadcast era, I think that's that mushes to the past and then the digital information era. Simple.

00:17:53:06 - 00:17:57:29
Unknown
I think it just continues.

00:17:58:02 - 00:18:16:14
Unknown
Yeah. There's no dates and stuff, you know, but it's. Yeah, it's it's it's passable as a, as a slide. It would be interesting. I wouldn't do it now, but it'd be interesting to press it and say, okay, I'll put in dates, you know, do what what Nate Jones did and don't let space limit you and, see what I could do.

00:18:16:16 - 00:18:22:11
Unknown
Oh.

00:18:22:13 - 00:18:59:25
Unknown
Include dates and details. Include dates, inventors and details. Just to drive it nuts. Nate said he had to actually go into a different version of, He had to use a key that he had to make those giant things.

00:18:59:28 - 00:19:05:29
Unknown
Right, right.

00:19:06:01 - 00:19:11:00
Unknown
Yeah.

00:19:11:02 - 00:19:28:28
Unknown
Same size, but. Yeah. Okay, good. But got it. Got dates in it. It redesigned it to be able to do that. It put in, the images of inventors. Yeah. Pretty good. Pretty darn good.

00:19:29:01 - 00:19:37:08
Unknown
Yeah.

00:19:37:11 - 00:19:51:08
Unknown
I was just think of that. Yes, exactly.

00:19:51:10 - 00:20:33:27
Unknown
Yep.

00:20:33:29 - 00:20:55:29
Unknown
This is not out of bananas, Gemini, but I use Gemini. The other night, somebody who was having health issues, and I took their history and their labs and put it in there, and it got a tremendous explanation of what was going on. And then I had made some mistakes. I said, well, actually thought that it was currently on one medication and they weren't.

00:20:56:05 - 00:21:07:09
Unknown
So back in I said, well, what is this change? And it changed it, right? And it changed it again. And it's pretty damn impressive.

00:21:07:12 - 00:21:14:04
Unknown
All right, Jason, you put a picture up on the rundown that's frightening me.

00:21:14:06 - 00:21:30:04
Unknown
Hey, is that real or is that I. I can't tell.

00:21:30:07 - 00:22:18:04
Unknown
I was.

00:22:18:07 - 00:22:50:27
Unknown
Well, I'm now very disappointed that the that the nano banana Christmas sweater is not swag that you got. Wow. That would be the coolest thing on earth, Yeah, yeah.

00:22:50:29 - 00:23:00:23
Unknown
In the matrix. In the matrix, Jason's wearing tacky sweaters.

00:23:00:25 - 00:23:56:11
Unknown
Yeah. Cool.

00:23:56:13 - 00:24:28:24
Unknown
Well, at the moment. At the moment, the sky did not fall at all. Nvidia showed tremendous growth. A tremendous confidence. The stock went up like crazy, recovering some of what had lost recently. And, so all was well. And then came the spoiler, which is Google. Google spoiled Microsoft for us last week, making its announcements just as Microsoft was, starting its conference.

00:24:28:26 - 00:24:59:29
Unknown
And then, and then, things look better for Google. Because it did a chip deal. Is this go too far down in the rundown? No, it did a chip deal with meta. So now Nvidia has, for tensor chips. So the market saw this as a major competition to Nvidia. So so but both you know I saw this The York Times had a story I think it was today or yesterday.

00:25:00:01 - 00:25:14:06
Unknown
Lead story was worry about the bubble bursting. And it's it's as if media coverage is wishing for it. And my view.

00:25:14:09 - 00:25:17:29
Unknown
Yeah.

00:25:18:02 - 00:25:33:15
Unknown
That and it's it's it's drama. It's a big story. Plus they don't like the technology. They don't like the technologies and the companies, they're they're hoping for it to happen. So there's this constant drumbeat of blah blah blah blah blah blah blah blah blah blah blah blah, blah, blah, right. And but at the same time, then you have Nvidia.

00:25:33:17 - 00:25:59:22
Unknown
The headline is if the Wall Street Journal, Nvidia strong was all show AI fears are premature. And then with the Google work why Google soaring stock is defying fears of an AI bubble. The reality keeps pushing against this media narrative of of immediate disaster. And there's also more competition here, which is a good thing. So Nvidia seem to own the world, and now Google is pushing up and that's, that's I think good.

