In theory the arpanet was restricted to someone with connections to ARPA (now DARPA): either the military, military contractors (typically being paid to support the net, like BBN and SRI) or arpa-funded research institutions like ISI, MIT, etc). However if you had a parent or friend with access they'd sometimes give you the phone number of a local TIP (dialing point) -- then pretty much the only place you could go was the machines at MIT tech square (AI, MC, ML, DM) because they didn't have any restrictions.
Also various people who had accounts on other machines on the net would connect to the MIT lab machines as tourists simply because it was fun.
Even once the lab caved and put in some login security it was only to access the machine -- once you were connected you could change to another user or whatever.
It was "the ARPANET" by the way -- ARPA's network rather than having a name. ARPANET was also an adjective (e.g. ARPANET TIP). I know RMS would have said that but the interviewer transcribed it without the article.
Back in the early 90's I remember reading a "Letter from a Famous Hacker" in an e-zine talking about getting access to MIT systems. The sentiments expressed in the text made it seem like the security posture of the MIT AI lab was rather laid-back (and it sounded pretty cool).
I dug up the article[1] tonight (thanks, Jason Scott!). The the perspective I have today (at the time I had no idea who RMS was) it seems pretty clear that RMS wrote it. It's a neat little time capsule.
Also various people who had accounts on other machines on the net would connect to the MIT lab machines as tourists simply because it was fun.
Even once the lab caved and put in some login security it was only to access the machine -- once you were connected you could change to another user or whatever.
It was "the ARPANET" by the way -- ARPA's network rather than having a name. ARPANET was also an adjective (e.g. ARPANET TIP). I know RMS would have said that but the interviewer transcribed it without the article.