Most Arc users have long since migrated to Zen. I personally have never liked either of them. It's hard to step away from Vivaldi. Although Vivaldi runs Chromium, it in 2026 is basically what Opera 12 was in 2016 (before they crippled the browser with the Chromium re-write). Vivaldi just has too many features. The only browser I've found that is comparable to Vivaldi's feature-set and customizability is Floorp.
i too like vivaldi for the same reasons. i use brave, chrome for google stuff, firefox if i need a third browser but i use vivaldi mostly. except on mac where i still use safari
> […] could have been using Proxmox all along without any of the licensing and fee nightmares.
As a user of Proxmox, and observer of VMware infrastructure in other silos in some IT orgs I've been in, there is no way that Proxmox was a viable solution 3-5 (or more) years ago.
Hyper-V could have been used in many instances. I know a few places that chose to start moving to Nutanix even before the Broadcam fiasco. OpenStack is useful for many large scale environments (especially with tenant isolation and usage tracking/billing/chargebacks).
Proxmox is good a good solution for SMBs, but (especially pre-9.x) there were a bunch of things that made it much less usable for large(r) scale deployments IMHO.
I too ran a large VDI, server, and scientific computing cluster based on Linux - libvirt in my case. I saved a bunch of money over VMware. When I quit to go to a higher paying job, they had to rip and replace everything with vSphere because they couldn't find anyone willing to work for .edu money with the skills to manage that despite the massive cost savings.
I had asked for vSphere to avoid this problem, and they balked at the licensing costs even in the good ol' days.
I emphasize most, not all. Most companies using VMWare even 5 years ago could have achieved similar infrastructure using Proxmox. It's been a viable alternative for years. It's always intrigued me how easily big companies get pulled into licensing nightmares because their execs get bullshitted into believing a sales pitch at a seminar. Anyone who has worked with Oracle's product lineup instinctively knows what I'm talking about.
> I was working on a pretty big proxmox setup back in 2016...
Define "pretty big". Proxmox officially supports up to 32 nodes in a cluster and the Proxmox folks themselves have "seen clusters of around ~24 nodes in production" (as of 2025) and for "bigger clusters than that, fine-tuning might be necessary":
There are folks talking about running things in the high-40s and low-50s, but that's not officially support. If you have beefy hardware you can fit a lot of VMs on a single hypervisor (I've always found running out of RAM is often the limit).
A single vSphere cluster can have up to 96 nodes (without vSAN), and vCentre can handle multiple clusters (2500 total hosts; 40k VMs). Proxmox recently released Cluster Manager, which helps with feature parity.
I'm a happy user of Proxmox for a number of years (and helped ran OpenStack before that), but I think for large(r) scale deployments (and even smaller ones) there's a good reason which VMware was so popular. (How many folks are actually at that (large(r)) scale is a different question.)
(Also: XCP-ng doesn't seem to get much press. Haven't used it myself though.)
I have professional experience with all listed (both on vendor and user side), other than hyper-v. Nutanix is probably the closest true solution to VMware, although not 1:1. VMware is an excellent piece of tech run by people who DGAF and just want the $.
I like Proxmox a lot, but it is technically and commercially several years behind. There is no moat, and when stuff slides sideways, you need to understand what's going on (e.g a competent sysadmin). That is changing with the recent Proxmox NA announcement, but still not there.
OpenStack is not the same product at all, and is about an order of magnitude more complicated. Much more powerful, but instead of a good sysadmin, you probably want a team (and maybe some core devs).
Anyone larger like that automatically also can afford the staff to simply address any shortcomings in-house and then own it forever, and perfectly tailored, and perfectly customizable at any time forever. There is no missing feature that a few hours or days of developer time doesn't erase.
Oh it's crude on day one? So what? You're still operating and on your own terms, and every week a few more minutes of work gets added as little pain points get addressed by whoever cares, and before you know it it's better than anything you can buy, and the total cost was less than your previous licensing, and you never had to deal with a serial number or license manager or account rep, never had to care one bit about counting cpu cores or client seats or vms or physical hosts or ips or backup copies or spinning up new dev boxes or more production without warning or any of that bs.
The bigger you are, the less you need the convenience of outsourcing, not more.
> The bigger you are, the less you need the convenience of outsourcing, not more.
Quite often the bigger you are the more the higher ups want risk to be mitigated, and so are often glad to cut a cheque to be able to yell at someone over the phone.
The other aspect is that they want fungibility. As a corporation, you use off-the-shelf solutions so that when you do your big layoffs every few years you can rotate in some other cogs who already know how to operate the standard solution. Building in-house means valuing your employees long-term and we can't have any of that going on.
It's been years i played with Proxmox. Back then it wasn't as easy as Vmware Esxi was/still is: Just boot a pretty small os, define a datastore somewhere, just grow disks, migrate them, etc. All done with some clicks of a mouse button in a nice web gui. Maybe i should start playing with Proxmox again?
Ed Zitron is right about the need for a market correction, but he's fundamentally wrong on AI itself. He'll get his 15 minutes of "see I told you so" when the market corrects, but over the long term he'll be proven wrong.
Ed's primary gripe is that he thinks the business models aren't viable for profitability. But Google's AI infrastructure buildout is already spending less then the depreciation value of the hardware, meaning it's inevitably going to become profitable - at least for Google. Microsoft has since adopted the same approach that Google is using, focusing on faster and more efficient models in order to reduce costs.
Ed won't acknowledge that the paradigm shift for programming and SWE has already happened. He won't acknowledge that roughly 30% of radiology labs in the US and 40% of dental practices have adopted AI.
I personally think Ed is digging a hole he won't be easily able to climb out of.
> Can you make a falsifiable prediction without moving goal posts and ambiguous timelines?
I find this demand dishonest. The inherent nature of bubbles is that you cant predict when they pop. The more they grow, the more damage they cause, but it is impossible to guess what exactly triggers the pop.
So yes, ambiguous timeline is actually correct way to phrase things. Anything else is pure gambling.
I'm a SWE and use AI all day every day. I also acknowledged the very real rollout of AI in other facets of the economy besides SWE. But a market correction is inevitable.
We are in the wild west of AI. No one truly has agentic workflows "down pat" yet. We have no idea which AI workflows, pipelines, architectures, and platforms are going to stick around long term. For all we know a kid in high school can be the one who comes up with a SWE workflow that FAANGs and the rest of the tech sector ends up adopting.
Recent numbers from Google suggest that Ed Zitron might very well be wrong about the data center build-out, which is the single biggest crux of his argument. But even if he's wrong on that, he'll still get his 15 minutes of "see I told you so" (while still being wrong) because even if the data center rollout doesn't implode from assets depreciating too quickly, a correction will still need to happen on the dirt end of the shovels. Even if you don't have too many shovels (which Ed thinks we do), you still need to be digging in the correct patch of ground to strike a gold vein. There's going to be winners and losers in this thing, and the market correction will happen once the biggest investors and hedge funds have correctly identified who the losers are.
What I can't do is provide a timeline. I don't claim to be a prophet. However, if Ed is correct on the hardware buildout, the sheeit will hit the fan once amortization schedules hit critical mass. Otherwise it will be whenever the big money learns who the winners and losers are on the implementation rollout. That can happen at any time. It's perfectly normal and is not a doomsday clock. That's why it's called a market *correction*.