00:25:59:22 - 00:26:26:10
Unknown
And then Nvidia came along and they had to say, oh we're happy for Google. So they had a post here that said we're delighted by Google's success. They've made great advances in AI and we continue to supply to Google. Nvidia video is a generation ahead of the industry. It's the only platform that runs every AI model and does it everywhere.

00:26:26:10 - 00:26:53:27
Unknown
Computing is done, Nvidia blah blah, blah blah blah. So Nvidia fascinatingly there, I think has to say, we everywhere we can run everything, and Google is a customer so they can't smash Google. They've got to say, good for you little Google. See if you can catch up with us. Fascinate.

00:26:53:29 - 00:27:11:21
Unknown
Yeah.

00:27:11:23 - 00:27:18:29
Unknown
Yeah, yeah.

00:27:19:02 - 00:27:48:01
Unknown
Yeah, it is a very big deal. It's a whole. Yeah. Yep yep yep. And I think the fear one of the fears or the bubblicious fears is the, limited, supply of chips. And now, if tensor chips are seen as credible competitors to Nvidia chips, then that means the supply is not as limiting as it seemed to be.

00:27:48:04 - 00:28:18:05
Unknown
At the same time you had you had, this is true. I didn't put in the rundown and think maybe this last week, Google must double AI serving capacity every six months to meet demand, double every six months. There I infrastructure boss told employees. That's just amazing. Just amazing. So so this demand is high, but there is there is supply and the supply is high, but there's demand.

00:28:18:07 - 00:28:33:22
Unknown
So happiness all around.

00:28:33:25 - 00:29:02:24
Unknown
And by the way, we're not making predictions. There could be a bubble. It could burst. Yeah we don't. And we've talked on the show in the last few weeks about whether it's an open AI bubble. Is it an LLM bubble. Is it a real bubble or not? Only the future will tell you.

00:29:02:27 - 00:29:14:11
Unknown
Yeah, I agree.

00:29:14:13 - 00:29:31:17
Unknown
Yes.

00:29:31:20 - 00:29:50:05
Unknown
No.

00:29:50:08 - 00:30:17:09
Unknown
Yeah. I have two major concerns. One is the we talked about this on the show is the, CapEx build out of all these data centers that if they're not needed this much, then there's a lot of money that's, that's hanging on that. And that could deflate the second concern. The bigger concern is do we really know what the business model of AI is yet do we really know where the revenue stands?

00:30:17:11 - 00:30:39:09
Unknown
There's lots of venture money going in. Lots of orders going on, lots of playing with it. Right. Lots circular funding. Is there have we yet seen and I don't think we have. Have we yet seen the data that says that there's real productivity gains, and real profitability gains outside of the AI companies in the, in other industries.

00:30:39:11 - 00:30:48:29
Unknown
And there is contention that there is. But we need we need data. We need to say, is this real or not? And I think so far it's iffy.

00:30:49:02 - 00:30:59:04
Unknown
So those are the two. Those are the two pins outside the balloon. To me.

00:30:59:07 - 00:31:13:12
Unknown
Okay.

00:31:13:14 - 00:31:23:29
Unknown
Yay yay.

00:31:24:01 - 00:31:33:12
Unknown
You're going to have to get a better a bigger board room meeting to have them all in. Yeah.

00:31:33:15 - 00:31:45:20
Unknown
It's a good thing.

00:31:45:23 - 00:32:11:22
Unknown
Thank you, thank you, thank you.

00:32:11:24 - 00:32:15:28
Unknown
Right.

00:32:16:01 - 00:32:19:24
Unknown
Genesis mission.

00:32:19:26 - 00:33:12:21
Unknown
All right.

00:33:12:23 - 00:33:36:20
Unknown
What's not to love? I think that government data is stuff that we have already paid to create. And, I think it should be made available to, the industry and to academics as well, and startups as well. So I think that's a great thing. And, the more of this there is, the better. I think.

00:33:36:22 - 00:33:57:27
Unknown
You know, I these days, I find myself often talking about how news companies are trying to cut off AI, and I think they're cutting off their nose to spite their face as a result. And the discussion that I want to see us have about the responsibility we have to the larger information ecosystem, I think media and news media are failing there, and we know why the copyright fight, blah blah blah, you're stealing my stuff.

