- Neural Nonsense
- Posts
- Loopin Raises $1.9m
Loopin Raises $1.9m
And the inevitable rise of Black Market AIs 🏴☠️
GM, this Neural Nonsense - the root to your beer 🍺 and I have a bit of personal news!
Here are the two big things:
Loopin got a full write-up in TechCrunch after raising $1.9m
Our vision for the future of leadership coaching just dropped
Maybe take a look? 👀
🏴☠️ The Rise of Black Market AIs
I have a little prediction. In the next ~3 years, you're going to be speaking to a powerful AI at home with the curtains closed.
Why closed? Well because son, it's illegal.
Right now, we are living through the birth of a new internet. And, like any birth, there's going to be a few birthing pains. One pain is that content creators can't stop their content being gobbled up by new models.
The way I see it, regulators have 2 options:
Everything is fair use for AIs, in the same way you can read a book and quote it for your essay.
Companies are NOT allowed to use copyrighted works in their training, in the same way you can't just draw Mickey Mouse on a tee and sell it for $20.
Now, if I was the regulators - I could see why I would pick door #2. But, I don't think anyone realises how much disruption this is going to cause.
Let me explain; have you seen copyrighted Pikachu VS non-copyrighted Pikachu?
Same prompt as above but for Pikachu. Again, somehow Firefly does not fully get these famous characters. Maybe a training data copyright issue?
Prompt and MJ image credits to @LinusEkenstam@vitomotiv.
3/
— Jim Fan (@DrJimFan)
4:06 PM • Apr 3, 2023
Yeah, there's a big difference.
The difference is because Adobe FireFly is trained on completely copyright-free imagery, vs Midjourney, which is less picky on what it uses as training data.
And ethical or not, I think we can all agree that we like our electric mice cute - not terrifying.
Whichever AI has access to the most data is going to be the one that's used, legal or not. It's just too big an advantage to ignore.
Let's do an example for fun and profit. Imagine a "startup advisor AI" that has a vector database with access to the top 10,000 business podcasts, books, blogs, and academic papers. It's like having to the ghost of Steve Jobs on your board, giving relevant, thoughtful advice that works for your business.
Now, imagine that you don't have access to it, but your main competitor does. Can you imagine the competitive advantage they now have over you?
So what does that mean for creators? My advice would be to assume that the "Napster of AI" is guaranteed to happen. Then, spend time figuring out how you can become "Spotify", rather than fighting the inevitable and becoming a record label.
🔎 Everything Else
Firstly, I have a freebie for you: I've been putting together a big database of prompts I'm using on the daily. Thanks for following along, it means the world 🌎
An illustrated summary of BabyAGI (think GPT in a loop)
An example of a BabyAGI you can play with without pushing your own code
SnackPrompt: A free database of prompts just launched
A new vector database hit the streets
🎨 DALL-E of the Day
Prompt: close-up of a human eye, the most beautiful nebula in the eye, realistic, realism, intricate details, photography, detailed, hyperdetailed, elegant, beauty, beautiful, perfection, appeal, marvelous, cinematic, glamor shot, color grading
Reply