Latest / In the Loop AI / ChatGPT Images 2.0: We Tested Everything (It's Amazing)

Transcript

ChatGPT Images 2.0: We Tested Everything (It's Amazing) · In the Loop AI

0:00
Josh: Just yesterday, OpenAI released ChatGPT Images 2.0, and the model blew my mind.
0:05
Josh: I was up until two o'clock last morning playing around with it because of how powerful it is.
0:09
Josh: As I was watching Sam announce this model, he was talking about how image gen
0:13
Josh: wasn't really that important to him. He felt like they already had a good image generation model.
0:17
Josh: When he was presented with the outputs of this one, he had his holy shit moment.
0:21
Josh: It's actually really phenomenal.
0:23
Josh: And through trying it ourselves, we have uncovered that it's actually true.
0:27
Josh: I mean, we've really frequently used NanoBanana as the go-to default image generator.
0:32
Josh: But now it's getting close to being indistinguishable from reality entirely.
0:36
Josh: And we have a series of examples that we're going to show you that are probably
0:39
Josh: useful for your actual applicable life.
0:41
Josh: Things like interior design or generating comics or generating sales graphics.
0:45
Josh: I don't think there's anyone who wouldn't find a beneficial use case of an image
0:49
Josh: generator model that is as good as this one. So let's get into the actual announcement.
0:53
Josh: Let's walk through the examples. It's pretty amazing stuff.
0:56
Ejaaz: Around midday yesterday, OpenAI tweeted this very mysterious post and it goes,
1:02
Ejaaz: This is not a screenshot, which is weird because it looks like a screenshot
1:07
Ejaaz: of someone's Mac desktop,
1:09
Ejaaz: except this is completely AI generated.
1:11
Ejaaz: And this was the precursor to their official announcement, which is ChatGBT Images 2.0.
1:17
Ejaaz: It's their new image model, and it absolutely blows every other image model
1:21
Ejaaz: out of the water. And I don't mean that as an understatement.
1:24
Ejaaz: It is number one across every single image benchmark.
1:28
Ejaaz: It's beaten nano banana to and any
1:31
Ejaaz: of the chinese image gen models just completely don't weigh
1:34
Ejaaz: up so what are some of the new things here well the fidelity and
1:37
Ejaaz: quality of these images are incredibly high you're seeing a demo video here
1:41
Ejaaz: where we have a chameleon in various different positions the wording text is
1:45
Ejaaz: typically such a hard thing for image models to nail especially when the ai
1:49
Ejaaz: world it would like jumble up the letters or it wouldn't spell things correctly
1:52
Ejaaz: now we have that completely and utterly resolved.
1:55
Ejaaz: And so you can see some of these examples come to life here.
1:57
Ejaaz: For example, look at the fidelity of this image of rice.
2:01
Ejaaz: Typically, this would just look like a gobbled white mass. And now you can individually
2:05
Ejaaz: see each grain, which is pretty nuts.
2:08
Ejaaz: And then you have examples which are a little scarier, where this looks like
2:13
Ejaaz: a real screenshot of a handwritten note in someone's stylistic way,
2:16
Ejaaz: but it is very much completely AI generated.
2:19
Ejaaz: So you can imagine this could be used for various different nefarious purposes,
2:23
Ejaaz: which might get more malicious.
2:24
Ejaaz: But there's a ton of different examples and we want to get straight into it,
2:27
Ejaaz: starting with ones that we've generated ourselves. There's one around furniture, right?
2:32
Josh: Yeah, but I actually want to start with the rice one, because what you mentioned
2:35
Josh: with the rice is that it's precise enough to show the grains of rice,
2:38
Josh: but it's also precise enough to write a single word on a grain of rice.
2:42
Josh: And that fidelity is new. So what I did is I actually went to ChatGPT myself
2:46
Josh: and tried to emulate this.
2:47
Josh: And I asked it to create a piece of rice with the word GPT Image 2 generated
2:52
Josh: on it. And this was the output that I got.
2:54
Josh: Or actually, this was the first output that I got. And I spent maybe five minutes
2:58
Josh: trying to find the grain of rice. I don't think it worked.
