The challenge: Build an art eCommerce store with 100 artworks and 9100 products. The deadline: 5 days of vacation from Monday to Friday. The methodology: Lean startup. The technology: DALL-E and Let’s Enhance. The platform: Society6.

That’s the thing about being an entrepreneur; we can’t sit still (always scheming); even on vacation, we set up startups in no time. Recapping the last blog posts, you know that 2022 is becoming the year of generative AI for image creation.

The explosion of models like DALL-E, Midjourney, Stable Diffusion, etc. They have opened up a world of possibilities; they have democratized the high creativity, the artistic imagination at the service of anyone, superpowers that until now only a few had access to.

Having the innate skills of the true artists we admire is no longer necessary. Instead, you must create an account and compose an excellent prompt to interact with these AI models.

After overcoming the experimentation phase and the wow! The goal of this Lean Startup exercise was to bring down the high technology until it became tangible, in the creation of a small business, a proof of concept; now it is time to implement and build with all this technological development.

My intention in this article is to demonstrate that AI creates more business possibilities than it destroys; I would dare to say infinite possibilities for new businesses that are not yet written.

From the respect I have for the illustrators and artists who are up in arms against these AI platforms, my advice is to turn this situation around; those of us who develop AI want you to produce much more with much less; that’s why we create AI tools, not to take jobs away from anyone, but to build new possibilities in a high productivity environment.

As more powerful Audio and Video generation models are released, we will all, in some way or another, be great musicians and producers, we will all be unique filmmakers, we will all be artists in some way, and we will all have access to superpowers, as I told you in 2015, the dilemma going forward is if we all have superpowers, who will beat the trivia.

Those who use Vizologi will be the best business strategists. We will not ask anything in return, no previous training, no extensive knowledge, just that you compose good prompts to take full advantage of the algorithm and invisible data; we will do everything else for you.

I met the Society6 platform when I decorated my house; I bought several paintings of different artists, little known but very talented, either Society6, Redbubble, Contrado, etc … They are print-on-demand platforms; from what we know as web 2, they integrate several different business models that characterize them as a platform.

eCommerce lets you create your store at zero cost for artists. Marketplace offers a tremendous supply both in quantity and quality and integrates a minimal social component; for them, artists, illustrators, or designers are the community that provides them with artwork and designs (the prosumers of which we have spoken so much in this blog).

They take care of everything else, product construction, printing, distribution, and customer service. It is worth noting that in the case of Society6, if you upload an artwork with the maximum resolution of 16000×16000, they automatically create 91 products with one image. Then, by creating 100 paintings at full resolution, I got 9100 different products.

About the incentive, it is ridiculous; it is a huge problem we have in the web 2 platforms, whether it is Youtube, Pinterest, Uber, or Society6; the payment or commission offered by the multilateral platforms in the age of creators on the Internet is insufficient. In any case, they have extraordinary communities working tirelessly for them; I do not know how they do it; in this case, it is a 10% commission for each sale.

In volume, you must have around 1500 published works to be able to start monetizing minimally or pay the mortgage fee; in the same case of an average Youtuber who pretends to live from the publicity generated on Youtube, the effort is enormous compared to the reward obtained.

Let’s not fool ourselves, these monopolies do not stop creating spaces of ultra competition downstream, where always wins the same upstream, and so they have let them run the SEC for years.

Let’s hope that all the heads already working on web 3 solve this massive problem, either by organizing ourselves in DAOs or any other way, with crypto or without crypto; the big problem is the reward, and it is the incentive; the second is privacy.

Not if I make a call with a 3D avatar in the metaverse; let’s solve these problems first; let’s get down to business now that we are rebuilding the Internet again.

MONDAY – Generation of images.

The most fun part of the whole process, it took no more than 1 hour; the exercise was to conceptualize an iterative prompt, then I created the following structure:

[Name of classic work of art] reinvented and reimagined by [Tim Biskup], [Action Painting], [Gond Painting]

With this composition, the painting of Women in Seville dancing together.

Then, to create the 100 works, I iterated the classic work field and left the same prompt structure in each generation. For the Siesta on the sleep swing work, the prompt was as follows:

[The Swing] reinvented and reimagined by [Tim Biskup], [Action Painting], [Gond Painting]

TUESDAY – Image labeling.

Once the works are obtained comes the labeling phase, assigning a name to each of the pieces, this part must also be creative; a good job without a good name is half-baked. Apart from being suggestive, the name must have a minimum of SEO; after all, the buyer has to reach this content somehow. The whole process took about an hour.

Turkish soldiers dancing the water dance.

Collage of the tower of humanity 1.

WEDNESDAY – Resizing of images.

The native images generated in DALL-E have a size of 1024×1024. To get to a high-quality image scaling to 16000×16000, I had to resize those images from 1024×1024 to 1000×1000 to get the desired multiple in the scaling; the whole process took no more than an hour.

Portrait of Turkish concubine.

Ethnic woman with jewelry and pearls.

THURSDAY – Image scaling.

Up to this point, in all the actions performed, I did not invest a single euro; the free tokens that DALL-E offered me to test the tool were enough, and there was no investment in the generation of the images.

To scale the images from 1000×1000 to 16000×16000, use a tool called Let’s Enhance; paying $12, the results are spectacular, with incredible quality; they also use Artificial Intelligence, which without AI, is very, very complex to do if not impossible.

The quiet man carrying precious flowers 1/4.

The rebirth of a woman on a blue background.

FRIDAY – Upload images to Society6.

The most tedious phase took about 3 hours; once you obtained the 100 images in high quality, you must upload them to the Society6 platform, including the title, the tags that label the works, and a description; I created a small script with GPT-3 to make the descriptions, then my idea was to use AI in all parts of the process of creating this micro business.

I dedicated one more hour to make the final touches and image scaling inside the products, and I linked the main works on my Pinterest board to generate a minimum of SEO. I asked my niece about the validation, and she thought it was great; then, the startup was validated ipso facto.

Black and white landscape with zebra woman.

The quiet man carrying precious flowers 4/4.

Copyright.

A tricky subject; just follow common sense; the authorship of the works belongs to me, I am the author of the prompts, and I can commercialize with these images as much as I want.

Like any other software, we don’t have to complicate the issue, the bureaucrats are reviewing these developments, and it seems that they think differently; In my opinion: The moment we assign intellectual properties to Artificial Intelligence (algorithm + data), then we will be lost.

I see AI as an evolution within a framework of software tools and data.

AI must always be at the service of the human using it, not the other way around; AI by itself does not generate anything; there must always be a person who gives it instructions.

It is just another very powerful and excellent functionality, but it must be conceptualized within the limits and parameters of this software environment.

My clients’ business strategies with generative AI in Vizologi are their property and for their exploitation; at no time should Vizologi as a brand, company, software, or me as an individual acquire any intellectual property rights over the content that my client has created.

Conclusions.

Saving time and costs.

The usual process of the Society6 analog artist, who must generate 100 artworks for his eCommerce store, could ideally take him more than one year, and without the 11 o’clock stop for coffee, paint the artwork, scan and digitize, and finally process the image and publish. However, in this article, we have compressed those 365 days into an 8-hour workday.

About the costs, it could have been zero, and the price was $12 to generate the 100 images in high quality. I insisted on obtaining that resolution to create 9100 products, from stickers to quilts. This cost is extra; I could have avoided it by getting less-resolution scaled images. So we could say this is a residual cost or close to zero.

Creativity is dead; long live creativity.

From my point of view, AI is not killing human creativity; on the contrary, it is an explosion of imagination and unprecedented possibilities; now, creativity begins and, most importantly, is accessible to anyone. Moreover, the first sensation that produces the generation of digital works is to draw it in analog, that is, to create the physical work.

Endless content on the Internet.

Today, we create approximately 2.5 quintillion bytes of data between humans and machines, not counting technologies such as IoT.

If I generated 100 images with reasonable quality in one hour, think about how we will multiply data and content with the incipient Generative Artificial Intelligence.

The more data we generate, the better the algorithm’s performance, developing a recurrent continuous and evolutionary feedback process between the same AI.

Endless business opportunities.

This proof of concept is about an eCommerce I have in the pipeline to release my first set of NFTs in OpenSea.

Self-consumption, of course, any of the 91 products that I need in my daily life will be directly customized. It makes no sense to buy a sweatshirt from Zara when I’ve already created my personalized one with a unique design. Fashion, jewelry, furniture, and many industries will change radically in the coming years.

I have another 400 works in the chamber that I will publish soon, Mediterranean art, abstract, modernist, etc.

Inverse approach, from the digital product to the physical development, and not the other way around; as I said, when you have the works in your hands, what you want is to paint that painting physically.

In short, endless possibilities; we have just started the implementation phase, and there is no doubt that those illustrators or artists who do not adapt to the change will see a significant disruption before their eyes that will make them lose their jobs.

Those who go ahead and take advantage of all the possibilities offered by these technologies will be able to generate 800 works of art a day at a cost close to zero and multiply by x20 or x1000 their current turnover; it all depends on the prism through which you look at it.

For those of you who are interested in buying any of the products I have shown in the article, here is a 15% discount https://share.society6.com/x/2tixK4

The link to the complete store, you can find it here https://society6.com/pedrotrillo