I have not been able to find any information about trusted computing on the Internet. I am reasonably sure that its software manifestation can be bypassed by a Linux live disk, but what about hardware versions?
However, an important principle of security "The system as a whole is only as strong as its weakest link." So how TPM is bypassed really depends on how TPM is being used. A good example is ChromeOS using TPM to manage encryption keys. However, the user's encryption key is based on your password, which could be obtained with a Hardware Keylogger, Phising, Rubber Hose Cryptanalysis, or other means.
Although TPM is probably a good solution for key storage, there are many ways to obtaining secrets.
There are many ways of using trusted computing. Depending on use would depend on what attacks apply, so far I know of TPM reset attacks which find some way to glitch the TPM chip to reset, sometimes requiring physical access to the processor: http://www.cs.dartmouth.edu/~pkilab/sparks/
Also there is was a talk at Blackhat DC in 2009 about a method to bypass Intel® Trusted Execution Technology which is supposed to enforce a boot to a known trusted system: http://invisiblethingslab.com/resources/bh09dc/Attacking%20Intel%20TXT%20-%20slides.pdf
As pointed out in the other comment, it depends on how it is used, by the "scourge of freedom, on the Internet" I assume you mean UEFI SecureBoot the stuff Microsoft has mandated for installing Windows 8 on ARM processors? If so this is a newish implementation and I don't think there are any specific attacks against it yet.
TPM stuff was never designed for the basic consumer to use or OEMs it was for admins to help secure their corporate networks by being able to verify the system that is running is a known system before releasing encryption keys.