3:01
Josh: So I asked it to draw a box around the grain of rice and it drew a box and then
3:05
Josh: actually etched it in the middle.
3:06
Josh: So there are some edge cases that don't quite work. I mean, that grain of rice
3:10
Josh: was not in the original one.
3:11
Josh: But for the ones that do work, it's pretty incredible. And you mentioned furniture.
3:14
Josh: I am currently living in an apartment that can use a little bit of extra furnishing.
3:18
Josh: This, unfortunately, is not what my apartment looks like. This is a much nicer
3:21
Josh: variant of something that I would like to aspire to.
3:24
Josh: So what I have prepared here is a reference image for ChatGPT along with the
3:29
Josh: prompt of what I would like it to do.
3:31
Josh: And that involves just doing things like...
3:33
Josh: Adding lamps and adding different furniture basically swapping out the existing
3:37
Josh: furniture that exists in this living room and moving it into a totally new vibe
3:42
Josh: and style that i think i would more likely appreciate and resemble so while
3:46
Josh: that's thinking i guess we can kind of get into some more of the interesting parts of this model.
3:52
Ejaaz: Well i have a example that i actually have ready to
3:55
Ejaaz: go here um i was kind of obsessed i don't tell anyone this i was obsessed with
3:59
Ejaaz: manga as a kid and so i was like you know what would be cool if we could turn
4:04
Ejaaz: our show you and i josh into a manga comic so i created this detail prompt i
4:11
Ejaaz: gave this beautiful photo of us but look
4:13
Josh: At those handsome guys.
4:14
Ejaaz: Look at those handsome very handsome guys um and uh i basically asked chat gpt
4:19
Ejaaz: to generate the prompt for me so i gave it a rough idea of what i wanted to
4:23
Ejaaz: create the scene as it were and it created a very detailed prompt with stylistic
4:28
Ejaaz: references details stuff that I wouldn't know because I'm not a storyboarder.
4:32
Ejaaz: I'm not a manga creator, but funnily enough, I have an AI that can do it for me.
4:35
Ejaaz: So I don't know if anyone is paying close attention to this storyline here,
4:39
Ejaaz: but if you're not, that's great because I want to show you the end output.
4:43
Ejaaz: So as you can see, very long prompt, and this is the finished result.
4:47
Ejaaz: So what you are looking at here is Josh and I,
4:53
Ejaaz: Let me explain this. Josh and I have been filming a podcast.
4:55
Ejaaz: As you can see, we've got us set up over here.
4:58
Ejaaz: But then we look out the window and there is a shadow.
5:01
Ejaaz: And we notice that it is Sam Altman, Godzilla size, coming down upon us,
5:07
Ejaaz: terrorizing New York City.
5:09
Ejaaz: This is, I'd say the time estimate is roughly five years in the future,
5:12
Ejaaz: maybe even three. Don't know how quickly AGI gets here. We grab our weapons.
5:16
Ejaaz: It is clawed. This is not a sponsored video, by the way. I just came up with
5:20
Ejaaz: this randomly. and it shoots out prompts that wrap around Sam Altman and eventually
5:25
Ejaaz: bring down GPT-5 from taking over the world.
5:29
Ejaaz: Now you know what's going on in my head, but if you just notice this,
5:32
Ejaaz: like look at the fidelity of this.
5:33
Ejaaz: This like took five seconds to create the prompt and then another two minutes
5:38
Ejaaz: to create the actual image.
5:39
Ejaaz: Look at the fidelity of this. Like the writing is all accurate.
5:43
Ejaaz: The, like this would cost like thousands and thousands of dollars and weeks,
5:46
Ejaaz: maybe months of time to actually create from scratch. and this did it in a bunch
5:51
Ejaaz: of seconds for a couple of cents like it's pretty
5:53
Josh: Impressive oh it's so good so if manga
5:56
Josh: isn't your thing we have the furniture example it's
5:58
Josh: ready to go so here i have the original that we're seeing on screen right now
6:03
Josh: this was the original living room i fed it the prompt and here is the new one
6:06
Josh: it totally maintained the integrity of the room while swapping out just a few
6:10
Josh: key pieces of furniture to change the vibe and i think it's a testament to a
6:14
Josh: practical use case that a lot of people might have,
6:16
Josh: They want to design things. They want to make things look good and maintaining
6:20
Josh: the personalized fidelity of whatever space it is.
