@jmoiron I wonder how pmem.io is going to change this picture https://t.co/pWtf0ucxf6
jmoiron plays the blues
@jmoiron @andybons @chimeracoder Yes, it's flouride. Also FOMO wrt awesome NYC gophers hanging out when I'm not around...
@jmoiron Trying to install a #python package that uses Numba (and therefore conda) is a really great advertisement for #Golang.
@jmoiron If a service isn't simple, it won't be predictable; if it's not predictable, it won't be reliable; if it's not reliable, it will be replaced
@jmoiron @bryanl A far more prevalent anti-pattern is over-designing things.
@jmoiron @jasonhevans @jimbartus yup! we worked together in like ... 2008, 2009?
@jmoiron #Golang's strength: core refuses to sacrifice implementation ergonomics for lang ergonomics. Community Generics rfcs ignore this balance.
@jmoiron @jimbartus How did you get access to our internal design documents?
created etod, epoch to date
@jmoiron @francesc @jessfraz I'll take my NY lifestyle + HK hikes :) https://t.co/76VmWYClRt
@jmoiron The answer is _always_ "An ORM was doing something unexpected without my knowledge." https://t.co/Ol12OER5w1
@jmoiron @jessfraz @FiloSottile @francesc Belotta <3
@jmoiron @chimeracoder not this year :(
@jmoiron @chimeracoder I think you might have just won #gophercon bingo?
@jmoiron A thing you know and a notification you will always approve without reading.
@jmoiron https://t.co/Dm2jMN4Ku3 The #Golang version, "Is Go too slow for webscale?", fits in a tweet: No.
@jmoiron software thought leader howto: downplay programming as much as possible, elevate soft skills, throw shade on people who code for fun
@jmoiron RT @ladykan: Hello old friend! 💁🏻I visited him 9 years ago in China and saw him again today in New York! #terracottaarmy #兵馬俑 #ageofempires…
@jmoiron @chimeracoder It's still a stack language masquerading as a serialization format so it still has all the same basic issues.
@jmoiron @amontalenti All these years the of the py3k shame campaign; barely anyone would still be using py2 if py3 had good CPU parallelism.
@jmoiron @amontalenti That or pandas data frames. It's frustrating how it's a ton of hassle & care to end up with something brittle & over-engineered.
@jmoiron @amontalenti I've been contemplating trying shared mmap regions + np.array to do proper sharing across processes, but nobody seems interested.
@jmoiron py3's concurrent.futures still uses multiprocessing, which means it still uses pickle, which means it's still an outage in wait