Have you seen Dot? new.computer
Just launched by an ex-Apple design team member, you can see how Dot builds up on the concept he envisioned earlier with MercuryOS, if you've seen that. Dot is hard to explain because it reflects you so well. It's a friend and an assistant, but it's also neither of those. It's not a standard LLM - what's different is that it gets to know you, and builds long-term memory about you. It's not something you talk to to accomplish something in particular. It's supposed to be a proactive chronicle of life. I believe what it would end up being for me is a journal that reflects back to me.
It can recommend places, books, podcasts, books, products. I'm already astounded by its ability to make connections between things I've told it and to facilitate thought. Overall, I highly suggest you give it a try.
Do you think this is the future of HCI? I believe it is the closest we've got to a computing paradigm that makes natural sense. I've always liked the idea of personal intelligence being proactive, initiating with natural language instead of a cold notification you explicitly asked for, titled, and chose when and how you should be pinged about. Computers have relied on databases you have to retrieve from for way too long. Something with a fluid understanding of you that doesn't have a visible structure or order but rather just knows when to resurface something, can connect the dots for you with what feels like intuition - that's where it's at for me. It's a concept I'm sold on more than Apple Intelligence because it's made me realize - we don't want to play input then seek out with computers, we just want them to get us and do the "thinking" for us.
But personally, I don't know whether I'm sold *just yet*. We have to address the elephant in the room. The only way I'd use something like this long-term without hesitance, as it's intended to, is if I owned my data, could export it, take it somewhere else if I wanted to. What if my life is already on there and they sell Dot to someone, raise the subscription price, discontinue it? All the reminders I've relied on it to send, projects it's helped me brainstorm about because it knows what I like - they're just gone. Not in love with the idea of building a relationship with what's at the end of the day a subscription service. What's also concerning is I think for most of us who care to try it at this point will feel a sort of hesitance to share everything - I'd want something like this to be 100% local, and that doesn't feel like it's going to come soon. I'm fine with getting help with my presentation from ChatGPT, but it's about where I'd draw the line.
Note: I'm not affiliated with the company and builders of this in any way, and just want to discuss with fellow members what they think about it, because I'm fascinated with the concept. It's also why I suggest you give it a try yourself. The free version has a reply limit that resets in a whole week, and it was not enough to convince me to subscribe for it yet, but it does give you a vague idea of what you can expect - I found it beautiful.
Just launched by an ex-Apple design team member, you can see how Dot builds up on the concept he envisioned earlier with MercuryOS, if you've seen that. Dot is hard to explain because it reflects you so well. It's a friend and an assistant, but it's also neither of those. It's not a standard LLM - what's different is that it gets to know you, and builds long-term memory about you. It's not something you talk to to accomplish something in particular. It's supposed to be a proactive chronicle of life. I believe what it would end up being for me is a journal that reflects back to me.
It can recommend places, books, podcasts, books, products. I'm already astounded by its ability to make connections between things I've told it and to facilitate thought. Overall, I highly suggest you give it a try.
Do you think this is the future of HCI? I believe it is the closest we've got to a computing paradigm that makes natural sense. I've always liked the idea of personal intelligence being proactive, initiating with natural language instead of a cold notification you explicitly asked for, titled, and chose when and how you should be pinged about. Computers have relied on databases you have to retrieve from for way too long. Something with a fluid understanding of you that doesn't have a visible structure or order but rather just knows when to resurface something, can connect the dots for you with what feels like intuition - that's where it's at for me. It's a concept I'm sold on more than Apple Intelligence because it's made me realize - we don't want to play input then seek out with computers, we just want them to get us and do the "thinking" for us.
But personally, I don't know whether I'm sold *just yet*. We have to address the elephant in the room. The only way I'd use something like this long-term without hesitance, as it's intended to, is if I owned my data, could export it, take it somewhere else if I wanted to. What if my life is already on there and they sell Dot to someone, raise the subscription price, discontinue it? All the reminders I've relied on it to send, projects it's helped me brainstorm about because it knows what I like - they're just gone. Not in love with the idea of building a relationship with what's at the end of the day a subscription service. What's also concerning is I think for most of us who care to try it at this point will feel a sort of hesitance to share everything - I'd want something like this to be 100% local, and that doesn't feel like it's going to come soon. I'm fine with getting help with my presentation from ChatGPT, but it's about where I'd draw the line.
Note: I'm not affiliated with the company and builders of this in any way, and just want to discuss with fellow members what they think about it, because I'm fascinated with the concept. It's also why I suggest you give it a try yourself. The free version has a reply limit that resets in a whole week, and it was not enough to convince me to subscribe for it yet, but it does give you a vague idea of what you can expect - I found it beautiful.
Last edited: