4,949 karma · joined June 1, 2013
Currently building the best managed Postgres service at Snowflake, by way of Crunchy Data. Formerly at AWS working on (likely) the largest consensus and replicated WAL systems in the world as well as macOS virtualization just for funsies.
Twitter @bitsandhops
What's sad, but to be expected with anything involving rigor, is how many people only read the title and developed an averse reaction. The Hacker in Hacker News can be a skilled programmer who can find their way around things. Maybe now it's more of the hack as in "You're a Hack", meaning you're unskilled and produce low quality results.
It's also very real that there's a shortage of equipment for consumers that support any QoS let alone fq_codel or CAKE at a reasonable bandwidth rate, unless you're a power user dabbling in "prosumer" devices or racking your own switches. You could probably use a Raspberry Pi to do this for a 1Gbps connection but very few people have the know-how to do such things.
Is it just me or does $100 million sound like it's on the very, very low end of how much training a new model costs? Maybe you can arrive within $200 million of that mark with amortization of hardware? It just doesn't make sense to me that a new model would "only" be $100 million when AmaGooBookSoft are spending tens of billions on hardware and the AI startups are raising billions every year or two.
As an aside, I’ve never enjoyed the defeatist culture that permeated operations and distributed systems pop culture, which this post seems to re-enforce.