6:23
Josh: If you have a piece of clothing, I know this works for try-ons.
6:26
Josh: It's really good at maintaining continuity throughout these images.
6:29
Josh: So I thought that was a pretty interesting thing.
6:30
Josh: If you have an apartment, if you have a closet worth full of clothes,
6:33
Josh: you can just place those clothes out, take a picture of you,
6:36
Josh: take a picture of their clothes, ask it to address you, ask it to redo your
6:39
Josh: living room, whatever it may be.
6:40
Josh: Super powerful and works fairly quick. I mean, this output took maybe a minute to generate.
6:44
Josh: And for those not sure this is actually available to all users of chat gpt i
6:50
Josh: believe very limited instances for the free users but if you have the plus plan
6:54
Josh: for 20 a month you can just go off and start creating images and it will look
6:58
Josh: just as good as this one yeah.
7:00
Ejaaz: I mean if you're a professional that has been toying around with using ai but
7:04
Ejaaz: it's just never been good enough it's always got some form of error whether
7:08
Ejaaz: minor or big now we have a tool that actually works for you so if you're a designer,
7:14
Ejaaz: if you're a floor planner,
7:15
Ejaaz: there's a bunch of other examples I'll show here. This becomes a practical tool.
7:19
Ejaaz: Like GPT Images 1 was very much a novelty and a toy. It was fun to see everyone
7:23
Ejaaz: in Studio Ghibli versions of ourselves.
7:25
Ejaaz: But now you can use this to create certain things.
7:28
Ejaaz: Now, not all use cases are good. If you're like me, I use social media to disseminate
7:34
Ejaaz: a lot of the breaking news that happens in the world of technology,
7:38
Ejaaz: AI, or whatever it might be.
7:39
Ejaaz: But you now have reached a point
7:42
Ejaaz: where we can't necessarily believe everything we see and images
7:45
Ejaaz: too from chat gpt doesn't make that any
7:48
Ejaaz: easier what you're seeing on the screen right now is not
7:52
Ejaaz: an official take or update on the bloomberg terminal that is also not my desktop
7:57
Ejaaz: monitor this is completely ai generated and you can probably tell parts of this
8:02
Ejaaz: like kind of gives it away it's a little too zoomed in unless of course you
8:06
Ejaaz: can like change the default settings in your the Bloomberg terminal,
8:09
Ejaaz: but some of these things are really good.
8:11
Ejaaz: This is exactly where this is on the Bloomberg terminal.
8:13
Ejaaz: The percentage mark isn't that large on the actual thing.
8:17
Ejaaz: It's got all the sections pretty much nailed. So you know that the model looked
8:20
Ejaaz: up official Bloomberg terminal layouts and like recreated it,
8:25
Ejaaz: but it added a completely fake kind of like bit of news.
8:28
Ejaaz: So you could change that bit of news to represent real news,
8:31
Ejaaz: but it would still be fake.
8:33
Ejaaz: So there's a lot of like avenues here for misinformation or disinformation.
8:37
Ejaaz: So like not entirely accurate, but somewhat accurate.
8:40
Ejaaz: You can imagine the kind of social media frenzies that this would create if
8:43
Ejaaz: people were to believe and buy into these things.
8:45
Ejaaz: Like imagine if you read an announcement that wasn't actually real,
8:48
Ejaaz: bought a stock, and then it realized that it was fake, and then it crashed.
8:51
Ejaaz: You could end up losing money.
8:52
Ejaaz: You could fake data. There's a lot of avenues of this go down.
8:55
Josh: Yeah, there's two points on this. One is that we're at the point now where even
8:59
Josh: if you pixel peep, it is almost indistinguishable from real life.
9:03
Josh: You can't really tell what is AI generated and what is not. And as that kind
9:07
Josh: of gap converges, I imagine it will create a lot of chaos where there's just
9:11
Josh: no way to tell what's real when these images are so good.
