I almost exclusively dev in containers as is, cant really imagine letting some AI model run free on bare metal no matter what claims of guardrails it might have.
Thats more or less what I did, and nix just made sense for the job. For 99% of people I'd say no its not worth it to tinker, just go with opnsense virtualized so you get at least some the benefits of the better linux drivers. By that I mean NBASE-T on various intel chips and while intel's igb is fairly solid on unixes many other vendors' drivers are less so.
However if you're willing to figure out configuring per your needs you definitely can get a lower latency router with all the same capabilities and more, with it's components more sanely isolated via containers.
I’m inclined to believe in a future a bit less binary and more MR in general. Devices that can handle a variety of experiences from 1% augmented to 100% virtual.
The average person didnt really care what tech was involved, they dont romanticize software in the same way as tech inclined people do.
People hated it when apps were glitchy, when it wanted "constant" updates, or how they couldnt share a page because the entire site was some bloody flash applet.
iirc theres ones that do. However I dont recall there being any clean fix to the amperage constraint issues. Especially when a lot of usb-c chargers will vary output as they heat up with usage.
Which is kinda part the issue, usb-c charging bricks, they aren't usb-c power supplies, there is no expectation of sustained output capacity. Thankfully at least some the multiport ones have renegotiation more or less solved cleanly rather than what is essentially rebooting the PD controller.
I mean, nginx dang well should? This is just an incredibly synthetic http(s)/1.1 test for what its worth.
Like you totally could turn off garbage collection for caddy especially since this is only testing incredibly short single response queries that would never need GC. Shockingly you would actually get better performance than either nginx or zeroserve, but like the uselessness of this benchmark it'd mean nothing to the real world usage of these web servers.
Systemd still has to have one of the greatest ratios of "genuinely helpful software to haters" of all time. I do NOT miss all the bodge of older init systems.
And as if avoiding AI is something exactly difficult in the lands of non-proprietary OSes.
If it gets people to enjoy photos they took even if transformed, I'd rather it over the alternatives of people just entirely generating slop altogether.
I know my way around photo editors, someone being able to easily remove a water bottle is the least of my concerns. Most of these are tasks few genuinely mourn having a required level of skill to do, and by large the vast majority of edits people will actually do with this are still relatively simple to anyone versed in photoshop.
Much as I have a not so great opinion on Siri's capabilities, I'm rather surprised how many people appear to use Siri/Apple Intelligence to search for rather niche hobby content that I run a site for. OpenAI's scrapers I expect volume from, but I didnt really expect apple's to be consistently rank second.
What if they like photography and have a need to both focus up close on the camera's displays, but also out to the distance?
Probably many such cases similar.
Cutter does work on the H series without laser upgrade, their marketing is a bit unclear. What you dont get is the "birds eye camera" to scan the work surface and overlay cut out locations onto in software, you're flying a bit more blind sorta speak relying mostly on measurements for many materials.
Printed stuff adds locating targets that the nozzle camera scans for. Although it doesnt scan too intelligently, just somewhat within the expected region according to paper size rather than letting you place it anywhere on the bed. Not sure if the "birds eye camera" changes that much.
I guess your phone takes the place of that camera for the A2L, probably makes more sense for the average person who cares for such a feature.
Theres a bit more to that story, leading up to Vista Intel's iGPUs lacked proper DX 9 support, just partial. Particularly 910/915 were marketed as "Vista Capable" but not "Vista Premium Ready".
And the later ones which did have full DX 9 support really weren't that much better. I recall drivers being an absolute mess for a lot of external peripherals too, many that weren't even dated took ages to get something usable for vista if at all.
Best "EDC" carry blades for me have all been 3d printed using reliable ol utility blades.
I dont need no survivalist tool or fidget not-toy, just something compact, always sharp, and way cheap in case anything ever were to happen to it. Sub $1 cheap.
Negativity bias imo. We all hear about the failures associated or attributed to AI. I say attributed as I'm curious how many whether LLM generated output or not get prescribed away from being any one individual or team's fault.
Of course we see larp success stories, marketing fluff promising the sky, but genuine realistic success stories in devops relating to AI are rarely exciting to read or write.