HN user

mentat

2,067 karma

You can contact me at jfburns ((at)) gmail ((dot)) com.

[ my public key: https://keybase.io/1mentat; my proof: https://keybase.io/1mentat/sigs/la6OVrEkrrjxpm0dCi276b2B5nUCMUw59rFH0EU_IXc ]

Posts6
Comments1,004
View on HN

You’re very sure because you’ve built one at scale? This is such peak HN. And you have plenty of other people who have never built a billing system at scale to agree with you. The OP is right, this is in fact very very hard for the reasons they say.

The idea that most people who are doing professional public speaking are reliable narrators is a bit quaint. There is a lot of room for framing that you have to allow for story telling. If you think that all story telling is about reliable narration, you're going to have a tough time being successful at it or interacting with others who are.

I ask a lot of questions on code reviews as it's usually less easily perceived as conformational and I have come to assume that I'm missing context when something is really off. Assume that coworkers are competent but may either have or lack context. If your coworkers are genuinely incompetent, then a new position might be better than fighting though PRs.

Saying this seriously a few decades after computers were room sized is I think a bit short sighted. Betting against technical progress especially in density / materials / weight is not a good bet.

What kind of firewall work didn't have default drop? Once you take that off the table you really do have a problem but I've only seen that one in 20 years of professional work.

"Remain obsessed" with transmission of a highly contagious virus that has shown significant ability to mutate at scale? That sure seems worth being obsessed about. When it mutates to spread past all current defenses then consistently kill because we failed to stop current spread it will indeed be "depressing".

You have to pass difficult standardized tests to qualify for these spots though. It's not like 50,000 people can just decide they want to be doctors. It depends on ability as measured by testing ability which generally still relates to economic ability in aggregate.

The major downside of having an “always syncable” data structure is that it gets really big, really quick

Perhaps it would be prudent to read the actual article before tossing off this sort of comment since it's indeed the whole point?

Do you know of any good writeups for this particular kind of process? I too did fuzzing of network devices before it was called fuzzing and am interested in trying it again with modern tooling.

Physical based attacks were the majority of the risk, and that was, practically, pretty minimal. If you want to watch from inside a VM, great, but mass attacks against watchers? I'm not seeing it. (Been going since DEFCON 9)