0xEFF

Genuinely curious why? Isn’t it just a large http server for static binaries?

Kubernetes is already fully decoupled from docker. Defaults for new clusters are containerd today.

You may be thinking of Twitter.

Apple TV is quite good these days.

> Really you should just tax income at high rates and investment at near-0%.

Does this address taxes lost from people being paid in stock units?

You can see the window size is zero though, so I think GP is suggesting sending a TCP reset or something similar.

Maybe this isn’t a good option because it would have too many undesirable side effects?

Seems selfish. Lockdowns reduce deaths.

Edit: I don’t mean to call anyone names, genuinely interested in understanding if you’re concerned about being perceived as selfish and if so how you reconcile the tension?

My exposure to BDB was through sendmail, spam assassin, and subversion. All three a big chunk of my job in 2006.

For context, before the rise of SQLite, BerkeleyDB was often used for local file databases.

I never programmed for berkelyDB but I don’t believe it was SQL, so I think that’s what made SQLite so successful. The combination of an SQL api backed by local files

Down voted you because of your about me.

My chase sapphire card number was stolen while on vacation out of the country before covid. Reported the fraudulent charges and got a new card and number.

Recurring subscriptions transferred for about 12 months then I had to update the card number.

I’m not a payments expert but presumably the subscription is tied to some authorization and not necessarily the number itself.

Fair though, today I learned getting rid of RSA and using opinionated DH parameters is a significant aspect of what defines TLS 1.3.

And a license.

> For over 90% of workloads kubernetes is an overkill.

It's not. Take any simple web app and deploy it into managed GKE with Anthos and you automatically get SLI's and the ability to define SLO's for availability and latency with a nice wizard. Takes a few minutes to setup.

The amount of engineering needed to achieve good SLO monitoring dwarfs the engineering needed to run a simple app so it just never happened. That's no longer the case.

> Only when company is reaching google scale kubernetes make sense.

Also obviously not true given the number of companies deriving value from kubernetes.

Note, I'm a Google Cloud partner.

It’s not that much. It’s about 1600 hours of work which seems like the right amount taking on Uber and their lawyers.

Also factor in the risk of working 1600 hours and getting paid nothing for it if the case was lost.

I'd consider taking the bet, but only against, "We'll be on Mars before we have generally available self-driving cars that do not require humans."

Both seem unlikely.

All of those places can also be read.

SSH agent is a good example. It’s effectively an environment var which is why this works fine:

  sudo SSH_AUTH_SOCK=$SSH_AUTH_SOCK git clone ...

Edit:

The reason I think it’s silly to make a blanket statement environment vars are bad is because too many containers have credentials baked into the image when they should be passed in another way.

How do you auth to vault?

You may already know, but in case not, you can also use this tmux feature from your editor. For example in vim

  :w !tmux load-buffer -

Writes whatever you have selected in vim into the tmux buffer through standard in.

Elsewhere you can paste like pbpaste.

  :r !tmux save-buffer -

The save-buffer command is’t the most intuitive name, but it’s saving to a file, stdout in this case.

What better place is there to store credentials?

Yes, but reasonable to leave over the actions of many people.

Managed services with state of the art IAM policies are more secure than lifting and shifting a Linux box running whatever PAM configuring was setup on it in 2005.

This works in reverse as well, which is useful to use your laptop as a proxy. Suppose you can SSH to a server that can't connect to something your laptop can, like a git server accessible only over VPN.

  ssh -D8081 localhost
  ssh -R8081:localhost:8081 some.remote.server
  http_proxy=socks5h://localhost:8081 git clone git@git.corp:foo/bar.git

I agree with this up to a point. It’s been valuable to me as an IC, and I feel it’s a value I provide when the problem is so hard there isn’t a clear solution.

It’s nice to be able to say, “what do you think I should do?” as an IC, and nice to be able to analyze the problem if necessary as a manager.

On the other hand if nobody knows what to do with a hard problem, everyone gets stressed out about it.

> If I'm creating tax filing software, do I need to spin up a local instance of the IRS to test with?

No, integrate directly with the test endpoints provided by the IRS [1].

[1]: https://www.irs.gov/pub/irs-schema/Portal%20Transition%20Gui...

This discussion thread is a good start.

A short system design doc with a working prototype is vastly superior to a 50 page design doc with no code.

Think of it as the scale model of the house in the original example. It’s a way for sake holder to see and feel what they’re going to get.

So if a new investor puts a stop loss order in at a 60% loss, brokers let them exit their position and then didn’t let them back in?

This seems unconscionable given the >300 point volatility today and the large number of brand new investors.

It’s hard to imagine how these actions didn’t benefit wallstreet at the expense of regular people.

One thing I’ve read is both a short squeeze and a gamma squeeze have never happened at the same time to this degree.

Is that true? I tried and failed to verify.

I had this experience too around 2015-2017, but I've noticed a big difference between 2018 and now. They've hired a large number of people whose job is primarily focused on understanding the customer's business and use case. Most of these people have empathy in my experience.

In the large enterprise migrations I've worked on the past 2 years, I haven't heard a Google employee or partner say, "you're doing it wrong." Sometimes there is a response along the lines of, "we don't recommend doing it that way for the following specific reasons. For your consideration, here are alternatives we recommend for your use case. We've filed your use case as a feature request with product."

Additionally, I've seen issues which block a migration to GCP get escalated and fixed quickly. Issues which in ~2015 might have garnered a, "you're doing it wrong" response.

I've noticed a real shift in attitude and execution. There is genuine empathy and an attempt to understand how enterprise customers run their workloads in the cloud, multi-cloud and hybrid-cloud.