00:33:57:27 - 00:34:27:09
Unknown
So on, so forth. But I think that, in general, we ought to be pouring more and more reliable data sources into this world because who knows what can come out of it. And, certainly government data should be out there. I think companies should be asking the same things. You know, some years ago, believe it or not, I was I was brought in to Switzerland to speak at a, pharma, company, like how to make a googly pill or something like that.

00:34:27:09 - 00:34:46:05
Unknown
I don't, I can't recall what it was. And I didn't realize it was dumb of me, but I didn't realize how much pharma is all about finding a molecule that does this or doesn't do that. And the part of the discussion was when they have a failure and they have many, many, many failures, they try something against a hypothesis.

00:34:46:06 - 00:35:04:28
Unknown
It doesn't work. The question that they were grappling with the time, and I don't know where it stands in the industry today, is should we share that? Yeah, in a sense where we're letting other people off the hook and we're going to the expense of having had the failure and others are not. But if they do the same thing, then we're not going to go down there.

00:35:05:00 - 00:35:36:18
Unknown
The dead ends, they went down. Right. So industries of all sorts have data that I think they should be making available in this world. And, you know, I was thinking the other day too, Jason, how, you know, now the Google has the has the halo and the the the gold medal for the moment on AI, the fact that it created labs, with Transformers, they could have held that they didn't write that paper.

00:35:36:21 - 00:35:59:00
Unknown
They didn't have to put that out there. They could have held that advantage for years. Like postscript never was, patented because they thought it was going to be too easy to then adapt. And they held the secret hope he held the secret in for years. Google could have held transformer in as a secret for as long as they could have.

00:35:59:00 - 00:36:59:10
Unknown
They didn't. They put it out there. Government data is out there. Company data should be out there. They're leapfrogging each other and it's good for everybody. So I think this is, something to really praise from the administration. Put that in your history books.

00:36:59:13 - 00:37:11:25
Unknown
There's that, there is that. Yeah.

00:37:11:28 - 00:37:30:07
Unknown
Yeah, I agree, I absolutely agree. And, you know, any research that is paid for by government should be open because we've already paid for it and should be open to other, other academics, but also and in companies, but also the AI. And one would hope we'd have a lot more of that. That's where we lead the world.

00:37:30:09 - 00:37:45:04
Unknown
You know, the Chinese are doing this. They're certainly taking everything that they pay for and putting it together. So yeah, I agree, Jason.

00:37:45:06 - 00:38:48:17
Unknown
Little irony here. Yeah.

00:38:48:20 - 00:39:17:20
Unknown
Trigger or the panic trigger the anger. But. Never mind.

00:39:17:23 - 00:39:32:21
Unknown
But they should have they should have called Google first. And they did have to eat crow and say they've updated the article. Realizing we contributed to a perfect storm of misunderstanding. Yeah.

00:39:32:24 - 00:39:36:06
Unknown
Not good.

00:39:36:08 - 00:40:11:05
Unknown
And once this impression gets out there, it's hard to erase it, as we know with all kinds of disinformation and misinformation. So it was irresponsible of them.

00:40:11:07 - 00:40:26:19
Unknown
Until until either they withdraw it or it ends up in court and that gets thrown out, or who knows? I don't know what all. I have read the suit, so I don't know what all they're claiming.

00:40:26:22 - 00:40:36:25
Unknown
So far. Yep.

00:40:36:27 - 00:40:48:01
Unknown
You can relax.

00:40:48:04 - 00:43:04:01
Unknown
No.

00:43:04:03 - 00:43:14:10
Unknown
So I put a story up there. Underneath that story which has been a little confused about the combination here. So pseudo

00:43:14:13 - 00:43:42:05
Unknown
Said that somebody got Ahold of the, according to Billboard, they got Ahold of, investor presentation materials. Which said that, sooner apparently doesn't, the company has paid only $2,000 in total on music used to train its models, said Billboard. So I'm not sure what the fight was with Warner Music. They were using the music that people created to train the model.

