Here is the Lion's Den #threadcast for today's Episode. It will go live at 1 pm eastern time.
Drop your questions and comments down below.
Here is the Lion's Den #threadcast for today's Episode. It will go live at 1 pm eastern time.
Drop your questions and comments down below.
Today we will be discuss how to build a database and why what we are doing on Leo is so different than what went before.
Building a database is the founding stone for so many applications nowadays. It should not be overlooked
Without a doubt. So few realize the value of it. This is the epitome of exponential.
Lion's Den Day should be spent at home in front of the computer, not out traveling. Shame on me
These start later and later at night where I live... Oh, well... Good I have nothing at that time.
Remember to support and welcome Dash and Meme communities that are coming to INLEO. They are also promoting ownership and lion values.
Couldn't access my PC, here using my mobile phone on cellular service with my second account.!LOLZ @ahmadmanga
I think LeoAI should make use of HafSQL too, it's more accurate that way.
So it's basically like the Magic Mirror in Snow White.
You're the fairest one of all, Task!
I'm not very happy about AI taking the art artists make and use it for learning... I know it's in their legal rights to do iit, but it still feels scummy
It blows my mind how LLMs maintain quality while having less parameters now.
this all being said, I think a lot of convos will go open source at the user’s discretion
i love you all, need to go back and punch the 9 to 5 clock lol
save a u n i c o r n <---- still cant type unicorn
What was the creative commons site you mentioned? It cut out at that bit on the recording
Was gonna ask what the main topic would be but I found the sub-thread
I work with a lot of ecommerce companies and they are literarily paying money to acquire better data on their potential customer base
Isnt that interesting?
It seems everyone understands how data is the new oil, except people on Hive.
They are intent on providing Big Tech with more of it.
Moreover, sometimes I help clients build models to predict something about their business and sometimes we deliberately opt for an inferior model because we just don't have the data to justify a more sophisticated one
There are the tradeoffs, for sure. You probably have to adjust the weighting.
It shows how the more data LeoAI has, the more sophisticated it can be.
Building a database besides LEOAI? Are we talking about building databases on LEO in general and every user building databases with their own content?
No we are talking about the vector database that Leo uses for LeoAI.
It is different than what we see with a traditional database.
Here is the space. Set a reminder.
The youtube-summarizer bot will go live this weekend. Its main goal is to make it more convenient for us to add more to the #database faster
Here's an example of it in action (see comments section of the parent thread)
https://inleo.io/threads/view/mightpossibly/re-taskmaster4450le-2puduwq8t?referral=mightpossibly
i was literally 2 years ahead of this stuff with Veews lol my goodness, my timing sucks at everything LOLOL
hehehe we will all feel like big shots saying stuff like "talk to my agent about it" or "have your agent talk to my agent"
I found a plugin called Smart Connections, that is great at connecting documents with similar topics. Imagine downloading HIVE posts and creatinga database of Connectionswith it!
I was going to try this today, but my internet plan ended, and can't do it with phone plan...
https://inleo.io/threads/view/ahmadmanga/re-taskmaster4450-jjmj2jtu?referral=ahmadmanga
The more I learn about AI, the more I feel people around me are in for a huge awakening by
ANY text-based documents!
https://inleo.io/threads/view/ahmadmanga/re-taskmaster4450-jjmj2jtu?referral=ahmadmanga
That's the thing! AI will be better at human feet pics than actual humans!!!
https://inleo.io/threads/view/khaleelkazi/re-taskmaster4450-mygip6mj?referral=khaleelkazi
Venice.ai manipulates all the weighting of the open-source LLMs it uses to have less censorship.
No, the results of LeoAI won't be on the blockchain unless the users posted them afterwards
Never-ending, you said that.
You’ll be able to flip on a setting - fully by user choice - to open source your queries / chats with LeoAI
The way this is achieved is that leoai will take your chats and autonomously post them onto the hive blockchain as plain text and then train based on it
Then other users will actually be able to see a feed of “Ahmad chatted with LeoAI about bitcoin” and click on it to see the conversation (picture it like sharing a chatgpt link)
All chats default to private unless the user chooses to help with training and open source their chats
It doesn’t exactly train from private prompts and conversations
Similar to how chatgpt does it
For example, if a user is prompting with specific data sets, they may not want LeoAI to remember those data sets so that another user could request it
Of course...
https://inleo.io/threads/view/winanda/re-taskmaster4450-pbykhcbs?referral=winanda
for example imagine you run a private company and prompt it with financial data
You don’t want LeoAI to be trained on your private financial data
This could actually become a selling point:
https://inleo.io/threads/view/khaleelkazi/re-taskmaster4450-8nfgyzaj?referral=khaleelkazi
LeoAI/Chatbots are definitely more appealing to the general public than whatever Web3.0 is!
Actually, LeoAI doesn't even need to have its own image generation pipeline, it could use normal Stable Diffusion inference, and most people will be finfine with that
Having any image generation tool is better for marketing than having none.
https://inleo.io/threads/view/ahmadmanga/re-taskmaster4450-35zmftm38?referral=ahmadmanga
I was partially present but it was a nice call. See you next week, I hope the space will be better too.
See you all on Threads. ✍