9:14
Josh: The second thing that I'll mention is from this model in particular,
9:18
Josh: anytime it's asked to generate a visual asset of a piece of software,
9:23
Josh: for some reason, it's exceptionally good at understanding the nuances of every
9:27
Josh: frame of every piece of software.
9:29
Josh: Last night, I had it do DaVinci Resolve, which is what I edit a lot of videos in.
9:33
Josh: I had it emulate Photoshop, and it got every single slider down to the correct pixel,
9:39
Josh: which leads me to believe that it appears as if
9:42
Josh: there was some training customization around the software
9:45
Josh: project in particular and you have to ask the second order
9:48
Josh: question why why is it so good at all the software and I guess the answer for
9:54
Josh: me is well it's probably because they want their agents to understand how to
9:57
Josh: navigate it and then eventually emulate it and then eventually replace it and
10:01
Josh: this nuanced understanding of how everything works is training for the image
10:05
Josh: generation model but also training for.
10:07
Josh: Just i mean the future of what these agents are going to look like so
10:10
Josh: there might be some hidden stuff going on behind this image generation
10:13
Josh: model as well so back to the demos in addition to these capabilities
10:17
Josh: we have another one teed up right here which is to create a premium
10:20
Josh: infographic poster another strong suit of this model is
10:23
Josh: text and how well it's able to render text that looks lifelike looks accurate
10:28
Josh: and is able to kind of create a like a storyboard if you will a poster it can
10:33
Josh: create multiple outputs what i've asked it to do here is create an editorial
10:36
Josh: infographic and this is the first time I'm actually seeing the output of this
10:39
Josh: and it seems pretty cool.
10:41
Josh: So this is for Limitless as you are familiar with and it kind of walks through our week in review.
10:47
Josh: So the things that Limitless mentions, this is the poster that serves as like
10:51
Josh: the weekly roundup, the weekly review.
10:52
Josh: It is pretty good.
10:55
Ejaaz: Is it accurate? Yeah, I'm curious.
10:58
Josh: You check the accuracy, I'll check the QR code, see if that works because word
11:01
Josh: on the street is that QR codes work pretty well. wow.
11:04
Ejaaz: I might need to replace the entire roundup newsletter josh with something like
11:11
Ejaaz: this just a quick a quick glance quick take away
11:14
Josh: You can imagine how this can kind of carry out
11:16
Josh: to other applications right it's like if we want to juice the
11:19
Josh: newsletter up a little bit we could just create a graphic with one prompt
11:22
Josh: by feeding it the context of everything we spoke about to give you
11:25
Josh: this detailed infographic this also applies to educators
11:28
Josh: and people who are teaching things it's really
11:31
Josh: easy it to make graphics on particular lessons or
11:34
Josh: mathematical equations or diagrams or anything you want visually represented
11:39
Josh: it's exceptionally good at that so i thought this demo was kind of fun it creates
11:43
Josh: the limitless we can review as a poster that's printable the qr code does not
11:47
Josh: work but i've asked it to make the qr code scannable so while it finishes that
11:51
Josh: up and we test that maybe we could go on to another example yeah.
11:53
Ejaaz: I was just going to say before we move on um the educational traditional
11:57
Ejaaz: point is a very precedent one, mainly because if you're like me,
12:02
Ejaaz: you could read as much text as you want, but sometimes a visual that summarizes
12:07
Ejaaz: everything really helps.
12:08
Ejaaz: You can now plug an entire book's worth of text into a single prop.
12:13
Ejaaz: A lot of these frontier models now have a million contexts, which is a couple
12:17
Ejaaz: of novels or many, many novels.
12:19
Ejaaz: And so if you can imagine, if you're trying to learn about something and you
12:21
Ejaaz: want the key points, you can not only ask the AI to summarize things and give
12:25
Ejaaz: you a bullet pointed list, you can get them to transform it into an illustrative
12:28
Ejaaz: poster that just you can look at in a glance before you go to bed and learn something brand new.
12:34
Ejaaz: So I can imagine this being used in science as well.
12:37
Ejaaz: Back when I had a biology degree or back when I was doing my degree,
12:41
Ejaaz: I remember we used to have these like research poster conferences and they used
12:46
Ejaaz: to be like, I don't know, A1 size.
