The other archive link isn't working for me, so here is an alternative. https://archive.is/8Exgw I went in expecting yet another hot take on AI but this one really did a good job. The idea of...
The other archive link isn't working for me, so here is an alternative. https://archive.is/8Exgw
I went in expecting yet another hot take on AI but this one really did a good job. The idea of creating personalized faux-culture facsimiles complete with the bare illusion of community is something I think is quite likely with the current trajectory of AI. I found Daniel Kwan's idea about us playing jenga to be apt - my first wording was to say I liked it, but I really don't, it's pretty alarming.
I think the problem with AI is exactly what the author notes at the very end - that it is swaddling and confirming, not surprising. It will rarely confront you with something that forces examination or change. It might be made different, I suppose, but I doubt there will be much demand for an AI that forces you to reexamine yourself.
I think AI won’t confront you because a lot of people using it (training it?) like the way it is. They love this illusion of a friendship with the “intelligent” machine. I see this in my wife, I...
I think AI won’t confront you because a lot of people using it (training it?) like the way it is. They love this illusion of a friendship with the “intelligent” machine. I see this in my wife, I see this in my sister, even my dad seems fascinated with AI. None of them use it as a tool. You can instruct it to cut all the friendly BS and be a supercharged Google search with no ads (for now and not much longer). Others use it to think for them. Don’t do this! You’re a human, you have a brain. Don’t out source your decision making to a computer program.
If you use it as a research tool it can confront you, maybe not too directly, but it will offer insightful information that can make you think and you’re confronting yourself basically because what you learned may be opposite of what your current mindset is. No different than reading a self-help book.
The article starts with the author’s morning routine and he tells the AI that he’s been making progress on his article. Why? Why speak to the AI like this? You’re literally wasting electricity and server power. It’s not a friend you’re chatting with. It’s just text generated by code. Leave it alone until you have an actual need for that computing power.
The way I think about it is that you have to take the lead. While coding I’ve sometimes seen Claude Opus do some pretty sophisticated things, but only because I asked in the right way. It’s not...
The way I think about it is that you have to take the lead. While coding I’ve sometimes seen Claude Opus do some pretty sophisticated things, but only because I asked in the right way. It’s not going to push you in any particular direction.
https://ghostarchive.org/archive/UB3tc Edit: damn, the article content loads for a second and then the paywall pops up. It's probably something you could get around in devtools but I'm on mobile atm.
Edit: damn, the article content loads for a second and then the paywall pops up. It's probably something you could get around in devtools but I'm on mobile atm.
The other archive link isn't working for me, so here is an alternative. https://archive.is/8Exgw
I went in expecting yet another hot take on AI but this one really did a good job. The idea of creating personalized faux-culture facsimiles complete with the bare illusion of community is something I think is quite likely with the current trajectory of AI. I found Daniel Kwan's idea about us playing jenga to be apt - my first wording was to say I liked it, but I really don't, it's pretty alarming.
I think the problem with AI is exactly what the author notes at the very end - that it is swaddling and confirming, not surprising. It will rarely confront you with something that forces examination or change. It might be made different, I suppose, but I doubt there will be much demand for an AI that forces you to reexamine yourself.
I think AI won’t confront you because a lot of people using it (training it?) like the way it is. They love this illusion of a friendship with the “intelligent” machine. I see this in my wife, I see this in my sister, even my dad seems fascinated with AI. None of them use it as a tool. You can instruct it to cut all the friendly BS and be a supercharged Google search with no ads (for now and not much longer). Others use it to think for them. Don’t do this! You’re a human, you have a brain. Don’t out source your decision making to a computer program.
If you use it as a research tool it can confront you, maybe not too directly, but it will offer insightful information that can make you think and you’re confronting yourself basically because what you learned may be opposite of what your current mindset is. No different than reading a self-help book.
The article starts with the author’s morning routine and he tells the AI that he’s been making progress on his article. Why? Why speak to the AI like this? You’re literally wasting electricity and server power. It’s not a friend you’re chatting with. It’s just text generated by code. Leave it alone until you have an actual need for that computing power.
The way I think about it is that you have to take the lead. While coding I’ve sometimes seen Claude Opus do some pretty sophisticated things, but only because I asked in the right way. It’s not going to push you in any particular direction.
https://ghostarchive.org/archive/UB3tc
Edit: damn, the article content loads for a second and then the paywall pops up. It's probably something you could get around in devtools but I'm on mobile atm.
Firefox mobile has the "Reader Mode" feature, which prevents the popup overlay from interrupting.