D_Malik comments on Open Thread: March 4 - 10 - Less Wrong
You are viewing a comment permalink. View the original post to see all comments and the full post content.
You are viewing a comment permalink. View the original post to see all comments and the full post content.
Comments (391)
I want to use computers without being exposed to the internet. I haven't been able to find a lot of practical information about this. I would appreciate the LessWrong hivemind's thoughts on this matter.
My current plan is something like have no internet access at home, record internet tasks on a todo list, go to a library once per week and do them (sample tasks include downloading fanfic reading material for the week, sending e-mails, downloading new versions of programs, checking my bank accounts, etc...). Lack of torrents is something I'll just have to live with. I'm also thinking of switching my phone to a Nokia 106, which has no internet access. There's a neat trick where you can get e-mails as text messages, which should be enough to deal with emergencies. I'll make sure to avoid careers which involve prolonged interaction with internet-capable machines, such as programming. Is teaching math safe?
I dread the day when wireless internet becomes omnipresent. It's a horrible, horrible supertimulus.
If you want to have absolutely no exposure to the iternet, this should be easy to accomplish by sabotaging the tools your computer uses to access the internet. If you're on Ubuntu, you can do this easily by going into /sbin and removing dhclient, and possibly if* and iw*. To get those back you'd need a liveCD, and to get a liveCD you'd need an internet connection.
I used that solution for a couple months and it worked well. But I often do have legitimate uses for the internet. So you could use something like LeechBlock to block unproductive websites but retain internet access. Unfortunately, LeechBlock and similar solutions are easy to circumvent.
So I built a system to flexibly manage internet access restrictions in a way that is very difficult to circumvent. The code is here, but it's hard to use without a tutorial, and I haven't written a tutorial yet - I'm planning to, but it might take me another month or two. Essentially, you have two user accounts on your computer. The one is a "controller" administrator account, while the other is a "requester" non-administrator account. Most of the time, you are in the requester account, and you do not have access to the controller account; the controller account's password is randomized and hidden from you by the system. You can issue requests to the controller account; an example of a request is "give me access to the controller account". These requests take some time to be enacted, which prevents precommitment-breaking. The controller account restricts the requester account's internet use by filtering packets and blocking certain domains. Again, this code will be hard to use until I write a tutorial/readme, which I'll probably do in a month or two.