Originally, the motivation was to get feedback and progress in research as publications procedures were lengthy. However, it has today other aspects as discussed below.
Even though we like to think we are a part of an honest community, it is an issue that some may collect ideas by reviewing papers. Today when it is such a focus on publications, it is also a problem that many are looking for ideas.
If you add your paper to arXiv, you definitely have the proof that you have the original ideas even though someone may send you on a major revision round to gain time to work. In such a case, it is the reviewers that may be a problem.
A possibly disadvantage by putting in in arXiv, is that you may help others in a well-developed stage to reach their conclusions before you and publish it. It is therefore a trade-off between revealing your ideas too early and to prove that you were first if the reviewing process becomes long.
Originally, the motivation was to get feedback and progress in research as publications procedures were lengthy. However, it has today other aspects as discussed below.
Even though we like to think we are a part of an honest community, it is an issue that some may collect ideas by reviewing papers. Today when it is such a focus on publications, it is also a problem that many are looking for ideas.
If you add your paper to arXiv, you definitely have the proof that you have the original ideas even though someone may send you on a major revision round to gain time to work. In such a case, it is the reviewers that may be a problem.
A possibly disadvantage by putting in in arXiv, is that you may help others in a well-developed stage to reach their conclusions before you and publish it. It is therefore a trade-off between revealing your ideas too early and to prove that you were first if the reviewing process becomes long.
I have noticed that some researchers in theoretical physics focus purely on "pre"-publishing papers in archives these days. They have their own communities where they get citation scores for their archive papers as well. It is fast for them to pre-publish. If they are trusted in the scientific community, they can easily publish in archives.
Answering Jonas Kristiansen Nøland 's question: yes, there are similar trends in other fields, e.g. there now is biorXiv for the life sciences: biorxiv.org
In the old days, before the web days, peoplewould send each other preprints of their papers long before they were published. In the department tgat I worked at, they had a preprint library, and we all read papers long before they were published. This accelarated the progress of research on an international scale.
Every thing was based on mutual trust. A preprint was viewed as a paper that would be published and referred to in papers in the future.
Today things are seemingly different, but not completely. First, owing to the exponentially growing number of written papers, most respectable journals take a very long time to publish papers. In fact, it looks as if the traditional system of publication is not able to provide the publication service at a reasonable pace and, often, the proper quality. Quite a few people interact scientifically through the network independently of punlication in journals. This is a service provided by several platforms, the Los Alamos archives included.
People who posted their earlier work pre-send to regular journal if they aim to recieve feedback or any usefull comments they may add later to revise their papers. On the other side, it is good to build a kind of realibilty and trust in the community of Research area.
With lots of papers locked behind pay-walls, having some version online, say on the arXiv or here, is useful to all researchers. It simply makes your work more accessible.
But I also echo some of the earlier comments. Putting papers on the arXiv allows for some informal feedback before publication. In truth, most of the feedback I have ever been given in really help with some of the literature - we all miss some papers.
Uploading to arXiv before sending to journal gives a chance for feedback. Also, if your work is accepted by a journal but not open access, your work will be visible to all and hence, increases citation.