00:43:42:07 - 00:44:07:14
Unknown
So I'm confused here. According to Billboard, which obtained the investor presentation materials, it also said that users are generating 7 million songs per day on Suno over the course of two weeks. They create as many songs as are currently available on Spotify. Yeah, yeah. And I don't know how much of it we would consider, you know, tuneful slop.

00:44:07:17 - 00:44:26:24
Unknown
Or is it a case where I, since I can't I'm not like you, Jason. I'm not a musician. Can I express myself musically in a way that I couldn't? And it's a good thing, you know, I don't I don't know, depends on what we create. It says the figure is the $2,000 is not low, because Suno hasn't used the work of other artists to train its models.

00:44:26:26 - 00:44:38:18
Unknown
No, the whole platform says Gizmodo. And the work that it spits out is built entirely on the music that was made by creatives who had to find ways to learn to play on their own.

00:44:38:20 - 00:44:58:13
Unknown
According to Pitch Deck, it was allegedly sent to investors. The company spent only $2,000 on training data because it has been scraping songs for free, allegedly by ripping them from YouTube. Okay, so that's the problem. All right, now we see the problem. So that's the claim of the labels. So they only spent $2,000. They should have spent more.

00:44:58:15 - 00:45:05:00
Unknown
Got it. Got it.

00:45:05:02 - 00:45:19:09
Unknown
Yes.

00:45:19:11 - 00:45:51:10
Unknown
Yeah. No.

00:45:51:13 - 00:46:06:24
Unknown
Yeah.

00:46:06:26 - 00:46:11:02
Unknown
No, no.

00:46:11:05 - 00:46:54:02
Unknown
Yeah.

00:46:54:04 - 00:47:11:14
Unknown
Yeah. Does.

00:47:11:16 - 00:47:36:22
Unknown
Yeah. And, I think that the copyright issues are unsettled in courts and, and so they realized they had to do something for now because as a song writer, you can sit down and you can listen to people's music. And as long as you don't replicate their music, you can be inspired by it. Right? And I even do it in the style of.

00:47:36:24 - 00:48:28:08
Unknown
And so if that's allowed for you, Jason Howell, why shouldn't it be allowed for pseudo?

00:48:28:10 - 00:48:40:16
Unknown
DJs.

00:48:40:19 - 00:48:58:07
Unknown
To.

00:48:58:10 - 00:49:02:29
Unknown
Yeah.

00:49:03:02 - 00:49:10:11
Unknown
Yeah. It would be a shame because again it this this democratizes music to an extent.

00:49:10:13 - 00:49:41:02
Unknown
Well that was that was the Gizmodo story. It was talked about how expensive it is to sing and and sooner reduces that cost of creation and music.

00:49:41:05 - 00:49:45:00
Unknown
Well, it also it's just buying equipment.

00:49:45:03 - 00:49:53:03
Unknown
Are you are you a drummer? Jason.

00:49:53:06 - 00:50:16:11
Unknown
So, so so if you had a if you had to try you. So you play keyboards. You play guitar, right? Do you play any other instruments like woodwind or brass? Okay, so the point is you could do tracks, right? You could, you could do those tracks, but you could just, at the end of the day, still be missing a drum track.

00:50:16:13 - 00:50:51:23
Unknown
And obviously there are plenty of, of, of, of, of electronic drum tracks that exist. Now, the difference here would be that it does it appropriately to your song. You could add that in, and you could have a fully realized song that you otherwise couldn't have. That's pretty damn cool.

00:50:51:25 - 00:50:56:11
Unknown
Yeah.

00:50:56:14 - 00:50:59:23
Unknown
So.

00:50:59:25 - 00:51:07:11
Unknown
Yeah.

00:51:07:13 - 00:51:12:22
Unknown
That's really cool. Yeah.

00:51:12:24 - 00:51:18:23
Unknown
Exactly. You can't, you can't pull in an orchestra. Right.

00:51:18:25 - 00:51:53:04
Unknown
Well now that you're out of the garage you can probably fit it in there. It's okay.

00:51:53:07 - 00:53:01:11
Unknown
moment.

00:53:01:14 - 00:53:20:23
Unknown
Well is it a prompt objection is telling you to disobey something that is there. Right. There's two ways. Is that nobody thought to make the guardrail. Then you're there. I'll have it. Give me a bomb rest because nobody told me not to write. So is it a prompt injection? Tell me if I'm wrong. Here. This is. This is my amateurish question.

