Jeremy Kun

> A further pass, e.g. polynomial-to-mod-arith should decompose the RNS polynomials into the individual limbs. It seems reasonable to define mod_arith ops to work on rns types as if they...

Agreed, this would be awesome 💯

Something like this may be necessary as dialect conversion is changing upstream, cf. https://github.com/google/heir/pull/2267

tbh seems like a lower priority right now since we are focusing on ML which is all dense, but still could be interesting to see what the potential for speedup...

> webmention.io is a service to receive (not send) webmentions By "send" I mean I am posting to the webmention.io API when it is discovered on the target website. I...

Looks like that fails. I tried http://php.microformats.io/?url=https://www.jeremykun.com/fhe-in-production/ which produces https://gist.github.com/j2kun/759804379117a19c8375c6ef580febae I try to post a webmention to one of those links and get but webmention.io reports ``` { "status": "no_link_found",...

The inner-most fraction has degree 1, and the product has `n` such degree-1 fractions. So each of the products has degree `n`. Then you sum `n+1` of those degree-`n` terms,...

I just started reading the paper, so no notes yet, but I noticed they claim in 1.3 that the bicyclic encoding method doesn't support power-of-two cyclotomics, and I think that's...

@lawrencekhlim do you have a public reference implementation?

Yeah this is not relevant for most cases since most back ends have ct-pt ops that can be matched at secret-to-scheme, after which point the type system includes sk/pk differentiation...