[libre-riscv-dev] AI Accelerator

Luke Kenneth Casson Leighton lkcl at lkcl.net
Wed Feb 5 13:40:30 GMT 2020

btw, immanuel: these are tough but not hypothetical / rhetorical
questions.  if there's a satisfactory answer then i'll back it 100%.
the discussion however has to take place transparently (obviously)
however more than that: if we enter into this area *without* having
satisfactory answers (publicly available for review) then we're
_really_ screwed.

one potential avenue of discussion: ways in _which_ proceeding with an
AI accelerator may not just be the "leader / innovator showing
proprietary companies that it can indeed be done in a libre/open
fashion" (because due to the Libre licensing that HDL can simply be
taken and deployed for [unethical] proprietary purposes) that *if* you
have a libre AI accelerator it stymies [unethical] proprietary AI
development in field X, or may be used to mitigate or nullify use of
AI in surveillance or other [unethical] field Y.

*then* we have an excuse / good reason to proceed.


More information about the libre-riscv-dev mailing list