> may not be algorithmically optimal, but better meet the needs of the community
Yeah, that sounds a bit like a blank check for someone to corrupt the generated output under the guise of a nebulous "It's better for our community" while scaremongering about the big bad algorithm.
She is only asked about it in the last question, and mentions a New York City politician wanting open source algorithms:
> Algorithms make decisions all through his borough, the East Bronx, including where children go to school, where police patrols are assigned, and the schedule for garbage pickup.
Which seems a bit bizarre, open source algorithms would be nice, but the tie in to 'the humanities' seems tenuous at best.
Although as someone mentioned below, they are rebranded Clevo machines, and really were designed to run Windows first. Station X is just offering a guarantee that they will run Linux properly.
You would rather have a self driving car in an undefined state, rather than having it shut down? A random glitch could be just as bad as an exploit; if some chunk of memory gets overwritten and your car decides that the brick wall doesn't actually exist any more, I don't think whether it was an an exploit or not really matters. The occupants end up injured either way.
Because there is latency and overhead involved when shuffling things to the GPU, and a lot of the focus on POWER these days is on OpenCAPI/NVLink. If there is an ARM chip fast enough to make OpenCAPI support worthwhile, I would be interested to hear about it.
After some more looking around, talking to someone on reddit.com/r/openpower, and looking at the new site information; the thought is that these are 50 mm × 50 mm Sforza sockets and the 4-core CPUs are 12-cores with 8 cores disabled.
http://www.civilbeat.org/2018/01/hawaii-distributed-phony-im...
might work.