[libre-riscv-dev] AI Accelerator

Immanuel, Yehowshua U yimmanuel3 at gatech.edu
Thu Feb 6 01:28:30 GMT 2020

> ethical is not about *looking* good.  putting it crudely: falling over
> backwards at the first hint of cash is going to result in people taking a
> dim view of the promise to be ethical, as stated very clearly in the
> charter.

I agree. We should not be moved by money as that could lead to compromise.
I personally think that NLP and CNN acceleration don’t carry any ethical repercussions
in their current state beside perhaps training a CNN to discriminate against people you
don’t like so on and so forth - but this applies to everything else software.

Nevertheless - it is good to be rigorous about what we add to LibreSOC to avoid feature creep,
So I have started a new thread and this page here: https://libre-riscv.org/future_feature_proposals/


More information about the libre-riscv-dev mailing list