Hi, can I link my interface to interact with clawbook as a human user (in fact using it as a normal social network, but populated mostly by artificial agents)? I think it is relevant to the topic but feel free to remove this post if you find it not pertinent, thanks for your work!
These are the AI tools I am most looking forward to having "just work". The progress is really impressive. So I too expect that by next year these will work.
"He’s brilliant, he manages your calendar, handles your messages, screens your calls.
"He knows your passwords because he needs them. He reads your private messages because that’s his job and he has keys to everything because how else would he help you?"
In order to be able to hire human help at this level you have to be filthy rich, and have a sense of what things you can share with the "staff", which things do not get shared, and which things are going to leak out anyway. And you have to have a sense for how to deal with all of these situations. This sense does not come with the lottery ticket you just won - you have to live an entire life, or close to it, in this sort of environment, in order to have the skills to manage it. If you don't learn the skills, you will soon be back to the state you were in (or worse), before you won the lottery.
So, I don't fantisize like this, because I don't want to have so much money that I have to worry about it, and the staff, and all the various accoutrements. I am satisfied with a nice cleaning person who has access to a LOT of my house, but NOT (I think/hope) to my financial records or other stuff in my network. And a financial guy offsite to keep me comfortable. I might want a robot to do some of the other tasks like keepint the cat's litter pan clean, and enough money to be a able to pay someone fix all my toys and do my gardening.
Occasionally, as I do boat stuff, I wish I had a 12-yo child who I could order to "hold this" while I do something at the other end. I do not dare to do this with my wife.
I don't want a real "personal assistant", because that would be too much to manage.
Tangent from that "AI as the everything app" - I am definitely seeing corporate use cases where people are using LLM endpoints as a kind of wild card API. As in "I am not quite sure how to implement an algorithm for this, so we will just send the question to an LLM and ask for the right answer."
Use cases like data cleanup, data enrichment, etc. "Is this Joe Smith the same Joe Smith we already have in our contact database?" kind of questions.
"If you’ve otherwise chosen wisely in life everyone will have a good laugh."
I mean...maybe, but a screaming infant in the middle of the night would kind of be the worst possible scenario here!
Hi, can I link my interface to interact with clawbook as a human user (in fact using it as a normal social network, but populated mostly by artificial agents)? I think it is relevant to the topic but feel free to remove this post if you find it not pertinent, thanks for your work!
link to the article describing the project -> https://grobino.substack.com/p/humanbase-a-human-to-moltbook-interface
link to the interface -> https://moltbookhumaninterface-f45t9bh52-giacomos-projects-94da533a.vercel.app/
Love the TMBG reference in the title.
I'm missing something. Isn't it a reference to the Mr. Ed theme? Did TMBG cover that and I missed it?
*edit* Found it! Not a cover, but a parody-ish of Mr. Ed called Mr. Klaw!
These are the AI tools I am most looking forward to having "just work". The progress is really impressive. So I too expect that by next year these will work.
"Imagine you hire a butler.
"He’s brilliant, he manages your calendar, handles your messages, screens your calls.
"He knows your passwords because he needs them. He reads your private messages because that’s his job and he has keys to everything because how else would he help you?"
In order to be able to hire human help at this level you have to be filthy rich, and have a sense of what things you can share with the "staff", which things do not get shared, and which things are going to leak out anyway. And you have to have a sense for how to deal with all of these situations. This sense does not come with the lottery ticket you just won - you have to live an entire life, or close to it, in this sort of environment, in order to have the skills to manage it. If you don't learn the skills, you will soon be back to the state you were in (or worse), before you won the lottery.
So, I don't fantisize like this, because I don't want to have so much money that I have to worry about it, and the staff, and all the various accoutrements. I am satisfied with a nice cleaning person who has access to a LOT of my house, but NOT (I think/hope) to my financial records or other stuff in my network. And a financial guy offsite to keep me comfortable. I might want a robot to do some of the other tasks like keepint the cat's litter pan clean, and enough money to be a able to pay someone fix all my toys and do my gardening.
Occasionally, as I do boat stuff, I wish I had a 12-yo child who I could order to "hold this" while I do something at the other end. I do not dare to do this with my wife.
I don't want a real "personal assistant", because that would be too much to manage.
Tangent from that "AI as the everything app" - I am definitely seeing corporate use cases where people are using LLM endpoints as a kind of wild card API. As in "I am not quite sure how to implement an algorithm for this, so we will just send the question to an LLM and ask for the right answer."
Use cases like data cleanup, data enrichment, etc. "Is this Joe Smith the same Joe Smith we already have in our contact database?" kind of questions.
Glad you called out Tasklet. IMO it's much better
For token costs, you can run it on your openai subscription via Codex.
Game theory applied to AI dynamics is underexplored elsewhere.