12:48
Ejaaz: It was absolutely massive. And you would have so much condensed information
12:51
Ejaaz: there and it took me weeks to make and the fact that i now have a tool here
12:56
Ejaaz: where you can just probably plug in a bunch of papers
12:59
Ejaaz: get out extract the right information and then put it out in a
13:01
Ejaaz: very visual way just blows my mind like we are condensing a
13:04
Ejaaz: lot of frontier research and education tools like
13:08
Ejaaz: with this one simple update it's very very cool but to move on to uh one more
13:13
Ejaaz: example that we generated um one thing that's cool about images too is you can
13:18
Ejaaz: play around with one image and make it into several different aspect ratios
13:23
Ejaaz: so what we have here is an individual i don't know who this individual is,
13:27
Ejaaz: but it has generated it looking out onto the greatest city in the world,
13:30
Ejaaz: in my opinion, New York City.
13:32
Ejaaz: And it's in like a nice little sunrise or sunset. I can't tell which one is which.
13:36
Ejaaz: But as you notice, it gives us different aspect ratios of the guy.
13:39
Ejaaz: Like over here, we see him on the left.
13:41
Ejaaz: Over here, we see him from a distance back. Over here, we see a panoramic view
13:44
Ejaaz: where we can see him looking out onto what is this? This is Brooklyn Bridge.
13:48
Ejaaz: So the details of it, you know, you can see some of it is like kind of like
13:53
Ejaaz: blurred aspect ratios as well.
13:54
Ejaaz: It's just very impressive and you could start creating like
13:57
Ejaaz: storyboard sequences from this or just kind of like pitching visuals to whatever
14:01
Ejaaz: you whatever kind of like idea or concept you want to make you could use this
14:06
Ejaaz: in the product realm if you're like trying to figure out whether a model looks
14:10
Ejaaz: good advertising your product in like let's say the product was a coat in this
14:13
Ejaaz: particular way or it could just be something advertising completely different it's it's very cool so
14:17
Josh: How does this model perform so well i think
14:20
Josh: is the question and one of the novel breakthroughs that this image
14:23
Josh: model has that others don't is the detailed reasoning capabilities this is an
14:27
Josh: image generated model that will think before acting and will reason through
14:31
Josh: the steps required to get the best image output so generally it's just pure
14:35
Josh: inference you give it a context you give it input and it just spits something
14:39
Josh: out this one actually reasons through the,
14:42
Josh: i guess the reasoning of why it's doing these things and that's part
14:44
Josh: of the reason why even though you're not giving it necessarily the
14:47
Josh: best prompt it's giving you a really powerful output and i
14:50
Josh: have another fun example here of um just like more comic books
14:53
Josh: that you can make this was a single prompt and that generated like
14:56
Josh: an entire comic book with a really accurate character that's carried throughout
14:59
Josh: another fun feature is the character continuity where
15:02
Josh: you can generate a character it will be prevailing throughout all the images
15:05
Josh: and then also one last example that we have here is of
15:08
Josh: anyone who's involved in social media or just creating any sort of
15:11
Josh: marketing material i asked it to create an ad package for
15:14
Josh: a monster shop in williamsburg called sagebird and
15:17
Josh: sagebird now has a full kit of various aspect ratios to be posted on any platform
15:23
Josh: that looks photo accurate if you'll notice there's even a street sign that says
15:27
Josh: bedford avenue which is a street in williamsburg which is very funny so i think
15:31
Josh: the the fidelity the quality the capabilities of this model.
15:35
Josh: Are really endless and again the constraint is your imagination with how far
15:39
Josh: you can push this thing because it's just it's so powerful i had so much fun
15:43
Josh: using this i must have generated at least 100 images so far just in the last
15:46
Josh: 24 hours and it is like it's so fun i recommend everyone go and try it and figure
15:50
Josh: out what use cases are best for you so.
15:52
Ejaaz: A question that came to mind immediately is okay it's good but how does it compare
15:57
Ejaaz: to some of its competitors primarily nano banana 2 from Google,
16:02
Ejaaz: who has previously held the number one spot here.
