It also gets in the way of our natural progress in reading, writing, and learning languages. I mean, of course, people could use AI to teach and learn, but why bother when a simple prompt instantly solves the problem?
It’s a double edged sword, and right now, I think it poses a risk.
If you're in a foreign country and can't communicate because some servers are down, that could put you in a tough spot. If the models were on the edge, that'd change things, but as it stands, you might find yourself in situations where the future™ totally fails you. AI will be great when its use is seamless, and you don't have to worry about the how or when.
Having to rely on a machine we can't control just to interact with people is a threat model I am not okay with. It's probably going to create a rift in society for the foreseeable future, at least if the pricing for ultimate plans stays at $200.
If the past is any indication, edge models will be the norm. I recently attended a panel at the Computer History Museum where Daniela Rus spoke about her company Liqued AI and it's edge models.
Redefining what’s possible on the edge.
Liquid’s edge-native stack delivers the fastest AI with zero cloud dependencies — powered by LFM2, built with LEAP, experienced in Apollo.
It’s a double edged sword, and right now, I think it poses a risk.
If you're in a foreign country and can't communicate because some servers are down, that could put you in a tough spot. If the models were on the edge, that'd change things, but as it stands, you might find yourself in situations where the future™ totally fails you. AI will be great when its use is seamless, and you don't have to worry about the how or when.
Having to rely on a machine we can't control just to interact with people is a threat model I am not okay with. It's probably going to create a rift in society for the foreseeable future, at least if the pricing for ultimate plans stays at $200.