Thoughts About LLMs, LeoAI, ChatGPT and "Public" Data...

"You don't own your data." It's a hard pill to swallow, but it's true. Once you post something on the internet, it becomes part of the internet forever. Once you give your data to a platform, it becomes owned by that platform forever. If you put your birthday on Facebook, you give them the rights to use that knowledge in any way they can.

Even on HIVE. Despite owning your account, more than you own it in these centralized platforms, once you post something here, you're giving that piece of information away to the blockchain.

Anyone can access it. Anyone can use the information you give in any way they want. Anyone, even the big players of technology, can train their own AI platforms on it.

This is a truth that Hive account@taskmaster4450 helped me realize: You lose ownership of any data as soon as you give it away! A secret stops being a secret as soon as you tell another person about it.

So, if your data stops becoming your own as soon as you post it somewhere... Why post anything at all?

I think it's a trade. When you post something on Facebook/X(Twitter)/WhatsApp, you're giving such data to the platform in exchange for a service. Maybe you want as many people to hear it as you want. Maybe you want to reach a person using a convenient service. Maybe you don't care, or don't know...

Meta/Facebook is training its AI on any public data posted on its platform. They're also probably training their AI on your non-public data in secret too, though that's a bit sketchy from legal standpoint. The point is, if your posts/images/personal information reached their platform. It's a fair game on how they use it. They even force you to give them permission in their TOS. (I had to educate my whole family on this years ago.)

I used to hate the fact that AI is taking everyone's work to make them obsolete. Actually, I still hate it. I just learned to live with it.

Adobe's has realized it has no future unless it uses Photoshop to train its AI service. Many users have migrated to other applications. I hate that Adobe is using every designer's work to drive designers out of jobs. Do they have any other options, though? ...The problem is that they have no choice. As much as it legally can, Adobe wants to be able to compete in 5-10 years.

We can't go back to AI-Free world! We have two options, Adapt or Lose!

Big corporations are racing to build more powerful AI tools. They are racing to gather as much data as they legally can. They are building their machines on any data stored on their platforms, all the while protecting as much data from their platforms from other big companies.

With the directions the world is going, there'll be a need for a public databases of information... One that's not controlled by big companies. I believe HIVE has the potential to be one such database.

That's why I think LeoAI, or any machine learning initiative that doesn't serve big corporations is important. That's why I don't mind that the data I post on HIVE be used for machine learning. That's why I need LeoAI to be as good as promised or better.

Currently, LeoAI which is based on Meta's LLM Llama, is being trained on all the data on HIVE Blockchain. For now, it mostly works on the backend of inleo.io Llama is most likely already trained on Hive Blockchain's data, but LeoAI is being trained to do stuff Llama isn't able to do out-of-the-box. Including a recommendation engine for the blockchain.

We're promised that eventually, (hopefully soon,) we would be able to access the AI and prompt it just like we prompt Llama or ChatGPT.

We're promised that LeoAI will be able to use the blockchain's data to find many information about users and articles. You can use it to remind you of things you posted in the past. You can use it to give you real-time information about what's being talked about on HIVE or transactions happening on the blockchain.

If these promises came true, I believe LeoAI could replace most of the "general" usecases of big companies AI tools. It'll also be cheaper compared to ChatGPT Plus or similar services. Though, with how quickly AI technology is growing, big companies will always have the upper hand at accuracy or in specialized use cases.

As much as I fear AI, dislike the direction we're going in, or sad about people losing their jobs because of it. I think it's inevitable... So, I'll just say "Exciting times are ahead!"

What do you think?

Disclaimer

While everything I said is my personal opinion. Most of what I wrote above are things I heard from Hive account@taskmaster4450 and I happened to agree with him.