Arxiv website is very good at collecting mathematical/science work updates. However, where were mathematical/science work updates posted before arxiv website existed?
Before 1991 - nowhere, there were no platforms to "post" preprints on. Some were distributed by mail, and even collected and catalogued by large libraries. US National Institutes of Health even briefly supported Information Exchange Groups that circulated preprints, from 1961 to 1967. It was shut down because journals refused to publish shared memos, see The prehistory of biology preprints by Cobb. In 1969 this pushback was enshrined by applying the Ingelfinger rule of non-duplication to preprints. Nonetheless, the practice spread informally in some fields, especially physics. As Klein reported in a letter to Science, it was established by 1980s:
"When I was a graduate student in particle physics in the 1980’s, preprints were well established. When we prepared a manuscript for Physical Review, we would also prepare a preprint version, which we would then mail to colleagues, and to a list of the major particle physics research centers around the world. Larger institutions had instituted a numbering system, providing a modicum of citability.
We also made preprints for write-ups of conference talks. Since bound printed conference proceedings took so long to appear, this was the main channel for propagating preliminary results. Preprints were standard for all of the large experimental collaborations, and for many, if not most theoretical groups. The process had some drawbacks – it was somewhat clunky and labor intensive, and it was elitist. If you weren’t on the right mailing list, you would miss important results."
In 1990 preprints began to be shared by email, starting with Cohn. In 1991 she inspired Ginsparg to create a central repository at Los Alamos that could be accessed remotely. First by FTP, then by Gopher, and from 1993 over the Web, see arXiv-ing the History of Preprints in Physics by Pell.
I remember those days. (I am in mathematics.) When I completed a paper, I would mail (not email) photocopies (called "preprints") to researchers I knew that I hoped would be interested in it. Then I might attend a conference or visit another university, where I would give a talk on my work, and bring a stack of preprints for interested participants to take home. An abstract might be published (in the publicity about a meeting, or a separate publication Abstracts of the American Mathematical Society) and interested mathematicians would send mail to me asking for a preprint. Sometimes (if it caught on) some of those recipients would make photocopies for their own interested colleagues.
Back then, photocopying was a real expense, and researchers would pay for it from their research grants.
All this would take place during the many months I waited for the journal to decide to publish the paper, and the many more months before that issue of the journal came out in hard copy.
When the paper was published, the journal would provide me (or sell me) copies of the printed paper. These were called "reprints" or "offprints" and the whole process continues: interested mathematicians, sometimes years later, who see the paper in the journal, would mail and ask me to send a reprint to them. (Since that was cheaper than making photocopies themselves.)
I had a large file drawer where I kept reprints and preprints of papers by others on topics that I was interested in.
Depending on what you mean by "science work updates", bear in mind that this might be covered by different types of journal: shorter works might be submitted to "Rapid Communications in..." with a quick turnaround, whereas the definitive write-up would eventually appear in the more prestigious "Journal of...".
Also, many institutions would formally issue "Technical Reports" (as a printed and bound series) that could by ordered either by individual researchers or a library (for example, here is a randomly-chosen example of a progress report from UIUC in 1966). These might only be reviewed internally, and covered the role of modern preprints. NASA used to have a service - "Center for Aerospace Information" - that would let one order microfiche copies of any report prepared for them.
The obvious other way of keeping people abreast of what one was up to was simply to talk. My personal impression is that the number of small, local (national/regional) one-day and half-day meetings has significantly fallen in the past couple of decades, possibly because of difficulty getting funding in academia for attendance at anything that doesn't result in published proceedings.
(There's also the option of writing letters, as in What type of letter is he talking about and why did he write a letter to other mathematicians?, though this doesn't scale well!)