Just a note on Rabbit. I just got mine yesterday. It's attempting to introduce a stop away from LLM to LAM (Large Action Model). The idea seems simple: why not let AI work for us. We do tasks all day long and AI can do them for us. We order food from DoorDash, we order an Uber to get to work, etc. Jesse's video is here: https://www.youtube.com/watch?v=22wlLy7hKP4
It's a compelling idea, but after 24 hours, it's not even close to being realized yet. It came with four LAMs, Uber, DoorDash, Spotify and Mindjourney. I can get none of them to work. I have either Google connected accounts or 2FA/MFA. Mindjourney I use through Discord. In fact, trying to connect Uber, it went around and around in circles thinking that I was a robot trying to log in. Even though I was answering the captcha's correctly, it still thought I was a robot.
The way this works is you go to the Rabbithole and RabbitOS LAM watches how you log in. Doing this triggers the websites robot mechanisms and I could never get in.
I'm still hopeful, we do need AI to do good and boring stuff for us.
Great piece! Putting these 4 side by side in this piece was nice to get a perspective on the current offerings and your thoughts outside of chatbot interfaces.
I should be getting me R1 soon. I was initially really excited about the 'teach' mode and had a bunch of ideas queued up, but later found out that teach mode will not be available until later this year (I think?). Crazy to think about how fascinated I was by what r1 was able to do when I saw the demo and now I think "hmm, yea cool."
Great article! I think that Apple is going to be a major player in this (big announcements coming at WWDC). They're going to help us think differently (pun not intended) about what AI actually means in our day-to-day life. Especially given their access to over 2 billion people.
Could you use an "agent" to go around (or phone around) and collect information, for example do interviews and pols with customers, voters, eccetera? Would people accept talking to a machine? The agent could be a "better" enterviewer than a human (no bias, remembers what you said the last time it iterviewed you,...)
If I had all the tools needed to build a wooden chair, I still wouldn't be able to do it.
While AI is ridiculously easy to use, I still think people need to learn how to do something without AI. At least the basics... Then use AI to improve, or get things done quicker.
I don't think it's a good idea to use it to avoid learning how to do something...
I agree Matthew. I bought mine because I saw potential in the R1. Yes there will be ridicule but remember the iPad and Watch faced a similar reception.
I was thinking it would be neat to be able to run a specific GPT on your R1
Great article. Your exploration of AI devices highlights an exciting shift toward more intuitive and integrated tech experiences. Despite their limitations, these gadgets preview a future where AI can blend into our daily lives, moving beyond screen-based interactions. This evolution promises to enhance our decision-making and creativity, opening up new possibilities for how we interact with technology. Have you seen any other aI device that excites you?
I came across this conversation with a chatbot which made me smile:
https://www.andrewfraser.nz/blog/hi-bot
:-)
Just a note on Rabbit. I just got mine yesterday. It's attempting to introduce a stop away from LLM to LAM (Large Action Model). The idea seems simple: why not let AI work for us. We do tasks all day long and AI can do them for us. We order food from DoorDash, we order an Uber to get to work, etc. Jesse's video is here: https://www.youtube.com/watch?v=22wlLy7hKP4
It's a compelling idea, but after 24 hours, it's not even close to being realized yet. It came with four LAMs, Uber, DoorDash, Spotify and Mindjourney. I can get none of them to work. I have either Google connected accounts or 2FA/MFA. Mindjourney I use through Discord. In fact, trying to connect Uber, it went around and around in circles thinking that I was a robot trying to log in. Even though I was answering the captcha's correctly, it still thought I was a robot.
The way this works is you go to the Rabbithole and RabbitOS LAM watches how you log in. Doing this triggers the websites robot mechanisms and I could never get in.
I'm still hopeful, we do need AI to do good and boring stuff for us.
Anyone manage to get the side button on their iPhone to work with ChatGPT voice? How do you do it?
You need an Iphone 15 to do so.
Thanks!
Great article. Loved your book. Keep it up!
Great piece! Putting these 4 side by side in this piece was nice to get a perspective on the current offerings and your thoughts outside of chatbot interfaces.
I should be getting me R1 soon. I was initially really excited about the 'teach' mode and had a bunch of ideas queued up, but later found out that teach mode will not be available until later this year (I think?). Crazy to think about how fascinated I was by what r1 was able to do when I saw the demo and now I think "hmm, yea cool."
How quickly we adapt, ha!
Great article! I think that Apple is going to be a major player in this (big announcements coming at WWDC). They're going to help us think differently (pun not intended) about what AI actually means in our day-to-day life. Especially given their access to over 2 billion people.
Could you use an "agent" to go around (or phone around) and collect information, for example do interviews and pols with customers, voters, eccetera? Would people accept talking to a machine? The agent could be a "better" enterviewer than a human (no bias, remembers what you said the last time it iterviewed you,...)
Actually AI amplifies the biases of the information it’s given. Like learning styles. Most LLMs will spit out that learning styles are accurate.
If I had all the tools needed to build a wooden chair, I still wouldn't be able to do it.
While AI is ridiculously easy to use, I still think people need to learn how to do something without AI. At least the basics... Then use AI to improve, or get things done quicker.
I don't think it's a good idea to use it to avoid learning how to do something...
Is it possible to get a peek at page 2 of Ethan’s LLM folder? It was really helpful to see and try the tools that he’s using!
I appreciate the pointers to self-contained and open source AI services, would hate to see the cloud subscription model dominate AI development.
Would also like to know how much AI content went into this article!
In other news
“A woman who "married" her AI husband says she's open to dating other people in the real world.”
This is legit.
Look it up.
I agree Matthew. I bought mine because I saw potential in the R1. Yes there will be ridicule but remember the iPad and Watch faced a similar reception.
I was thinking it would be neat to be able to run a specific GPT on your R1
Great article. Your exploration of AI devices highlights an exciting shift toward more intuitive and integrated tech experiences. Despite their limitations, these gadgets preview a future where AI can blend into our daily lives, moving beyond screen-based interactions. This evolution promises to enhance our decision-making and creativity, opening up new possibilities for how we interact with technology. Have you seen any other aI device that excites you?