In the SBC/SOC world, Free DOS and MS DOS are king.
Windows 95/98 are trivial to run on a 486 with full DOS support, but isn't common due to the RAM requirements. Windows compact still works fine.

Anything with an NT kernel would be both surprising and very specialized indeed.
At that level the more common options are Linux, QNX, OS/2, and VxWorkx instead.

That said, you're quite right that this will have little to no impact.

Embedded systems are still manufactured with the 486 core. It's an easily implemented core and licensing it is surprisingly affordable.
However these systems are rarely network facing in any way a kernel security patch would be helpful.
We already custom built user-lands since their purpose is very different from what desktop distros aim for.

Perhaps you're thinking of the "workgroups" flavors of client OS?
"Workstation" was used somewhat interchangeably with "Client" starting with the NT kernel.
The head developers of the NT kernel were previously DEC engineers, where the client/server model used in the PDP/VAX world was long solidified before Microsoft existed.

I still have a few boxed copies of NT 3, 3.5, and 4 up on my shelf as mementos.

Each one has a number of fancy certificate style papers inside that are titled "Client Access Licenses", or CALs for short.

Some are called Device CALs that license device clients, aka windows desktop OSes, to access the NT server.
Some are called User CALs that license a user account from a client to access the NT server.

Client access licenses is still the term used today:
https://www.microsoft.com/en-u...

The PC as in "Personal Computer" the "branding" was pushed by IBM, and the idea of a computer that is personal was pushed by many companies in the old 8-bit days.
Yes, this was originally to break from the Mainframe model.

Microsoft however has been pushing for the old mainframe licensing model from the very beginning of NT.
Or at least the mainframe model of getting paid per client device and per user account.

I don't think there is any vestige of "personal" at the root of any of their software since the golden MSDOS days.

The point however is that using "workstation" and "client" interchangeably is as old as the PC industry itself, and doesn't carry any hidden meaning or is evidence of any recent changes in ideology.

Your correction is too sparse and confusing. Do you mean the one minute this month was actually during last month? Or was the minute this month during next month?

In other words this bill doesn't require filtering similarly to how the DMCA doesn't require responding to takedown requests.

I wonder then if auditing of the filters for accuracy will be treated equally to the DMCA counter-notice process too?

You mean your phone doesn't automatically background convert H.264 into GIF, Adobe Flash, MPEG2, and ANSI cursor control characters to render in ASCII art??

Mine is already uploading the augmented overlay of a frog matching my facial expression of disappointment in response.

You must have the December model, or worse. Get with the times old man!

How? It should be your router redirecting their web requests to the captive portal server.
It has no need to ask their browser to connect anywhere except to the portal directly, so there's nothing to break.

The internal server that runs the captive portal should connect to other internal servers directly, like authentication systems or the NMS that will remove the routers redirect rules or whatnot.

Sounds like a quick method to disable cellular fully.
Fair enough, since the cellular toggle in the drop down screen only turns it off for 24 hours.

I have no idea why SMS doesn't work right over wifi calling for you though.
On one hand, AT&T intends for it to work.
On the other, it's AT&T. Just one more randomly broken feature among so many others.

I have AT&T too, and at work I have to completely rely on wifi calling as the nearest tower is 4 miles away. The phone doesn't even display the cellular signal bars on the right, only what looks like an underline. It's also clearly showing "AT&T wi-fi" scrolling along at the top left of the status bar.

I receive SMS, MMC, as well as iMessage just fine over wifi calling.
My boss uses Android and his SMS's show up fine. I don't remember his carrier other than it isn't AT&T. They even show as green, just as expected.

Curious that you say you have airplane mode enabled yet are on wifi.
Airplane mode completely disables wifi as well as cellular.
I can't imagine what setting you may have toggled instead, but your description as-is simply is not possible.

There are two much larger titanium capacitors just below and to the right of the capacitor in question, each of those others being titanium cans with two different farad markings.
Those other two have all the signs of being noise filtering caps.

The one in question is appears to be an aluminum capacitor providing smoothing to the switching FETs that burnt up due to the caps incorrect polarity.
That sounds like a single smoothing capacitor.