16:04
Ejaaz: Now, if you look at this image over here, it's not just number one,
16:07
Ejaaz: it's number one by a far mile.
16:09
Ejaaz: I think it has like 150 point increase on Image Arena.
16:13
Ejaaz: If you don't know what this is, this is like the number one benchmark to test these image models.
16:17
Ejaaz: GPT Images 2 isn't just number one overall, it is number one across every single
16:22
Ejaaz: category that is measured within this benchmark. By a long shot.
16:28
Ejaaz: So it has a very distinctive lead. And if you're looking at this and you're
16:31
Ejaaz: saying, okay, well, whatever, people can like orient benchmarks around this.
16:35
Ejaaz: So like, we don't know if it's real.
16:36
Ejaaz: I have a direct comparison for you. So the same prompt fed into GPT Image 2 versus Dano Bonata Pro.
16:42
Ejaaz: And you can see that there is quite a lot of differences. You can see GPT Images
16:46
Ejaaz: 2 over here on the left, the lighting is much brighter, the fidelity is arguably a lot better.
16:52
Ejaaz: And as you can see, like, you know, there's more expression on her face, she's smiling.
16:56
Ejaaz: And there's a lot more things in the background. Like if you look at the plants
16:59
Ejaaz: in the back, it's way more hyper-realistic and harder to create for an image model.
17:03
Ejaaz: Now, if you look on the right, Nano Banana 2 is very good, but there's less
17:07
Ejaaz: complicated things going on behind them.
17:09
Ejaaz: The lighting is a little bit off. And you can kind of tell that
17:12
Ejaaz: I don't know, like maybe still on both sides, you can tell that they are kind
17:16
Ejaaz: of slightly AI generated.
17:17
Ejaaz: I would actually argue that images too, now that I'm looking at it for longer,
17:20
Ejaaz: looks like the glisten just seems too glistening-y.
17:24
Ejaaz: But Nanobanana 2 can get away with it because the lighting is a little less.
17:26
Ejaaz: But the point is these models are getting way, way better.
17:29
Ejaaz: And the examples keep coming, but it's not just visual things.
17:32
Ejaaz: Like social media influencers don't have to be worried here.
17:36
Ejaaz: You can start using this for very practical purposes. Now, there was this awesome example over here.
17:42
Ejaaz: Where a guy took an image of a book, right?
17:46
Ejaaz: And he said, could you generate me a barcode for this book?
17:51
Ejaaz: And he generated the barcode. And when you scan the barcode,
17:55
Ejaaz: it takes you, it's basically an embedded link.
17:57
Ejaaz: It takes you to a page where you can then purchase or buy the book.
18:02
Ejaaz: Now, this is very impressive for if you are like trying to sell a particular
18:06
Ejaaz: product, especially if it's physical, you now don't need to go through the complicated
18:09
Ejaaz: process of generating barcodes, getting it printed.
18:11
Ejaaz: You could feasibly create your own design book cover, print it out,
18:15
Ejaaz: and then wrap it around your actual product and it actually works.
18:18
Ejaaz: It works with your internal system. So I just thought this was pretty cool.
18:21
Josh: Yeah, it's amazing. The clarity. And again, I think this is a testament to the
18:26
Josh: reasoning where it can actually recent its way through and generate an accurate
18:28
Josh: barcode in a world where it previously couldn't.
18:30
Josh: So now not only can it make infographics, but it could link these dynamic elements
18:34
Josh: to real world artifacts, to a custom domain, to your book.
18:38
Josh: They're actually usable without needing to take it into Photoshop and take it
18:41
Josh: that final mile. And that's like a really cool unlock.
18:44
Ejaaz: We have another example here as well of the front page of the New York Times,
18:50
Ejaaz: which of course is entirely fabricated um or
18:53
Ejaaz: at least partially so like this isn't a real article this
18:56
Ejaaz: isn't a real image of a paper but all the information on
18:59
Ejaaz: it so if you actually dig in here and read it um all the information about open
19:04
Ejaaz: air unveiling gpt image 2 is accurate they pulled it from the blog post they
19:08
Ejaaz: didn't you didn't have to provide the blog post they independently did it it
19:11
Ejaaz: reasoned through it pulled out the most important points and then wrote it in
19:14
Ejaaz: a stylistic manner of a New York Times writer.
