HackerTrans
トップ新着トレンドコメント過去質問紹介求人

peterwwillis

no profile record

コメント

peterwwillis
·6 年前·議論
Because people shouldn't have to increase their risk of death significantly just to have the privilege of having housing and food. The comparison is ridiculous. It's like saying, you already have a low risk of death by colon cancer, so why not eat things that dramatically increase your risk of cancer. Do you want them to die?
peterwwillis
·6 年前·議論
The point you've just tried to make is that it's fine to force teachers to risk death in a classroom, because they might also get hit by a car driving into work? You need to work on your rhetoric.
peterwwillis
·6 年前·議論
The teachers will work themselves into an early grave just to help the kids. Nobody's doing that for teachers. Teachers are more at risk of actually dying when in-person learning is forced on them as some school districts have done. And they have their livelihood to think about (many live a little more than paycheck to paycheck).

Comparatively, the kids aren't dying, they don't have bills to pay, they don't work overtime. Their parents may even be able to pull them from school, or homeschool. Worst case they might be held back a year.

The students will survive this, but the teachers literally might not.
peterwwillis
·6 年前·議論
This is what I think of when I think "innovation": a different perspective. Not lost in the weeds of details, but a thousand miles above them.
peterwwillis
·6 年前·議論
I like ordering groceries online, but I think the idea of instant grocery shopping is flawed.

1) You're depending on & supporting an oppressed segment of the labor force (people who need multiple jobs to make ends meet, because no 'gig economy' startup pays a living wage), 2) You're fighting with "surges" of use and paying wildly varying prices, with wildly varying levels of service, 3) You don't even get the items you're looking for half the time (in my experience), 4) It's more expensive!, 5) you don't actually need groceries delivered to you instantly.

Since COVID-19 I have switched to buying from local businesses and getting food from Co-ops and farm shares (and growing my own). I can schedule regular deliveries of produce to my door in bulk (in reusable containers), which not only reduces waste, it actually helps me plan my week/month better, I still get all the things I wanted, the price is fair, and I'm helping local farms and businesses.

I'm pretty sure the reason most people don't do this is that they're lazy, selfish, greedy and entitled. My evidence is the last two weekends I've ventured out into the world to look at apartments to rent, with the occasional jaunt to a hiking trail. I found virtually nobody was wearing a mask or socially distancing, unless they were forced to by a greeter in front of a business like a grocery store (many customers were trying to walk in without a mask, and thankfully were rebuffed). Many probably could have ordered online the same way I did, but would rather the immediacy of 'do it yourself' - even during a pandemic.

Sadly it's the poorest and most vulnerable people that need grocery delivery the most, because they have the hardest time getting to a store. A poor person in a food desert may need to spend over two hours round trip on transportation in order to get food, and an elderly person may just not be mobile enough. I doubt we would ever fund a public delivery service for them (though in theory the postal service could).
peterwwillis
·6 年前·議論
What use case are you thinking of? Two peers, or routing through networks of distributed decentralized peers?

For the latter you could take your existing model and add an extension "gateway protocol" that would connect servers to servers, and through that protocol support advanced routing, which would in effect become a distributed decentralized peer to peer network.

(obviously all that is a lot more work than you need for your use cases, but it seems like an interesting way to give people sexual agency without either a) commercial infrastructure or b) potentially unsafe public services)
peterwwillis
·6 年前·議論
Would it be terribly slutty for me to suggest a Peer2Peer protocol? Not everyone's into Master/Slave connections...
peterwwillis
·6 年前·議論
A lot of the problem with using Kubernetes is it appears to be the only option for running microservices in a cloud environment. People choose it because they think there's no other option (and they're somewhat right). But there's Nomad, DC/OS, Docker Swarm (for a little bit anyway), ECS, GKE, etc. That's still not a ton of options, but there are options.

That's just microservice orchestration. That's a small part of the totality of things needed to implement a full-out SDLC. You can't just build Kubernetes and think you're done; your code will need to integrate into a lot more stuff, and you'll end up writing 10 layers of glue because that's just how many use cases you have to support.

And it's weird that all that glue doesn't use standards. I mean, we have TCP/IP & RPC & REST, we have pipes & filehandles, we have the OCI specs. That gets us to a point where (at most) half of the stack of an architecture is portable and interoperable with any system following those standards. But then there's every other component of the architecture that connects all the pieces together, meaning you're writing glue that will only work for one implementation. Change your implementation, and you have to change your glue, and probably more stuff.

I think a lot of that non-reusable glue could be erased if it all followed standards, such that the configuration and operation of each part followed a standard interface, set of data types, etc. Tools and libraries could just "talk container orchestrator" or "talk load balancer" or "talk object storage" or "talk secrets management", and virtually any component could be integrated into any other, by virtue of either a system-wide or application-specific configuration.

You could argue we have something like that now with a "kubectl file" or similar, but that's not only still platform-specific, but other tools don't speak it, so K8s has to do everything, because it's the only thing that speaks its language (config file/backend data store/IAM/secrets/roles/etc).

Rather than resign ourselves to those limitations, we could bundle everything in an implementation-agnostic standard way with standard interfaces. The exact same configuration (as code) could be used to run the same complete architecture on a dozen different platforms, because every component would speak the same language and handle all the other components in the same ways. The backend services could all translate the standard based on how they were configured, such that generic instructions are then translated into implementation-specific actions. You could really write your architecture once and run it anywhere, without the caveat of "anywhere on this platform only".

