3 ms·
Besides for the novelty factor, I can't really see specialized personal AI devices taking off. Nearly everyone already has a phone which can handle the process
by gorbypark 2y ago
Besides for the novelty factor, I can't really see specialized personal AI devices taking off. Nearly everyone already has a phone which can handle the processing (some on device stuff should be coming "soon") and network connection, there's really no reason to have a separate device. I could see some "companion" devices becoming popular (watches, smart glasses, etc) but not really standalone devices that more or less just connect to an API to do inference.
That being said, I did almost order a Rabbit R1 this morning when I read that it came with a free year of Perplexity Pro (which is $200/yr) and the device costs $200. However, it was for the first 100,000 preorders and it's not available anymore!
- barfbagginus 2y agoI imagine an "Inference Backpack" with a generator and 2 gpu rig for some kind of heavy handed realtime AI use case, like tactical, security, gaming, AR/VR, or communications
- MichaelMug 2y agoHow does Perplexity Pro compare to others?
- gorbypark 2y agoI haven't really tried other "LLM based search" products, but Perplexity (even the free tier) is pretty handy. It's fairly good at summarizing a complex topic, and of course gives you all the sources so it's easy to click through. It confidently only failed me on one search, which was "What is the maximum RAM for an M1 Mac Studio?" and it told me 192GB (which is correct but for the M2), then I said something like "No I meant the m1 version" and it corrected itself wrongly and told me 64GB (which is partially correct, 64 for M1 Max 128 for M1 Ultra). The source page at least for the second results was a direct link to Apple's spec page for the Studio with M1 Max. I might retry it with a Pro search and see if there's any difference.