Introducing: LetzAI V4.
We're excited to release the newest version of LetzAI with Video Models, a new Image Editor, a new Upscaler, an improved LetzAI Chat, and a better UI.
All at $0.01 per high-quality image.
Create what you love.
The Seedance 2 API terms got me thinking a lot.
In short: API customers and platforms are required to upload signed contracts from a character/person's image rights holder in order to unlock their likeness. You also cannot resell or wrap the API.
When we launched LetzAI almost 3 years ago, we built our platform on this exact assumption: that people would care about their IP rights, and would eventually want to monetize them in AI. We invented model tagging because of that and were the first widely available platform to have such an @ tagging system in place.
However, reality for the past years in gen AI has been very different: many tool, many memes, and a technology that matured so quickly that it was extremely hard to control.
But most of all: Together, the world discovered the power of a new medium. We unlocked a new culture and created a transformative new era of art.
Seeing this rapid development, with Letz, we pivoted away from our initial hardliner idea, loosened our stance on the rights question and played along of what foundation model providers offered.
I ended up giving public speeches about this, explaining my thought process. I talked and wrote about why brands would be highly incentivized to join AI systems and become available in AI for marketing purposes. "If AI doesn't know who you are, it cannot create information about you". With Text, we now call this "GEO", and brands are already competing to rank.
My speeches, however, were not about text, but about audiovisual media. The title of that talk was "AI as a medium".
Today, I believe: AI should give everyone the means to be creative in the ways they want.
Create what you love.
It's in the distribution and advertising where we need to enforce contracts, not in the production. It has been like that with any medium, forever.
I wrote an article on how this could work with AI last year, called "Token Name System" (TNS), where I started to describe how a protocol could look like that embeds these mechanics, and in which people could secure their token names, just like they can secure their DNS domains. I see early suggestions like this enter the markets, like .agent domains.
I still believe an open protocol is needed in order to facilitate communication among gen AI platforms. I believe that protocol or standard needs to come from the big companies, like Bytedance and Google, similar to MCP.
I do not believe that we should reinvent the wheel and build on a 'censor-first' mentality.
The idea of gatekeeping models through complex and expensive contractual agreements will age badly and will be looked back at just like we look back at Red Flag Laws from early 20th century.
That's why, under the current conditions, we will not proceed with our Seedance 2 integration.
Free the models. Free the technology. Increase price per generation and let the market decide, if GPU costs are the concern.
But, you will not go in there with hopes of winning over the crowd the first time you prompt. To do so would be arrogant! And arrogance is for fools, not warriors! So you will continue with the sole purpose of improving your skills.
Most AI haters I read on X are unable to distinguish between one-prompt slop outputs and composed AI work.
This outs them as people who don't understand the medium they are criticizing and furthermore, as an archetype that got historically proven wrong.
They are the same archetypes that said photography was "just pressing a button" in the 1800s, CGI was "cheating" in the 1980s and now that AI content is "just prompts."
The ethical question around training data etc. is a separate topic, but when it comes to the craft question, most AI haters are missing the plot.
AI does not make uncreative people creative.
To use Nano Banana 2, simply select it from the new Dropdown Menu when generating images.
Image Edit uses Nano Banana 2 by default at 2k resolution. More resolutions and improvements will be added soon.
i dont post much – but when i post, its stuff like a video of otters presenting LetzAI V4, made with LetzAI V4.
it took me a total of 15 minutes to make this, pls dont be too harsh 🦦
can't wait to see what people with actual video editing skills will create!
THE INTERVIEW
if you could step into another version of yourself, who would that be?
→ LetzAI Video is leaving Early Access and is now available to everyone
→ img + vid: @letz_ai – Cinematic Mode with VEO3 ✨
@Jake_Joseph Yes, Misch is so gen AI OG that it’s nuts. He was there before even someone like me showed up.
So so early that it’s nuts. mad respect and eye for this stuff too