I was at Oberlin College in 1965 when they got an IBM 360 "Cadet" which took up the entire basement of the Physics Building and had its own A/C. The manufacturer's rep said that "Cadet" stood for "Can't add, doesn't even try." Any smart-phone today can do much more.
We programmed the thing in Fortran, not so different a feel from BASIC. What felt really different is that each line of the program had to be typed into a card punch machine which would then output a deck of punched cards, one for each instruction. We had to carry that to the machine's card reader and insert it AFTER the compiler deck that told the machine how to interpret Fortran into machine language. If the compile failed, we would be told only the number of the first invalid instruction. After replacing that card, we would try again.
When our program finally compiled, we got an "object deck" from the output side of the machine's card reader-punch (which was the size of a nice bathroom cabinet, as I recall). We would then carry that back to the input side and feed it into the machine. Output, if any, was on a printer across the floor, about the same size as the card reader-punch. The output printer typed on green-and-white striped paper a couple of feet wide with a row of sprocket holes down the side that was conveniently perforated so you could tear it off. Storage was on massive magnetic tape drives along the back wall that stood at least six feet tall. Data access was slow, since the tape had to be searched by spooling.
There was no connection to anything outside the room other than through power cables. We used slide rules for ordinary calculations, since electronic calculators of the era were large, expensive, noisy, and required a power cord plugged into the wall. Looking stuff up meant walking across campus to the library, where actual physical books had to be retrieved and indices consulted to find page numbers with the data we needed.
Fortunately, climate change and dark matter were not of immediate concern.