I feel like we're not talking about doing that because we keep getting caught up in "Fuck, Kubernetes is pretty hard" conversations. Yeah, it's hard; building and operating an 18-wheeler is hard. But what about the roads? What about the gas stations? What about the containers we put on the trucks? All that stuff is standard, and so we don't have to worry about what implementation of gas station or road we use. I feel like we still don't have those things in the cloud, and it's just weird.
peterwwillis
·6 年前·議論
It sounds cool, but it also extends the amount of components that have to be made resilient to failure and attack. Your HA vault+consul clusters, HTTPS & OAuth2, key generation, and automation pieces (inc. message passing & load balancing) all need to be working correctly. Compare that to a single stateless server which spits out an OAuth2 login url to a client, receives a token once the client is authed, and opens a connection with that user's specific network authorization.
peterwwillis
·6 年前·議論
> It’s as if Mint, with 13 million-plus registered users, were a resource-constrained startup instead of a property of Intuit

Somebody doesn't get how business units work. Your OPEX has to constantly decrease while the revenue has to constantly increase, but your revenue mostly goes to other BUs' CAPEX. So there's no money to fix tech debt, unless it's buying crap to scale with, if you can show growth. Doesn't matter if your daddy is M$, your allowance will still suck unless your lemonade stand is making bank.
peterwwillis
·7 年前·議論
Real talk: if you haven't kept yourself accountable for 6 years, you probably won't now.

I say quit your job, but don't have a plan; figure things out afterward. Catapulting yourself out of your comfort zone is the best way to get to know what you really want to do, and force yourself to care about what it is you want to do. Nobody here can tell you what that is.
peterwwillis
·7 年前·議論
My job is managing infrastructure, services and products for large corporations in the cloud. It is managed hosting, and SaaS/PaaS, and yes, there are plenty of providers who have alternatives that we regularly use, in addition to being able to build our own with mostly open source tools. It's part of my job to find alternatives so we don't get into a jam when the one solution we use becomes unusable.

The most "cutting edge" things I remember announced at re:Invent 2019 were machine learning and AI stuff, but it was all stuff integrating with existing 3rd-party services and products. They know you want to be able to work with somebody else's tool, and they want to make it easier for you to spend money using that tool on their systems. The bulk of AWS's income is still EC2 (and data transfer); the rest is window dressing to keep you from looking elsewhere.
peterwwillis
·7 年前·議論
"The Cloud" is just managed hosting and SaaS/PaaS. There are literally thousands of such providers in the world, and until everything about providing managed hosting and SaaS/PaaS becomes impossible to compete on, there always will be.

The only "duopoly" will be in terms of the hyper-scale size of such providers, but very, very, very few customers actually need such hyper-scale. Not to mention, you literally get to pick and choose how much of their tech you use, limiting any potential risk of lock-in. Not to mention, the hyper-scale providers are probably the most expensive ones!

I think this is all very over-blown. AWS's services are fantastic for scaling up services quickly, and their customer focus is great, but if they were all gone tomorrow we'd still be doing the same things as before, just slower and with less reliable infrastructure.
peterwwillis
·7 年前·議論
Amazon sent out 3.3 billion packages in 2017, and saves $2-$4 per package shipped if they handle it themselves. USPS, UPS and Fedex are all raising their rates. You do the math...

They also have 50 Boeing 767's, btw.
peterwwillis
·7 年前·議論
Yes, and that's a great general purpose solution. The problem comes when you try to market this as a replacement for a car, and then it doesn't work. The platform is too narrow, some require lots of fiddling to get right, you don't have a comfortable seat, you're open to inclement weather, you have to already have a decent e-bike, etc. Taking a lot of time to find your own solution is one thing, but providing everyone a one-size-fits-all (remember, the intention is to prevent people from needing a car) involves a lot of considerations. With every change you have to reconsider the whole design, because even small changes significantly affect the ride.
peterwwillis
·7 年前·議論
This is a great point. I see people bringing up lots of questions and concerns about these devices, but it's missing a critical point: these were designed with specific use cases in mind. They are only useful for people in cities to go short distances. And within that scope, they are an excellent addition to other transportation options.

What you mention is actually what I want to build next: micro-mobility for larger errands. If I buy a trunk load of groceries from the Wal-Mart 6 miles from me, I want to be able to bring it home. I'm hoping to start designing devices that fit this use case. Once they exist, they'll provide tons of benefits: more affordable than a car, the ability to perform trips faster than with public transit, and more compact "traffic" on roads compared to cars (to say nothing of emissions).

Another use of these devices is actually changing people's life decisions. I moved to my current address because it's very close to a subway and trolley line. But now that I use an electric scooter to get to work, I could move anywhere that's the same distance. I think there'll be tons of these little wins found as the devices expand more. (I should note, my city does not have any rental scooters, but I see people riding their own all over the place)

As batteries get better, I think we're going to start seeing more variety of devices tailored for more use cases. The scooter will own the short trip into work, whereas larger devices will be used for cruising off-road, even hauling a sheet of plywood from Home Depot.
peterwwillis
·7 年前·議論
They've also had these weird devices called 'telephones' for a while now, seems you can accomplish a similar feat. (Also, do people really need to say goodbye to their loved ones randomly in the middle of the day on their watch?)
peterwwillis
·7 年前·議論
I don't really care if it costs $1, it's still ridiculous that people today strap two supercomputers to their wrist so their wife on a business trip in Beijing can tell them over live high-def video to feed the cat.
peterwwillis
·7 年前·議論
And all of that requires 2 Cray supercomputers, unlike, say, the Apollo flight computer, a 2mhz cpu w/ 2048 words of ram and 36k of rom, in 1966
peterwwillis
·7 年前·議論
And the software on these supercomputers is used to tell the time and display a calendar.