I heard about LaTeX for the first time around 1993, when I was working with Prof. Diana Farkas on a manuscript we published in Acta Materialia. I fed her with my text contributions, just a few, since I was then just a master of sciences student. I watched her insert my text in the tex file and compile the source to produce the manuscript. I found that strange, since I was used with a text editor from an unknown company called Microsoft which was introduced just one or two years earlier in our university (we were late).
Then I forgot about that, I wrote my master dissertation using word (and suffering from compatibility problems related to fonts when I printed the text, there was no pdf at that time) and went to Düsseldorf to work in my doctoral thesis. Nobody used LaTeX there, so I wrote my thesis in Word too. I learned how to use the program, so it was less painful to print the text. And , of course, pdf already existed at that time.
I continued working with Word when I returned to Brazil in 1998, for my post-doc at USP. All my connections worked with word, so it seemed good to keep on using the program to write my manuscripts. Then I wrote a paper with Prof. Ryoichi Kikuchi, full of equations. When I received the proofs I noticed that many equations had strange typing errors which I didn't produce in the original manuscript. This was registered in my memory, I corrected the errors returning the proofs and kept using Word.
Some time later I decided to start using LaTeX, I must confess, it was pedantry, an ideologically motivated decision (I support free software since I first read about this in the heroic 1990' s). I believed I was not able to work with a tex file, so I started with LyX. Then the late Prof. Ibrahim (Himo) Ansara came to visit us. He saw me using LyX and told me he was a LaTeX adept. I asked what he used, and he said he edited the raw tex file, because LyX introduces many useless lines in the source. I thought "Is this possible"?
With the time, I discovered he was right. There is no problem using LyX to produce a letter, but if you try to produce a manuscript to be sent to a magazine you will end up with a lot of garbage in the preamble. By the way, soon after that, I discovered what were those errors in the proofs I mentioned before: missing backslashes in a LaTeX math formula. The publisher took my word file and converted it to LaTeX! Finally I started working with the tex file and I still do it today. In the present days I select magazines for my manuscripts based on whether they accept tex files or not, I primarily choose those who accept and use the "word-colonized" magazines only in a last case scenario.
I told all this only to show that the transition from the editor which carries its ultimate usefulness in the name (it is designed to be used in an office) to LaTeX is not easy, but can be done. As I told, I was ideologically motivated and took more than 10 years to make the transition. What bothers me is that my colleagues don' t even try it.
LaTeX is the better option to produce academic texts. Its ability to produce nice quality math equations cannot be even matched by any other text editor. The fact that you can edit the tex file in the raw version (it is ascii coded!) is also an advantage. Everybody who ever needed to insert an extra line in a matrix will agree with me that this is better performed editing the tex file rather than using the infamous "equation editor" of winword. Combined with gnuplot, LaTeX produces a manuscript with a professional look which allows you to publish your own manuscripts
if you want (and now, with ResearchGate, you can). My book was entirely typeset in LaTeX, in the final form, by me. The publisher had only to send it to the printing machines.
You don' t need to be a hardcore LaTeX user, though. Today there are plentiful tools to help generate tex files (and which does not generate garbage in the preamble). I use Kile in a linux system. There are many useful resources, like autocompletion of commands and a previewer of equations which help a lot in finding mistakes in very complex formulas. I also used the Latex Editor when I have to work in a Windows system. I'm not implying these are the best tools, they are only the ones I use.
So, if you are a scientist, give LaTeX a try. It does not hurt, and you will enjoy the result.