00:53:20:26 - 00:53:59:17
Unknown
Isn't a prompt injection, then basically a way to get it to ignore a guardrail that it has?

00:53:59:20 - 00:54:14:03
Unknown
See? Right. Okay. That's the other scenario. Yes. You're right. Okay. Got yes.

00:54:14:05 - 00:54:20:23
Unknown
Execution.

00:54:20:26 - 00:54:37:23
Unknown
Yep. Okay.

00:54:37:25 - 00:54:40:16
Unknown
Right.

00:54:40:18 - 00:54:46:15
Unknown
So what does it take to to to recognize and prevent that. Right. Got it.

00:54:46:17 - 00:55:00:10
Unknown
And that's that's in the code. That's because of these executable. So this is true in a coding context more than in a chat context.

00:55:00:12 - 00:55:04:23
Unknown
No one could do it in either place probably. Yeah.

00:55:04:25 - 00:55:27:21
Unknown
Hey Janice. Right okay. That's helpful. Thank you.

00:55:27:24 - 00:56:34:05
Unknown
Yeah.

00:56:34:07 - 00:57:08:15
Unknown
And certainly I understand the caution and the concern around young people using these tools, but this strikes me as a version of the stranger danger worries that came to the internet. And by cutting off young people, I wonder whether there could be greater damage.

00:57:08:17 - 00:57:32:03
Unknown
What if you feel terribly lonely, terribly depressed and ChatGPT helped you through that? Or character? AI or any of them helped you through that? And yes, there are the edge cases of something terrible happening, but what's the more common case? Are young people who felt lonely and felt depressed and felt as if they had an outlet. They're now cut off.

00:57:32:03 - 00:57:58:27
Unknown
What happens to them? And I don't think there's been there's been a lot of media panic around this and not nearly enough good research and good data. And that's concerning to me. This is comes out of a moral panic. So I'm afraid we're going to see from experience what happens to these young people, who were very sad today.

00:57:59:00 - 00:58:55:14
Unknown
Yep.

00:58:55:16 - 00:58:59:01
Unknown
No, no.

00:58:59:03 - 00:59:09:21
Unknown
It was delicious.

00:59:09:24 - 00:59:17:07
Unknown
Kind of set it differently. Yeah. I mean I remember the first time I held an iPhone, I said to myself this is hand candy.

00:59:17:10 - 00:59:28:03
Unknown
Right. Yeah. Yeah. You knew you had something that was, that was you wanted you wanted to touch it. You wanted to be with it.

00:59:28:05 - 00:59:48:24
Unknown
It depends on what it actually does. Will be the question we're waiting for.

00:59:48:27 - 01:00:10:25
Unknown
No.

01:00:10:27 - 01:00:42:08
Unknown
Right, right. And does it time to classes? Does it tie into separate screens? Is it just audio. What do I yeah I don't know what we're going to talk about.

01:00:42:10 - 01:01:12:13
Unknown
Yeah.

01:01:12:16 - 01:01:25:00
Unknown
Well a video of, you know, makes racks and makes data centers too. Yeah. I it's very I need to see a Venn diagram of who does what in this world.

01:01:25:02 - 01:01:43:13
Unknown
I was just thinking that I think I might, I might be back to work.

01:01:43:15 - 01:02:01:24
Unknown
He put up a post that acknowledged it, thanking his bosses and so on, but said that's taught.

01:02:01:26 - 01:02:12:26
Unknown
Oh yeah.

01:02:12:28 - 01:02:32:10
Unknown
I would certainly imagine meta wants to and will invest in what everyone is doing. Yeah, yeah. And if you go quickly to line 150, Jason, so this is a new paper by, two authors, including Yann LeCun. And so I said, oh, as I go through my papers over week, I see, oh, here's a paper by Yann.

01:02:32:17 - 01:02:54:04
Unknown
Let me take a look at it. Couldn't get past, like the first since, the the GPA. Provable and scalable. Self-supervised learning without heuristics. So I took the whole thing, and I put it into Gemini, and I said, explain it to me, and it kind of did, but I, I couldn't repeat it to save my soul.

