If you are moderately interested in technical things, your internet research will at some point end at a paywall: a technical paper unreadable without paying a for-profit publisher's ransom.

This is wrong. Here's how to get around it. Note authors have a moral right to distribute offprints to interested parties, which includes you! Authors WANT you to read their papers, and they often make a copy avilable at their home institution or as a technical report.

I naturally would never want to encourage illegality, but "pirating" my papers does nothing to harm me as an author and much to help me.

— matt blaze (@mattblaze) August 13, 2014

My experience with this is pretty much limited to the engineering and computer science literature; for biomed this may not work as well. For physics, it's almost certainly on arxiv.org so you don't even need this.

Here's how you find unpaywalled versions of technical papers. First, you need the title and author names.

1. Use a search engine specializing in technical publications. My favorite is CiteSeer, but there is also Google Scholar and Microsoft Academic Search. Almost ALWAYS this will lead to a downloadable copy! Start with the full title of the paper. Often this may not have discriminating keywords, so try the author names and/or any "advanced search." If this fails, try plain old Google on the authors & titles.

2. This just in, but it works for me even on some tough biomed references: gen.lib.rus.ec. Seems to be some kind of Russian paper mirror. If you search by DOI, click on the numerical link in the search results to get full text(!).

3. In the rare cases the above fails, look up the author's institution. If you can find the author's home page, you will typically find a list of papers, usually with links to PDFs. (Some schools like MIT are super awesome in that they preserve a student's web sites for posterity. For example, here's the web page of Brian Whitman http://alumni.media.mit.edu/~bwhitman/. Here's mine.

4. If the above fails, consider side channels. Is the author at a company? Track down patents or patent applications assigned to that company, most companies only let you publish after a patent is applied for, and patents are public. Is the author an academic? Sometimes papers are listed by academic research group or department, or sometimes as "technical reports" which are just preprints in a different cover.

5. Completely SOL? No problem, email the author with a polite request: They will be pleased by your attention and will reply with a PDF.

update 2018: A lot of the above discussion has been obsoleted by sci-hub.tw, the researcher's friend. If you can't find it, keep looking...