Dallaswho

I’m having to use Google to search Amazon. Looking for 696 2-gang masonry box today. Amazon shows trash. Google finds 20 at various prices right away on Amazon.com.

Yesterday I was looking for the guts of a Polaris reverse valve. I put in the exact part number, nothing close. Google found 10 right away on Amazon.

Nobody wants semantics when looking for things. This is garbage.

Imgur.plen.io takes an Imgur link and makes it a raw image link.

Exactly. Web2 set the internet back so far that we finally have a perfect use for cloud resource sharing and people don’t trust it.

I don’t care. I send free openrouter models pics of me naked almost every day.

Not a gamer but if you’re not getting a dedicated GPU, the integrated GPUs on mobile chips are about 4x larger than what is typically found on desktops. A little Core Ultra H mini might fit the bill without making the office look like a kid’s room or making noise or heat. AMD might have something like that too but I wouldn’t know.

Your hardness isn’t an issue. It is an issue that your alkalinity is way too high for that hardness level and pushing your pH to the moon. Higher hardness means Ph should be 7.2ish but you can’t keep it that low with too much baking soda.

CYA is higher than I like it but maybe that’s ok for salt pool.

I don’t see any reason to drain. One treatment of phosphate reduces can bring 700 to zero if you even care about that. Calcium isn’t a pool enemy, it just requires alkalinity be a little lower or else you go through a lot of acid.

Lots of shiny buttons. Dahua integrations are basically worthless but going through frigate yields 600ish sensors and endless ability to add more like classify if gate is open or closed or trash is out or what time mail came or where each dog was seen last.

I only tried Dahua integration to activate white illumination only when someone was in my driveway and then change max exposure to 16ms but it couldn’t even do one of those things right. Had to use rest_service instead.

Define compatible. Most PoE cams will share a stream or two or more. Some share voice and audio channels. Modern cameras share detection metadata, face recognition, zone crossing, and most features with any pro grade VMS or NVR and are 100% interoperable. This is called profile M.


Look out for retail brands though. These are VERY stingy with their features. Ubiquiti is probably the absolute worst. Apple-like walled garden. They share nothing with third party equipment. You can technically pull a sub stream but you can’t configure it and it’s a shitty VBR stream that’s no good to run tools on anyway. Same with main, plus you need their NVR to pull either, just like the analog days. Reolink is in a very similar boat allowing streams out that are only lightly configurable and generally don’t meet requirements of third party systems. If you already have reolink, unfortunately, your best bet is to stick with them.

If you don’t mind running a server, there is an entire software ecosystem built around making Reolink suck less with other equipment. Check out go2rtc and scrypted. They can turn Reolink streams into solid interoperable feeds.

Ya PVE was super great when I ditched the big gpu machine and went to minis, I just threw the hard-to-migrate, non-accelerated VMs right onto an old nuc. It took just a few minutes.

Problem is a year later I was still running this loud old nuc8 up in the IT area. It took the ram apocalypse to finally motivate me to migrate off of PVE because that old nuc had 32GB ddr4 and I got double what it used to be worth within 3 hours of posting it.

I mean if you’re using a mesh system, and it’s not wired, there is a much higher likelihood that the problems you experience are manifested by the local network as opposed to a service provider.

I’d at least step up to air with Apple. M3/m4 is fine. Don’t need m5. Neo is crippled by a 8GB RAM on a 64bit memory bus but only heavy media, gaming, and AI are really going to be affected. I wouldn’t want it but it’ll probably be fine for school work.

If she’s doing data analysis on an Neo, even with python, she’ll be limited to a couple million rows and have to get really good about deleting unused frames.