We have zero Honeybadger errors, performance is acceptable for all our routes in the application, and all of our key stakeholders are ecstatic about what we've built.
Is there some other metric I should be measuring our code by?
I can name features, but everything I can think of are technical features rather than obvious surface level stuff: DX12, better support for SSDs (Windows 7 doesn't natively support TRIM), HDR (I guess, but it still seems broken to me). And none of these are things that couldn't be implemented in Windows 7. The UI has nothing to do with these things, and there's no reason we couldn't have them without the trouble Windows 10/11, other than the fact that MS doesn't want to do things that way.
This was true at least as of 2023-25 when last checked, having been super annoyed by HDR just looking flat out wrong on my PC vs my Mac (which is probably the only OS that gets everything right given their large audience of Creative/visual professionals and mobile experience).
On Windows 11, when you reconnect to a monitor or set of monitors that you've connected to before, it will automatically return your open windows to the layout across those monitors that you had when you last disconnected (assuming those windows are still open).
This is extremely nice and saves me time on a literally (not figuratively) daily basis, to the point that I generally forget that it hasn't always worked that way.
I wish this worked! I have to go to the office on my hybrid schedule. When I switch between home and office, Windows is "smart" enough to keep the windows on the correct window in the task bar while generally placing the windows on the opposite window where I want it (and literally opposite where it is on the task bar!) It's so annoying and I dread the days I'm switching between office and home for this reason, as I have to drag each window to the opposing window before things are back to how I want them. It would be less bad the old way, where they were just stupidly thrown on the "primary" monitor and I only had to drag half of them over.
"Well, they turned the entire OS into a tracking, sales and ad/propaganda delivery service, but they managed to make a single feature non-dumb, so guess we're even."
(propaganda - Windows 11 default widgets are "offering" a lot of russian-biased media, because Microsoft is too dumb to recognize that and they take any news source - and russian connected outlets are happy to use this delivery vector that most gullible people leave turned on)
I don't think that any of the news-oriented default Windows application since W8 had an option to provide a custom RSS channel. It was always a default pool of sources they were bringing.
For my dual monitors, they have a conflict with this feature where they do not detect signal and then switch inputs and eventually power down. Then windows sees a different config and switches again causing an endless spiral. I have to turn both monitors on to the correct input while plugging in the laptop to the dock. I wish there was a way to save specific monitor setups and manually toggle them.
Yet when I switch between home and work I have to fully restart my laptop about half the time in order for it to even detect the monitors. I also find this feature has an issue with certain programs (Obsidian in particular) where it opens the window almost off screen.
Even aside from the malevolence, Windows is rotten from the thirty-year old metaphor that it started with: windows themselves. The job of positioning and resizing applications is a confusing mix of responsibility between the user and the system.
Once you've switched to tiled window managers, examples like these sound like Stockholm Syndrome.
I hate tiling window managers. After I start a program, I move and resize its window to the perfect position, and it stays there for weeks. I don't ever want it to be moved or resized automatically, which is what tiling window managers do by default.
I will offer that you can resize/move/float in most tiling managers. Remembering your modifications is usually possible too. It's the default behavior that separates the experience.
I can't see a practical world where the OS doesn't need to take control of window positioning in certain situations. As a core example, there is full screen. Minimize is another, but that doesn't have a clean analogue in the tiled universe.
There's a natural strong reaction folks have to window managers, because it forces you to mentally remap at such a foundational level.
I prefer tiled managers because the user offloads most responsibility. Open something and by default it uses as much space as is available. If you have a special need, you can float or resize, but the vast majority of cases it makes the right call.
At heart, it's offloading cognitive load. They're more predictable and require less faffing around.
DirectX 12, better SSD support, HDR, restoring window layout when reconnecting monitors, WPA3, DNS over HTTPS, WSL2, Windows Sandbox, per monitor DPI scaling, QUIC, dark mode/tabs/previous session/better encoding in Notepad, Windows Terminal.
All these sound really nice to be honest. The things I want my OS to get.
The only problem is: these are the improvements what this OS got in the last 20 years, not last year. Microsoft is a $3T company.
Wine is more like emulating Windows API behavior on Linux, while WSL is Microsoft throwing their hands in the air and saying "Lets just VM Linux wholesale".
Both aim to avoid Windows, neither replace Linux but instead tries to move more to Linux.
>Both aim to avoid Windows, neither replace Linux but instead tries to move more to Linux.
I don't agree: WSL is an attempt to use programs developed for Linux in Windows. It is clearly for people who want to use Linux programs but don't want the headache of setting up Linux or dual booting.
> WSL is an attempt to use programs developed for Linux in Windows.
Then I'd think it be available as a "right-click > Launch Linux Program" or something like that, like WSL1, rather than the VM approach WSL2 takes which gives you entire environment. Even Microsoft themselves market WSL like that:
I agree with your last part though, it's for people who want to use Linux without the headache of dual-booting or managing their own VMs, so they use predefined packaged VMs ala WSL instead.