01:02:54:04 - 01:03:10:10
Unknown
And so I put that up on online. I said I had to ask Gemini what this was, on Facebook. And John liked the post like, because I couldn't understand it otherwise.

01:03:10:13 - 01:03:55:29
Unknown
Oh, that's another step beyond. Okay.

01:03:56:02 - 01:04:04:22
Unknown
Hahahahahahahaha.

01:04:04:24 - 01:04:26:08
Unknown
Oh I see oh, okay.

01:04:26:11 - 01:04:49:13
Unknown
So I have a download of the sample. Let me ask you, Jason. So this is a browser. So it operates within the browser windows you have open on your phone. It doesn't it doesn't operate within other apps on your phone. Just within the browser and and I'm curious since you are you are doctor Android the king of all Android.

01:04:49:15 - 01:04:59:07
Unknown
Do you? I don't have any. I probably say this this way. I have no sense of tabs on Chrome on my phone.

01:04:59:10 - 01:05:19:29
Unknown
Or windows. I just I just use it and then go on to do something else. I don't have any sense of having tabs across and navigating among those tabs and knowing what's on those tabs. So for them to come up with a browser on Android doesn't seem terribly useful to me, because it's limited to interacting with just the tabs that are on your browser.

01:05:20:02 - 01:05:43:08
Unknown
And that's not the way I use a phone. How about you?

01:05:43:10 - 01:06:28:20
Unknown
Yeah.

01:06:28:23 - 01:06:35:21
Unknown
Got it.

01:06:35:23 - 01:07:28:09
Unknown
This is a little off topic, but how much are you hearing about the aluminum British pronunciation and spelling os. Finally bringing together Android and Chrome. Do you think that's going to make a difference in just this kind of functionality?

01:07:28:12 - 01:07:44:12
Unknown
Yeah.

01:07:44:14 - 01:07:57:28
Unknown
Yeah, yeah.

01:07:58:00 - 01:08:24:05
Unknown
It's it's probably a remake of both OS isn't if you're going to have to go to that fundamental a structural redo of one or the other, you might as well do both together.

01:08:24:08 - 01:08:28:27
Unknown
But thank you for that.

01:08:29:00 - 01:08:37:27
Unknown
Oh well.

01:08:38:00 - 01:08:49:27
Unknown
Yeah. Probably.

01:08:49:29 - 01:09:03:29
Unknown
It was too but I, it looks we feel better. It was too much for Jim and I, I don't feel so stupid.

01:09:04:01 - 01:09:18:05
Unknown
On that. Are you in the Christmas sweater?

01:09:18:08 - 01:09:42:17
Unknown
Okay.

01:09:42:19 - 01:09:54:26
Unknown
Yeah. Why don't you have your arrow there? There you go.

01:09:54:28 - 01:10:33:22
Unknown
No. You get it under Gemini, not under no. No banana. Probably.

01:10:33:25 - 01:10:38:00
Unknown
Thank you. Boss.

01:10:38:03 - 01:10:53:29
Unknown
That's still thinking. It's right. All right, all right. It's getting somewhere.

01:10:54:01 - 01:10:58:02
Unknown
So what's the next step? Look at the next step.

01:10:58:05 - 01:11:14:24
Unknown
Would you like me to generate Python code using this?

01:11:14:27 - 01:11:55:28
Unknown
And it says this loading nano banana.

01:11:56:01 - 01:12:05:25
Unknown
Thank you. Carsten.

01:12:05:28 - 01:12:12:17
Unknown
We are so grateful.

01:12:12:19 - 01:12:37:07
Unknown
I think it's, Shall we say, an oversimplification? The first. The first, part is whack a mole with a hammer animal. The next is a globe. And what looks like a rock. Then. Then the globe has things sticking out of it. Then it has a process which means nothing. And then it has a bar chart, which means nothing.

01:12:37:09 - 01:12:54:07
Unknown
Okay, y'all. Look, one has stumped that whole banana.

01:12:54:09 - 01:13:03:18
Unknown
Yeah.

01:13:03:20 - 01:13:21:06
Unknown
Always a pleasure, boss. Always. Yes. You too.

01:13:21:09 - 01:13:21:14
Unknown
Bye.