Originally sent this content as an email about SNARKs to pmetzger's cryptography mailing list, included here for safekeeping and dissemination.

SNARKs in general:
https://github.com/scipr-lab/libsnark
https://github.com/Zerocash/libzerocash
https://github.com/pepper-project/tinyram
http://people.xiph.org/~greg/simple_verifyable_execution.txt
http://www.pepper-project.org/
http://diyhpl.us/~bryan/papers2/bitcoin/snarks/

Typed presentations on the topic:
http://diyhpl.us/wiki/transcripts/simons-institute/a-wishlist-for-verifiable-computation/
.. last one has good video, https://www.youtube.com/watch?v=Z4jzA6ts2j4
http://diyhpl.us/wiki/transcripts/simons-institute/snarks-and-their-practical-applications/
http://diyhpl.us/wiki/transcripts/mit-bitcoin-expo-2015/zerocash-and-zero-knowledge-succint-arguments-of-knowledge-libsnark/
http://diyhpl.us/wiki/transcripts/scalingbitcoin/snarks/

Some history of probabilistically checkable proofs (PCPs):
http://diyhpl.us/~bryan/papers2/bitcoin/snarks/pcp/pcp-history.pdf
http://diyhpl.us/wiki/transcripts/simons-institute/zero-knowledge-probabilistic-proof-systems/

Quadratic arithmetic/span programs (non-interactive zero-knowledge proofs without probabilistically checkable proofs) (GGPR):
https://eprint.iacr.org/2012/215.pdf

Recently I gave a presentation with a very high level and general overview of ways to (ab)use SNARKs for bitcoin scalability reasons:
(start at page 43) http://diyhpl.us/~bryan/irc/bitcoin/scalingbitcoin-review.pdf

Some discussion about which CPU architecture to use for a SNARKs prover, whether to use RISC-V or moxie or some other CPU design:
http://gnusha.org/bitcoin-wizards/2015-09-29.log

Still no word on SNARKs with trustless setup.