I guess I was more contesting that WSL is for people to get away from Windows, when it is actually the other way around; it reduces the friction between tools developed to only work on Linux and Windows users, so that the Windows user can stay using just Windows. Back when I used Windows, this was always a point of contention for installing most dev related apps, and trying to use MinGW was such a pain (WSL was broken on my computer then due to Hyper-V being BIOS disabled). I used Linux now on my main computer, but I recently tried WSL on a family member's computer and I can see how if you just do all dev work in WSL, you would never have to go through the process of migrating to an entirely new OS and still get all of the benefits.
If I run WSL it’s because I try to avoid Linux - but I want to run something that needs a Linux environment. I think the argument about what’s avoiding what is pretty strange.
WSL1 seemed great, until it wasn't. Then WSL2 came along, which is just a VM and works identical to VirtualBox et al. Still huge hassle to deal with various things that get confused when you run it in a "Linux-but-not-really-but-also-Windows" environment.
Better to just go straight to what you actually want, which seems to be a proper Linux distribution, everything just works as expected then.
I used to run all three major OS' where I saw no real difference in me using it for the apps etc I needed. As I leaned more heavily on the development side, linux kind of prevailed. Windows 8.1 and Yosemite were the last of the other two I've used for real. Never had to look back to other two since, be it for work, games or whatever.
Even occasional need for Adobe things stopped. I would still really like to see Adobe suite on linux, but if they don't want my money that's cool too I guess. I suspect the software tools people use for work is what's holding them back mostly, like Altium, CADs etc. Funnily enough, Microsoft office is just fine without OS native version most of the time.
This may be unpopular, but they made three legitimately GOOD changes to Notepad, before they went absolutely batshit.
- Dark mode
- Tabs
- "Continue previous session" (restore after restart)
But they also made a HUGE regression which isn't talked about enough:
- Previously, if you delete the underlying file, Notepad didn't "notice." Meaning you could Save to recreate it. The file existed in an ephemeral state, as long as the Notepad window remained open. Right now, if you delete the underlying file, Notepad notices immediately, errors out, and closes that tab -- content gone.
VSCode handles this correctly, it puts a line through the filename in the tab (and red-font), but doesn't close the tab on you suddenly. Ephemeral state retained.
In my ideal world Notepad needs exactly ONE new feature:
- Right Click Tab: "Copy Path."
Then just REMOVE: Spell Checker, Formatting, CoPilot, Markdown(!), and Autocorrect. Completely inappropriate functionality for Notepad, and the Vibecoded Markdown implementation already added a security vulnerability(!) to freaking Notepad. Branch off into a product called "Wordpad" and then create all of that garbage in there.
Notepad also understands line endings and text encodings now, instead of always writing as CRLF UTF-16. That eliminates a whole class of “save file in notepad, corrupting it” trouble.
Yeah, I thought it was an odd couple of recommendations. Alacritty is an awesome terminal emulator that'll run on Windows (although I'd say Windows Terminal is still pretty close to decent)
> In AWS eg. bucket can be deleted only when empty. Deleting all files first is your confirmation.
That wouldn't have helped in this case - the agent made a decision to delete, so if necessary it would have deleted all the files first before continuing.
The question that comes to mind is "how are people this clueless about LLM capabilities actually managing to rise to be the head of a technology company?"
> The first delete would fail: “bucket not empty”. This might make the agent question the deletion (“bucket should be empty”).
This is actually not a bad test case for evaluating an LLM: give it a workflow that has an edge case requiring deletion, then prevent that deletion, and see if it:
Yeah, I've run tests similar to this while evaluating gpt 5.4 vs claude 4.6
Claude is more likely to figure out workarounds and get things deleted if I tell it to delete stuff, so it performs much better in this benchmark and I prefer it.
GPT is more likely to stop and prompt you "I got an error deleting this, should I try another way?", and since the operator gets more of these prompts, they'll hit continue more withut even reading it, so it ends up being more annoying for the operator and not really reducing the chance of it happening imo.
If your workflow for your llm says "delete the ec2-instance", and the ec2 api gives back "deletion protection is on", I want my llm to turn off deletion protection and delete it.
I feel like you're implying that the reverse result, prompting the user, is better, but I disagree with that.
This can still be done programmatically without any kind of confirmation from aws-cli, bringing this back to, an API can (and probably should be able to) take certain destructive operations that someone’s blocked from doing in a UI, such as in your example.
My s3 buckets are backed up with Nakivo (and immutable for 7 days) just in case, and that’s just to protect me from myself and my s3 provider either failing or deciding they don’t want to do business with me anymore for some arbitrary reason. I’m not even turning an LLM loose on it.
I don’t see the problem here. These people will be pushed out of the industry quickly and their business taken by other people, who are using agents, but are smart enough to run them sandboxed without any permission to production or even dev data/systems.
reply