ragona

> But benevolence is transient and brittle. I don't want to depend on it.

Yup, exactly.

Let me be clear, I do not. I think the U.S. government in its state can barely run itself. It appears to be actively working to be as net harmful to the general populous as possible.

I genuinely believe in Amazon and its humans. But at some point in the distant future internet infrastructure will be so ingrained in our lives that it IS our society, and I want to live in a Democracy.

I want our representatives to continue asking powerful companies how they're handling our data, and if they're becoming too big, and all of these things. I see this as a fundamentally good thing.

I don't believe that AWS should be handed to the government, that would kill AWS. I believe that when something becomes critical to the public we should talk about what that means, and how we stay safe.

Amazon holds themselves to an incredibly high bar, and I really applaud them for that. It made me proud to work there when we talked about how critical the work was, and I was awed to see how far we would go to recover from disaster.

But Amazon has this saying about relying on mechanisms and not best intentions, and relying on Amazon's best intentions is foolish as a citizenry; we should continue to question them. And if at some point in the future we realize that AWS is more important than the self-governing society that we live in we should ask whether that is okay.

Tech companies have become some of the most powerful entities in the world in short decades. In generations they could become the world. We shouldn't rely on their best intentions.

I think power is an interesting example because everywhere that it is privatized it tends to be a mess. Shit, PG&E burnt down half of California... what, on behalf of its shareholders, trying to cut corners and make them a buck? Who the fuck did that help? So yeah, I'd really rather see basic internet infrastructure be made into a public utility. Maybe 100 years from now S3 might make sense too.

The problem is that we're so busy arguing about whether we should even HAVE a government it's incredibly difficult to do anything smart with ours.

The bit of this that I think is easy to overlook is that if this is the status quo indefinitely I assume this is a very long term problem.

Amazon is a baby, it's a new-ish company in the grand scheme of things. It's a huge, huge baby, it has accumulated power incredibly quickly. Where is it gonna be in 200 years? Will the federal government still be more powerful?

No one is saying nationalize Amazon tomorrow. But we haven't even managed to give our citizens good internet service yet, we're way behind here, and we do not appear to be catching up.

> Fantasyland

I love that in a comment about fucking space aliens and competent governments the thing that made you call it fantasyland was the latter.

Okay, the French have an incredible train system that goes down sometimes when the union feels the workers are being treated unfairly. That compares poorly to... what? To our total lack of a train system? To our anemic unions and abused workers? I read this and it sounds like a success story for government competence.

I don't know how to think about the idea that the government is totally useless for technological problems. If we just accept that I think we're probably just doomed to be overtaken by more advanced societies. Agreed that we can't just dump AWS on the USDS or whatever tomorrow, but if we never plan for public digital infrastructure we're screaming towards some kind of Gibsonian cyberpunk future where the tubes are privately controlled.

Yeah, this stuff isn't secret. Note how if you encrypt your S3 bucket it uses a KMS CMK; that's an dependency between S3 and KMS, and yep it's the same KMS. This is true for probably over a hundred AWS services that rely on KMS; it's a "foundational" service. (As is S3.)

I'm curious how many companies would survive it at all.

Sure, you could _migrate_ off of AWS if AWS was still alive. Could you restore your data and start your business from scratch if you lost everything? I suspect many rather large businesses could not.

I love this.

Imagine, your company hits $1T. POTUS gets you on stage. They tell you you did SO GOOD. We're really proud of you. We're gonna make you a statue in the Hall of Really Good Capitalists, and we have the best news. You are SO GOOD at this that we're gonna let you do it again. You're poor now, but we're gonna put your name in history books! If you get this rich again on hardcore mode we will make your statue bigger!

Well I'm glad you thought my argument was strong!

Yes, that is what I believe. I see a path where after Amazon creates something that becomes a de facto public good they hand it over to a highly competent government agency full of the world's best operational engineers. Those people turn it into a public works project, and it has legendary 25-nines uptime.

Amazon engineers are excited to see their work become part of the fabric of the Galactic Alliance of Good and Kind Beings and no one does oncall like a Quarkian, am I right, 700-hour sleep/wake cycles how cool is that, love working with them.

Do I see a path to get there from here? Uh, not at the moment. We're still working on "hitting each other is bad" so I'm feeling pretty depressed about our outlook.

> how much faith do you have in the perpetuity of that benevolence

None! Literally none. This is a huge risk for our country, and we should thinking about what it means for the next 1000 years, not the next 10.

Bezos appears to truly, truly believe in Customer Obsession. The culture definitely softened while I was there though, and I can't imagine how it'll change in 200 years. I can't guarantee his great grandkids won't be dicks. I don't really want to live in the corporatist city state of Seattle By Amazon with its "benevolent" AI overlord and its drone stazi in 2234, ya know?

The data center operations team is probably still in the building. But the software engineering teams which is what I’m familiar with are all working from home. We don’t have to go to the data centers anyway usually.

Disclaimer: I worked at AWS from 2012-2020. I am not a current employee.

> I came to think of Amazon and Google as the providers of the very infrastructure of the internet, so embedded in the architecture of the digital world that even their competitors had to rely on their services.

This is the right take. I disagree with the disappointment around this article.

It’s about the fact that tech giants are running the infrastructure of the internet. Its the federal highway program of the 21st century. It’s a critical piece of public infrastructure. I can’t even begin to wrap my head around the death and mayhem that would ensue if AWS and Google simply packed up and turned off the data centers.

