HN user

stormbeard

566 karma

https://www.linkedin.com/in/tony-allen-a2082721/

Posts7
Comments99
View on HN

Between the two zip files, these files have had their content changed:

VOL00001/IMAGES/0001/EFTA00000160.pdf

VOL00001/IMAGES/0001/EFTA00000177.pdf

VOL00001/IMAGES/0001/EFTA00000214.pdf

VOL00001/IMAGES/0001/EFTA00000215.pdf

VOL00001/IMAGES/0001/EFTA00000216.pdf

VOL00001/IMAGES/0001/EFTA00000746.pdf

VOL00001/IMAGES/0001/EFTA00000822.pdf

VOL00001/IMAGES/0001/EFTA00000824.pdf

VOL00001/IMAGES/0001/EFTA00000827.pdf

VOL00001/IMAGES/0001/EFTA00000888.pdf

VOL00001/IMAGES/0002/EFTA00001046.pdf

VOL00001/IMAGES/0002/EFTA00001069.pdf

VOL00001/IMAGES/0002/EFTA00001213.pdf

VOL00001/IMAGES/0002/EFTA00001217.pdf

VOL00001/IMAGES/0002/EFTA00001218.pdf

VOL00001/IMAGES/0002/EFTA00001353.pdf

VOL00001/IMAGES/0002/EFTA00001375.pdf

VOL00001/IMAGES/0002/EFTA00001523.pdf

VOL00001/IMAGES/0002/EFTA00001524.pdf

VOL00001/IMAGES/0002/EFTA00001528.pdf

VOL00001/IMAGES/0002/EFTA00001932.pdf

VOL00001/IMAGES/0003/EFTA00002063.pdf

VOL00001/IMAGES/0003/EFTA00002064.pdf

These are files that only exist in one of the zip bundles, but not the other:

VOL00001/IMAGES/0001/EFTA00000164.pdf

VOL00001/IMAGES/0001/EFTA00000165.pdf

VOL00001/IMAGES/0001/EFTA00000167.pdf

VOL00001/IMAGES/0001/EFTA00000229.pdf

VOL00001/IMAGES/0001/EFTA00000384.pdf

VOL00001/IMAGES/0001/EFTA00000467.pdf

VOL00001/IMAGES/0001/EFTA00000468.pdf

VOL00001/IMAGES/0001/EFTA00000470.pdf

VOL00001/IMAGES/0001/EFTA00000476.pdf

VOL00001/IMAGES/0001/EFTA00000503.pdf

VOL00001/IMAGES/0001/EFTA00000531.pdf

VOL00001/IMAGES/0001/EFTA00000656.pdf

VOL00001/IMAGES/0001/EFTA00000657.pdf

VOL00001/IMAGES/0002/EFTA00001051.pdf

VOL00001/IMAGES/0002/EFTA00001052.pdf

VOL00001/IMAGES/0002/EFTA00001053.pdf

VOL00001/IMAGES/0002/EFTA00001055.pdf

VOL00001/IMAGES/0002/EFTA00001056.pdf

VOL00001/IMAGES/0002/EFTA00001124.pdf

VOL00001/IMAGES/0002/EFTA00001423.pdf

VOL00001/IMAGES/0002/EFTA00001424.pdf

VOL00001/IMAGES/0002/EFTA00001931.pdf

I didn't look to see what the content was.

I can't believe people are paying these crazy amounts for what is basically a fleet of firewalls. What is the difficulty in running VMs with nftables rules?

Are you an actual bot? 4-month-old account, 0 karma, and that's not even what the article was about.

NASA Goddard is being illegally dismantled during the shutdown. The events/timeline is shown and the full emails sent to Goddard employees are provided. Don't dismiss this as people complaining about a WFH policy--the buildings are being demolished.

I read this comment, went back to the article, and then came back to this comment. I have no idea what niche political agenda you're talking about- the message of the article is basically "solve problems your users are actually facing, not problems you think they have".

You can apply the concepts the author talks about to _literally_ any group that would make use of social media.

So what? People drinking less alcohol is something we should strive for, but I don't see the problem with taxing it and using the income to fund schools. Yes, if people drink less alcohol that will be less money for schools, but that does not automatically mean it will be less money than they receive today.

I worked at Amazon in 2021 and rage-quit after 9 months. The sign-on bonus I received was paid out monthly, so I didn't have to pay anything back. If it's large enough, they pay it monthly because they know it's very likely you won't make it to the 2nd year.

[dead] 2 years ago

This is an advertisement and doesn't have much substance.

I'm familiar with the "modern load balancing" take. Modern load balancers, like you'd find in any cloud vendor or OSS project like envoy/nginx/haproxy, learn about the endpoints it is balancing across via some service discovery mechanism. DNS is usually what you find, but there are other ways like envoy's xds mechanism.

When you generalize this by saying "cloud computing", it glosses over the fact that there's still a fleet of load balancers somewhere. When you use an NLB or ALB in AWS, there are many machines behind the scene and a very complex control plane providing those machines with the information they need to balance load.

The problem I'm talking about still exists here. I _know_ they have this problem because I'm familiar with how those systems are built and I know what shortcomings they have.

I didn't mean to imply that this is a problem with arbitrary machines that have GPUs. There are specific kinds of applications that use GPUs and because of how they work have hard limits on the number of tasks they can concurrently process on those GPUs.

Either way, I feel like these details are orthogonal to my original question. Do you think it matters?

From what I've seen, there's usually some hard requirement that a single instance of the application can handle N in-flight requests. This is fine when you have a single load balancer that uses some flavor of "least request" or "least connection" balancing algorithm. The problem arises when you have a fleet of load balancers- they don't share state, so each LB can't know the state of any individual backend without sending a request. This results in a failed request if the backend is already at full capacity.

Your tone is coming off as condescending and I'm not sure if it's intentional. I guess I should mention that I'm _very_ familiar with how load balancing works. This is a real problem and you can't "tune" your way into a consistent global view of the backend states.

Internal restructuring of the company may even take you from working on backend web apis to distributed databases.

This isn’t true. I’ve worked in multiple FAANG companies’ infrastructure orgs, including distributed kv stores, as a SWE. Anybody joining those kinds of teams are either specialists, very junior, or already had some kind of experience in the domain before joining.

Recruiters that say stuff like “we want strong generalists… blah blah” are not the people who are sourcing candidates for these deep systems roles. It looks a lot like the ML roles.

There's not much context here that explains why this is noteworthy. Arena allocation by itself isn't anything to write home about- is anyone able to explain why this is special?