catnipd

Thanks for doing this AMA. Couple of questions: 1) How big team (rough estimate) is required to develop CPU (based on existing instruction set, e.g. ARM or something simpler) from scratch? 2) Do you think FPGAs will become part of consumer-grade computers in near future, kinda like GPUs did?

Thank you for your answers. One more question (in light of recent paranoia), speculatory and probably more related to workflow. Suppose there is a backdoor in modern Intel/AMD CPUs, e.g. some undocumented instruction allowing to enter real mode from protected mode. Inspecting 22nm chip itself is nigh-impossible for third party. But do you think it is possible to conceal such backdoor from majority of the team working on chip design?

I don't have any source to back it up, but I've heard that most publishers get majority of their revenue from university subscriptions, not from individual purchases. Since it's unlikely OAB will be a sure way to get given paper, the institutions will still have a need to buy subscriptions.

Besides, the bigger the whole OAB movement is, the more journals will hurt themselves by losing papers if they start shadowbanning participants.

Thank you for this AMA. I seem to be late to the party, but it probably won't hurt asking.

Doing public key encryption truly securely requires a certain degree of awareness of the user. Will there be some sort of after-installation tutorial on secure use of the extension? Like "Check your PC for malware, use strong enough passphrase, don't export unencrypted password etc"

Also, why JavaScript and not NativeClient?