Skip to content
Enhancing Image Generation with Imagen 3 episode artwork

Nov 9, 2024 7 min

Enhancing Image Generation with Imagen 3

from Hard Fork AI

Google Gemini has made major upgrades to its AI model, including the introduction of Gems, Assistants, and Imagen 3. Gems are custom AI experts that can be used for various tasks. Gemini is known for its enterprise use cases and developer-friendly features. Imagen 3 enhances image generation capabilities, although there are restrictions to address concerns about deep fakes. Google has implemented Synth ID watermarking technology to verify the authenticity of generated images. Gemini faces competition from similar products by OpenAI, Microsoft, Meta, and Hugging Face. While Google's image generator is unique, open source models used by competitors may offer similar capabilities. Get on the AI Box Waitlist: ⁠⁠⁠https://AIBox.ai/⁠⁠⁠ Join my AI Hustle Community: ⁠https://www.skool.com/aihustle/about⁠

Episode details come from the podcast's RSS feed.

Embed this episode

Ready to play

Enhancing Image Generation with Imagen 3

0:00 7:51
of matches

Transcript Auto-generated, may contain errors

When it's time to scale your business, it's time for Shopify. Get everything you need to grow the way you want, like all the way. Stack more sales with the best converting checkout on the planet. Track your jichings from every channel, right in one spot.

And turn real time reporting into big time opportunities. Take your business to a whole new level. Switch to Shopify. Start your free trial today.

We have some major news out of Google. Gemini, their AI model has gotten a whole bunch of really powerful upgrades. They've added something called GEM's Assistance and of course, ImageN3. So a ton of big updates to talk about today.

I'll cover all of the podcasts. Before we get into it, I wanted to let you know. If you have ever thought of starting a podcast or you have a podcast and you grow it bigger, you need to take my podcasting course. I've recently launched a podcast course that takes you through every post-starting a podcast from start to finish.

Everything that I have learned in the last five years of podcasting, I have paid a ton of mistakes and most of them I would not wish for you to have to repeat. There's things that I've learned. I've spent, you know, tens of thousands of dollars on gear, equipment, building on entire podcast, studio and production team for my podcast and for a lot of the stuff that I work on. So if you have ever wanted to take a start a podcast, grow it.

I have a course that teaches you everything I learned in starting my podcast. How to grow from zero and how I got to four million downloads. I was able to raise hundreds of thousands of dollars, grab thousands of customers for my businesses. Everything that I've been able to do with podcasts and just think it's one of the best mediums for growth, marketing and setting yourself and your network apart.

So if you're interested, there's a link in the description. Normally it is $300. But for this week, I have set the price to be 50% off with the discount code AI chat. So there's a link in the description where you can use that.

So if you've ever thought of starting a podcast, I would grab that link and join the course. I think that this will be something that is amazing and really helps you in that regard. So let's get into what Google Gemini is doing. The first thing I wanted to bring up here, with everything that they've announced, they had a whole Tuesday announcements on some really new different updates that they've been doing, Logan, Kilpatrick over on X.

He's with their developer relation team. And he's been tweeting out a bunch of really amazing updates. The first thing I wanted to talk about is something that they've introduced called gems. So comes over on Google Gemini, essentially is one of their latest innovations that I think is quite interesting.

Essentially it's custom AI experts or agents, I guess. So you can have these pre-made gems, essentially chat GPTs, their GPTs. You have desk researcher, code reviewer, French tutor. So they call them gems.

They're essentially experts, but it's a lot to me. It's a lot like the GPTs that opening I has. And you can essentially start with these pre-built gems or these pre-built uses and use them to kind of start working on things right away. You can search for them and find them, or you can just go and scroll around and kind of get some inspiration on what you would want to use chat GPT or, I guess sorry, Gemini for.

Now, also very obviously famous for their enterprise use cases. This is something that they put a big focus on, especially with Logan Kilpatrick over there working on developer relations. In some of the startups I'm looking at, many of my developers have told me that Google Gemini is one of the easiest platforms to actually work with, integrate with their API is very straightforward. And they have went through great lengths to add very, very developer-friendly features to this.

So like massive context windows, millions of tokens, free tokens, all sorts of really powerful technology that usually be very expensive for a company to run, but it's Google so they're kind of throwing everything at it, trying to make this really, really popular. So right now I think this big kind of shift towards the specific kind of AI assistance with these gems that they have is interesting. It seems like it's kind of in response to what OpenAI is doing. So I think that's very exciting.

But the other thing I think that's very exciting is their image in three. So they're upgrading their image generation capabilities with this latest version. It's currently available to all users of Gemini and it has higher quality image. You're essentially just using text prompting it, getting an image.

And they are also including human image generation, although they do have some restrictions previously. They had a lot of controversy with previous image models and generating historically inaccurate images. So if you asked it for a World War II soldier, it was going to generate a World War II German Nazi soldier, it would generate a black person and obviously that never existed back then. So they got a lot of controversy flat and went viral for a lot of reasons.

And I think because of that, they kind of pulled some of the image generation capabilities for humans. Now that seems like they're bringing it back. But they do have some restrictions. So yeah, anyways, this is very interesting.

In order to essentially address concerns about deep fakes, Google has implemented safeguards. They have something called synth ID watermark technology that is going to essentially kind of watermark. So you can know if an image is real or not the effectiveness of this, though, we don't really know. There's a lot of debates going on around this, but it'll be interesting to see how effective that all plays out.

It definitely is a crowded market right now. Like Google is not releasing this without any competition. There's been very similar products and things coming from opening eye, Microsoft, and that anthropocugant faces. They are building kind of these customizable AI chat platforms.

So I think that their new gems or their GPTs are not completely a new thing. Image N3 I think is really where they're hoping to catch up and maybe pass some of their competitors. Not everyone is doing image generation. But that being said, even companies like XAI on Twitter using on like Grock their AI image thing over on X, they don't have their own image generator, but they were able to grab an open source image generation platform and embed it into Grock and essentially use that to get image generation right in their model without having to develop it themselves.

So Google is trying to get past everyone and come up with something new and unique by having their own image generator, which is fantastic. But I just think that some of these open source ones are so good right now that I don't really know if it's actually given them that much of a competitive advantage because all of their competitors can essentially just do what XAI did and grab some of these open source models that are powerful and then they immediately have image generation inside of the text models as well. So I think that's going to be really interesting. Google is definitely making a big push.

I'll keep you updated on everything that they roll out as far as the quality goes. Overall, I think that I would not count out Gemini, especially with some of the developers that have over their Logan Kilpatrick is doing a phenomenal job. If you're interested in what's happening over at Google, I'd follow him on X to see what's going on. And again, if you are interested in starting a podcast, check out the link in the description.

I'd love to have you as a member of the course. And this week is a 50% off discount. So if you're interested in doing it, now would be the week. I hope that you all have an amazing rest of your day.

Frequently asked questions

How long is this episode of Hard Fork AI?

This episode is 7 minutes long.

When was this Hard Fork AI episode published?

This episode was published on November 9, 2024.

Is there a transcript available for this episode?

Yes, a full transcript is available for this episode. You can read the complete transcript on the episode page.

Can I download this Hard Fork AI episode?

Yes. Use the download control on the episode player to save the publisher-provided media file.