sargun

This is massively different from DynamoDB.

I accidentally dropped something into an iron mountain bin years ago, and just pried it open with a screwdriver.

I accidentally took down our corporate network multiple times ~20 years ago, because I was setting up virtual networks with Linux machines acting as l2 bridges, and with proxy_arp to on. The DHCP server only checked for an arp reply, and I caused our server to run out of IPs. It had a cache and nobody knew how to clear it, so we just had to wait it out.

Eh, at least one of those -- the "crashing the node" with messages argument is weak-ish, you can setup:

  erlang:process_flag(message_queue_data, on_heap),
  erlang:set_process_info_limit(memory, 1024000).

Isn’t the highly connected nature of the model layers problematic to build into physical layer?

I find it a little funny how much the government spends on these dead end investigations. We never will know precisely how much is wasted.

What ethnicity are you? I went through an airport -- and nobody else got screened except me. What was special about me? I was the only non-white person in the airport. Upon complaining, this was the response:

> Random selection by our screening technology prevents terrorists from attempting to defeat the security system by learning how it operates. Leaving out any one group, such as senior citizens, persons with disabilities, or children, would remove the random element from the system and undermine security. We simply cannot assume that all terrorists will fit a particular profile.

Andy Pavlo absolutely seems like the kind of guy that I would want to get a drink with.

So, you’re essentially washing the soil in DMSO, and DDT is more soluble in DMSO? — curious, what does it take to wash all that soil?

I have wondered why the likes of McKinsey, KPMG, and PWC do not put up candidates (don't even sponsor them, just say you're electing _well known consultancy_).

What's the US Customs ruling in question? > This update was enabled by a recent U.S. Customs ruling.

I really like the manycores approach, but we haven’t seen it come to fruition — at least not on general purpose machines. I think a machine that exposes each subset of cores as a NUMA node and doesn’t try to flatten memory across the entire set of cores might be a much more workable approach. Otherwise the interconnect becomes the scaling limit quickly (all cores being able to access all memory at speed).

Erlang, at least the programming model, lends itself well to this, where each process has a local heap. If that can stay resident to a subsection of the CPU, that might lend itself better to a reasonably priced many core architecture.

I think the most interesting thing about this is the origin of the project - multikernel. It seems like someone is building Solaris zones for Linux.

Investment gap is what I'd say too. While Rust, Go, Python, etc... have had massive backers that have managed to invest a ton more into things like static analysis, type checking, and developer ergonomics, the Erlang ecosystem hasn't necessarily had the same love, and instead the major users have typically chosen to pivot, or build something outside of the BEAM.