When we consider the future and safety of our country we should keep these things in mind.

Imagine an unusually well-attended AWS all-hands was hit by a freak meteor. What do you even do? Take enough engineers off the oncall rotation and things will eventually fall over. I think the government would actually need to take emergency action at a level we don’t understand to prevent the collapse of our financial system beyond god knows what else. It would be utter mayhem. Globally.

That said, no one takes this more seriously than AWS. I really believe this. I sat in meetings regarding how to recover if we lost entire states, like literally what if we lose the Virginia to a catastrophe. (I worked on the Key Management Service or KMS which underlies much of the rest of the infrastructure.) But it is good and correct that our representatives are questioning our tech giants about their role in the future of our civilization. They’re playing perhaps the most critical role, and that should scare you.

I actually believe in the benevolence of AWS; I spent nearly a decade there. Do you?

Oh that’s neat, I did enjoy it. Thanks. It makes sense to me. Lying is work. Humans usually take the easiest path. Sometimes lying is less work than actual work and then humans lie.

Oh I don’t. All digital security problems are eventually physical security problems.

Regarding your last sentence, I think that’s true for, well, basically the entire world. Our security sucks in general. Humans aren’t as mean as they seem, or maybe they just generally suck at it.

Hey the images turn out great. That’s really cool!

Yeah Roblox is nuts, when my son talks about a “new game he’s playing” he’s usually talking about Roblox.

Ha hey, Greensock was and was and is the best, and it’s still going strong for Js. Frankly we could probably get half the experience of flash if we all just went and worked with that library!

They were! We had much better build and art pipelines than many of the PC games I’ve worked on. And the metrics data tracking and ugly anti user behavioral stuff was years ahead of itself.

Yeah that’s a huge oversight. AS3 was in fact the highest paid language in the industry for some time, and it was all of the social games. Then we all mostly jumped ship to mobile or PC or VR or whatever.

But it was a very significant industry, every major games company had divisions building FB games, we had games hitting 20+ million DAU overnight, it was big business.

I made Flash games (among other things) at Sony, Disney, PopCap, and Amazon. I’m not sure people understand how widely it was used.

Now... I also saw the writing on the wall and made a hasty exit in 2012 because it was clearly dying. But for a while it was a vibrant and well-paying niche.

Major FB games were all AS3, which was a very standard language. If you’re familiar with C# or Java you’d be totally unsurprised.

In fact, Adobe hoped it would become the new JavaScript. The history of ECMA 4 was fascinating[1].

1: https://auth0.com/blog/the-real-story-behind-es4/

YO, same, I bet we hung out on Kirupa at the same time haha.

I worked on flash games at Disney, PopCap, Sony and Amazon. That was the shit for a while for 2D games. But I actually went from social to mobile to PC games and found myself less and less interested. Don’t get me wrong it’s very cool code and quaternions solve a lot of the pain so some of the math is frankly less fiddly than 2D, but I wasn’t in love.

It’s been nice to fully pivot towards cryptography and security since I get to work on a lot more projects and do a better job on them. No one cares about bugs in games. Everyone cares about bugs in crypto. It’s nicer for the engineers.

Plus it’s frankly so deep that I have very little risk of mistaking myself for an expert. Even PhD’s only have a narrow specialty, it’s very comforting how hard it is.

I didn’t really find anything I felt like specializing in for the first decade-ish of my career. I was working on games, I don’t think 3D math is that cool, and I wasn’t excited. I was tired of coding games over and over. I don’t think I was an expert beginner but I wasn’t an expert in anything I wanted to continue doing.

I took a break to be a manager for a while, which turned out to both be an incredible learning experience AND enough pressure relieved from the daily code grind that I was able to rekindle my passion for code.

Started writing a lot outside of work, realized I really like cryptography, studied that, became competent enough to even figure out what TYPE of beginner I want to be, etc. it’s very satisfying.

I’m around four years into that journey now, I’ve returned to a coding path where I get to actually learn while building, and I’m excited.

The expert beginner is a surprisingly easy trap to fall into. I also think we do this TO people when we trap them in a role and give them very little flexibility in execution.

They will begin to specialize in doing a bad thing well without understanding what they’re doing, especially if they aren’t given mentorship.

Even better, WSL isn’t required these days. There is a full version of ssh in the standard cmd terminal.

They didn’t launch it with a big splash or anything, but it has saved my butt before.

Exactly, it’s always both. These conversations always suppose a degree of coordination and agreement I’ve never seen at a large corporation.

I’m guessing they’re referring to the fact that the CEO is very wealthy [1] and has self-funded the company, and suggesting that he’s not as concerned with the financials.

1: https://www.forbes.com/profile/adam-dangelo/#657861427cb4

I think that’s a little lacking in nuance. The team they have putting this together say to me they’re putting their money where their mouth is, at least. That paper has a fucking legit list of contributors.

Very similar.

Did you read the article?

> “Well, I'm certainly not talking about any kind of scenario where tomorrow someone just flips a switch and there are no police. What I'm talking about is the systematic questioning of the specific roles that police currently undertake, and attempting to develop evidence-based alternatives so that we can dial back our reliance on them.”

Vitale offers a nuanced perspective, and your response is a barely-related strawman.