19:17
Ejaaz: So you can start imagining what this could do for press and media.
19:21
Ejaaz: If you are a reporter, you might be thinking, huh, so you're telling me I could
19:24
Ejaaz: just feed this the bullet points that I want it to make, and it could write
19:27
Ejaaz: it in my voice, in my DNA that I stylistically write an article for?
19:32
Ejaaz: That's amazing. You could also ask it to generate the image for you.
19:35
Ejaaz: So there's this metro approach where you're talking about the product,
19:38
Ejaaz: but then you use the product to generate an example image that you then put in.
19:41
Ejaaz: This is, of course, also generated by images too.
19:45
Ejaaz: So there's a lot of applications here. Again, I mentioned earlier,
19:48
Ejaaz: disinformation is a very real thing. So you can imagine people sharing fake
19:52
Ejaaz: news articles about things that aren't real, that might sway markets or inform
19:55
Ejaaz: people in the incorrect way, but cool nonetheless.
19:59
Josh: Yeah. And then there's more examples for people who are involved in architecture
20:02
Josh: at all, if you're doing floor plans.
20:04
Josh: I mean, this one was cool where you fed it an image of a house and then it generated a floor plan.
20:08
Josh: But the next example I think was even cooler because this was a digital rendering
20:12
Josh: of a large building that had all of the specs listed next to it.
20:17
Josh: And using that spec sheet and using that 3D rendering, it
20:20
Josh: created a fully rendered floor plan
20:23
Josh: that you can actually use and send to an architecture to
20:27
Josh: send to an architect to to actually make blueprints and build the
20:30
Josh: building i'm not sure if this is up to code i'm not an architect but
20:33
Josh: i imagine you can probably iterate your way through this with a proper architect
20:36
Josh: to get it to be compliant to get it up to spec if it's not already and train
20:40
Josh: it to do that so this there's this unbelievable unlock that happens for pretty
20:44
Josh: much any profession that's generating any sort of image all you need to do is
20:47
Josh: put a stamp on the bottom it looks like it it already stamped it with some fake stamp approval.
20:52
Josh: But I'm sure if you do this type of work, you can kind of, you can put your
20:56
Josh: own spin on and throw your own stamp on there.
20:58
Ejaaz: If any of you are architects listening to this, I encourage you to try this
21:03
Ejaaz: out because I'm actually curious whether this is accurate or if not, like how accurate is it?
21:08
Ejaaz: Because obviously like architects in training, like train for seven years at
21:12
Ejaaz: school, which is just insane.
21:13
Ejaaz: They have to understand the physics behind the buildings that they're designing.
21:18
Ejaaz: And I'm wondering, is this physically accurate?
21:20
Ejaaz: Are the estimates, like, do they make sense? Or is this completely made up and
21:25
Ejaaz: we still have a long way to go?
21:26
Ejaaz: It looks legit to me, but then I'm not an architect. So if you're listening to this, let us know.
21:31
Ejaaz: There's another cool thing here where, again, I mentioned earlier,
21:35
Ejaaz: if you are a visual learner, sometimes you just, there's too much information.
21:39
Ejaaz: You can create these posters bracketed by a particular subject and it kind of splits it up.
21:44
Ejaaz: So like with here, we have like all the things going on in AI,
21:47
Ejaaz: you got AI models and agents, robotics, semiconductors.
21:49
Ejaaz: And you just have images which explain the start to end process of creating
21:53
Ejaaz: these different things and what they actually do with a few words underneath
21:56
Ejaaz: it, which I thought was cool.
21:57
Ejaaz: And then there was this final example over here from Matt Schumer,
22:01
Ejaaz: where I can relate to this because I formerly worked at a big four consultancy
22:06
Ejaaz: and we had to create slide decks and it would take so long because you had to
22:11
Ejaaz: move things in a specific way or reformat the text.
22:14
Ejaaz: And Matt Schumer one-shotted an entire slide deck by just providing it a bunch
22:19
Ejaaz: of information and it created it in the style of Spotify by the looks of it.
