I’d happily wager any amount of money I have access to that the people actually doing the implementation of these things are among the userbase.

Someone has to write the code and I doubt many people would quit their jobs over it.

By that logic just about anything the tech industry does could be attributed to HN

It’s exhausting enough to deal with services that change around on an annual/semi-annual basis with pricing and expectations.

Now the expectation is that we should tolerate goalposts being shuffled around on a weekly/daily basis with the added requirement of digging into bug tickets because there’s no attempt at transparency? The tech is cool but this is absolutely insane.

If you’re an individual developer paying $100-200/mo for a service that keeps changing, there is a LOT of reason to keep an eye on other products.

I’m not saying that there isn’t a reason to keep an eye on other products. I’m saying that every other product in the space has the same unit economics and will eventually need to charge enough to be profitable - and to continue training and hardware expansion.

Honestly a developer paying $200 a month is a nothingburger and if using their service to the fullest is losing them money.

For context, the company I work for gives each consultant a $2000 a month allowance and I think there are probably around 500-700 people with that allowance. I’m sure everyone doesn’t use it all.

If they have limited hardware resources, where do you think they are going to focus?

If someone is still using the “remembering IP addresses” argument in 2026 (or at any point in the 21st century), I question their technical competence in configuring a network correctly.

It also seems to be a learning curve thing because IPv6 addresses have their own versions of memorable mnemonics. If you are in a LAN space manually configuring LAN addresses, you just need to remember one of the local address (ULA) prefixes like fc00 and then start numbering your devices as ::1 and incrementing (fc::1, fc::2, fc::3, etc). But also in LAN spaces you could just rely on mDNS (devicename.local), it's gotten quite good in most OSes today.

If you need to remember random WAN IPv6 addresses without being able to use DNS or at least a hosts file you've probably got a bunch of other more pressing problems.

Fiduciary duty is fun to define because I’d bet it could be argued both ways here. If you want to consider Costco’s low margins as a core factor as to why consumers choose them, opting for a decision that makes their customer base run off wouldn’t be very responsible to shareholders.

Consider the Target backlash last year. They’re since down 14% vs Walmart (up 30-ish%). Regardless of anyone’s political beliefs, I don’t think a 14% loss seemingly caused by behavior that a segment of customers considered hostile is thinking of the shareholders.

Right but they're not being sued by their shareholders, they're being sued by a handful of customers and "on Behalf of All Others Similarly Situated".

> they're being sued by a handful of customers

To be fair, they’re being sued by customers who were marketed memberships.

To be really fair, they're being sued by lawyers hoping to take 50% of the proceeds, or 50% of some settlement that they get by shaking down Costco via threats to its reputation.

> To be really fair, they're being sued by lawyers

Is that the case history? Or bullshit assumption? Because this looks plaintiff sponsored.

Adding "bullshit" to a sentence does nothing to hide this kind of ambulence-chasing vulturism and exploitation - in fact it rather highlights it.

I mean, one of the legal firms behind this is Milberg PLLC, https://en.wikipedia.org/wiki/Milberg, who has been charged with illegally paying plaintiffs to sue in order to enrich themselves.

Whether an appliance OS uses SystemD or not is as silly of a concern as “does the lead developer prefer cheddar or brie”

What about performance characteristics? Recoverability of workloads?

I’m interested in a FreeBSD base OS because it seems ZFS is better integrated and ZFS has a lot of incredibly useful tools that come with it. If Bhyve is at least nearly as performant as KVM, I’d be hard pressed not to give it a whirl.

It's not silly at all.

I've been repeatedly burned by systemd, both on machines I've administered and on appliances. In every situation, the right fix was either "switch distros" or "burn developer-months of work in a fire drill".

In fact, I just decided to go with FreeBSD instead of proxmox specifically because proxmox requires systemd. The last N systemd machines I've had the misfortune to touch were broken due to various systemd related issues. (For large values of N.)

I assume that means anything built on top of it is flaky + not stable enough for production use.

I have never really understood the systemd hate. It sure as hell beat the sorcery that was managing init.d scripts for everything.

I managed the distro upgrade on hundreds of remotely-managed nodes, porting our kiosk appliance from a pre-systemd debian to a post-systemd debian, and out of all the headaches we suffered systemd was not one of them, short of a few quirks we caught in our development process. It pretty much just worked and the services it provided made that upgrade so much easier.

Curious how you got burned, I hear a lot of complaining but haven't seen a lot of evidence

I don't understand the init.d script hate ;)

It absolutely is silly. I’ve been responsible for managing low-thousands of Linux servers with systemd and it’s standardized a lot of things that otherwise would’ve been a lot of bespoke scripts.

Yeah, I’m kind of in the same camp. I never really had issues with systemd either. It mostly just works, even if it’s a bit heavy.

For me, moving to FreeBSD wasn’t about escaping systemd, it was more about the overall system design and how cohesive everything feels. That said, I’ve tried to keep Sylve neutral on that front. I don’t really position it as “systemd vs not”, just focus on what it actually does well.

It’s still early and not as feature complete as Proxmox yet, but I think it already stands on its own as a solid option.

“does the lead developer prefer cheddar or brie” Quite right but given I live in Somerset (UK) I can have both: Cheddar is in Somerset and where the eponymous cheese originated and quite a lot of brie is produced here too - it's not the French original effort but rather good.

I have quite a lot of customers that we have migrated from VMware to Proxmox. Some of them are rocking zfs instead of vmfs. Mostly these are Dell servers. Proxmox with zfs seems to be more aggressive about disc failure warnings, which I think is helpful.

Pick what OS works for you.

I agree that this is what everyone should strive to do but this quickly hits a limit.

For example, IAM/S3/SQS policy evaluations can have profound impact on an application running but an abstraction wouldn’t help much here (assuming the developer is putting any thought into securing things). There just isn’t an alternative to these. If you’re rolling out an application using AWS-proprietary services, you have to get into vendor-specific functionality.

My point is you should not build on top of their native services if it incurs this problem.

it's a trade off, a risk. (sure, for many people it doesn't make sense, because they don't have scale or growth anywhere near the numbers where generic/abstracted is not efficient enough)

No, the opinion section was absolutely not pushed towards libertarianism.

Have you read it recently?

Libertarianism is just an euphemism for "authoritarian righ-wing, but dont want to admit it out loud" in most cases.

by now [1], yes. but in my experience not a very popular view on HN. expect downvotes.

[1] i find peter thiel's speech at libertopia in 2010 a great early reflection of that shift: https://www.youtube.com/watch?v=KgH7Lv2gQdk

Yes.

Agreed. AWS is downright hostile about giving you any idea about what resources you actually have deployed, to the point where it must be deliberately malicious. Even their billing page is terrible for tracing down the root cause of usage with the default configuration.

You have to go into third party tooling if you want any chance of seeing what’s actually going on, especially if there’s any odds of you deploying stuff in another region and even moreso if you have more than 1 account.

At this point, I’d say it should be a best practice of owning 2 AWS accounts, even as a hobbyist: one payer account with a HEAVILY locked down SCP and then a child account with the stuff you’re deploying.

This reads like vague posturing instead of accepting (or even just looking at...) the reality on the ground.

I have about a dozen friends spread across 8 different mid-to-high level universities around the country in biomed. Europe and Canada are definitely a preference but China is entering conversation and has been for the last few years.

The alternative is to abandon an entire career or field of interest because the funding is held up by irrational national political policy.

There’s a size of enterprise where you can get away without PSTN integration but do need an answer for SSO and account provisioning/deprovisioning.