22:23
Ejaaz: So very cool, loads of different applications and I can't wait for more people
22:27
Ejaaz: to actually use this for professional purposes.
22:30
Josh: Yeah, the model's awesome. And I guess the ask is to share whatever you're using it for.
22:35
Josh: Because again, like those prompts, those examples are the only limiting factors
22:38
Josh: to really what this can do.
22:39
Josh: Because it has the reasoning, because it's so capable, it has the like pixel perfect fidelity.
22:44
Josh: It's really just a matter of massaging it with prompts to get the output you
22:47
Josh: want, not really a limitation of the model anymore.
22:50
Josh: And like to Sam's point early in the episode, it seemed like it was great before.
22:55
Josh: Now this is just unbelievable.
22:56
Josh: I can't imagine going back to Nano Banana Pro knowing that this exists.
23:00
Josh: And it's just a testament again to how fast we're
23:02
Josh: going and like what the downstream implications of this may be in
23:05
Josh: the future when you can generate infinite images for
23:09
Josh: cheap that are pixel perfect and indistinguishable from reality
23:12
Josh: what type of downstream effects does that have on
23:15
Josh: every visual artifact that we interact
23:18
Josh: with on a day-to-day basis i mean there's no way you could be sure and this
23:21
Josh: this has a lot of implications that i'm not sure we're fully aware of now but
23:26
Josh: will surely become known well as we kind of navigate through this it creates
23:30
Josh: a weird dynamic that seems a little uncomfortable like i and now i have to navigate
23:34
Josh: the internet with such a strong filter to just try to parse through what's real and what's not.
23:38
Ejaaz: I'm curious whether this tool can be used to generate visuals that humans hadn't
23:44
Ejaaz: thought of before necessarily.
23:46
Ejaaz: Like as the AI becomes smarter and is trained on our prompts and largely our
23:51
Ejaaz: flaws, like, you know, you can ask an AI to generate a detailed prompt to then
23:56
Ejaaz: prompt it itself because we don't know how to prompt it itself.
23:59
Ejaaz: Like it can do the same with images where it's like, I get that Ejaz probably
24:03
Ejaaz: missed this point. And so maybe if I create this visual in this particular way,
24:07
Ejaaz: it's one that he hadn't thought of, but now it like breaks new ground for it.
24:10
Ejaaz: So I wouldn't put it past this model to like the model that we have to today
24:15
Ejaaz: to generate something, a visual artifact that will soon be kind of like groundbreaking
24:21
Ejaaz: for humans to use. Like maybe it's not a poster.
24:24
Ejaaz: Maybe it's not a slide deck. Maybe it's something completely new that we haven't
24:26
Ejaaz: seen before. Pretty exciting stuff.
24:28
Josh: Yeah. So that's ChatGPT Images 2.0, the newest and hottest image gen model in the world.
24:33
Josh: I encourage anyone to try to displace it because uh that would
24:36
Josh: be amazing if it gets better than this but it's worth trying
24:39
Josh: it's worth sharing what prompts you use that
24:41
Josh: give you some specific outputs that you may find helpful interesting
24:45
Josh: the use cases are the currency please share yours in the comment section down
24:49
Josh: below if you enjoyed this video don't forget to share it with a friend who may
24:51
Josh: also want to generate some images perhaps they're involved in social media perhaps
24:54
Josh: they just want to redesign their hypothetical apartment whatever it may be it's
24:59
Josh: fun it's worth testing it's It's worth trying to just like feel it and understand the intelligence.
25:04
Josh: But yeah, I think that's pretty much it for today's episode.
25:06
Josh: Are you there any final thoughts here?
25:07
Ejaaz: Nope. If there's one request that I have, I want to see the images that you
25:13
Ejaaz: generate and try and surprise us, try and do a use case that we haven't covered
25:17
Ejaaz: on this particular video, because I'm curious of the creative purposes around this.
25:21
Ejaaz: Our social media profiles will be linked below. DM us there.
25:24
Ejaaz: And yeah, I look forward to seeing what you have to make.
25:27
Josh: Awesome. Cool. All right. We'll see you guys in the next episode.