And while we don't really know what exactly was the root of the problem during manufacturing, it is still interesting that the boards silkscreen has no polarity markings at all for this cap.
Power smoothing can be done with ceramic capacitors, which have no polarity. Those are almost never used to filter noise.

It feels like too much of a coincidence that the silkscreen doesn't account for a polarized capacitor.
If it originally called for a ceramic capacitor that wasn't polarized, a polarized aluminum capacitor *can* be substituted for the purposes of power smoothing (it would be in parallel)

It would explain a lot of this was supposed to be a power smoothing capacitor that was non-polarized, since there is no need to mark pin 1 or concern with which way around it is placed.
If somehow a polarized capacitor was used in its place, and pin 1 wasn't marked, with no polarity indicators on the board for an AOI operator to check against, it would explain why the pick and place machine didn't detect pin 1 and perform the proper reorientation of the part before placing it.
That could result in some feeder spools being one way and placed correct and other spools being backwards and no corrective rotation being applied.

This might be why so many different pictures of the same board seemed to be 50/50 which way the cap was actually placed.

Filtering caps need to be polarized, and will always have their pin 1 marked. The two can caps each with different farad values seem to be all correctly oriented in all the motherboard images he found.

I know this is all pure speculation, but it just seems so fitting on how exactly this mistake could slip through and nothing detect it until the boards get out in the wild.

This type of capacitor acts like a (very) tiny battery.
They are placed near the power/ground pins of every component and chip they can, each acting as a small reservoir of power.
When some component kicks up activity and draws a large but brief spike of power from the rest of the system, that is seen by every other component sharing its power bus as a huge but brief dip in power.

Various components handle huge drops in voltage differently, and for different amounts of time.
Their respective capacitors are there to supply constant power to (hopefully) cover the duration voltage dip, and afterwards recharge themselves just as rapidly.

It's like how all the lights in a house dim when someone in the back shed turns on the table saw.
These capacitors are akin to putting an APC UPS on each and every light bulb in the house.
Except a UPS that costs fractions of a penny...

As for not being detected in testing, a component needs a large enough dip in power, and for a long enough amount of time, just to maybe result in "some random glitch"
That glitch might not be noticeable each and every time, assuming that component even has a problem in the first place, which is triggered only in non-trivially reproducible conditions.

Remember when Scotty said he always multiplied repair estimates by a factor of four?
Most engineers also aim to add much better tolerances in their components than the datasheet will say.
If the datasheet guarantees the component works with 2.0 - 3.3 volts, that only means at 2v it won't screw up, it doesn't mean at 1.99v it instantly will cause a problem.
Nor does it mean 1.5v will actually be a problem, just that if it is you can't complain because it was never promised to work at that point.

It is 100% possible for a component to work through extreme conditions and never need a power smoothing capacitor in place at all.
But it is so cheap and so much easier to simply put one on everything purely so you don't have to worry about it.
Were talking tenths to hundreds of a penny. Throw them everywhere and forget about it.

There is a very real possibility one can just remove this cap and everything will work fine.
There's also the real possibility it won't, and the cap should be replaced.

"I am deeply sorry and am committed to learning from this situation and doing more to be the leader that you expect me to be,"

Mr Garg continued "Not your leader of course, you are all still fired. However I remain deeply committed to leading all of the money we won't be paying you directly into my yearly bonus."

Ohio here. Joined the beta in January '21, ETA was "late 2021"
Just last week got an update email, ETA is now "Mid to end of 2022"

Go look at the coverage map you linked. Ohio is 80% within one cell with no ground stations. Another 10% in another cell to the east with no ground stations.
Only a 10% slice to the north shared with PA has one ground station, which isn't in view of birds over 90% of the state.
This was the same back in January as it is today.

Now zoom out and look at all the cells without any ground station within them.
Something like 60-70% of the USA is *NOT* covered.

I'll assume you're equally incorrect about Europe and less than half coverage is there too.
That's a single digit percentage of the globe where starlink is available.
That does not "already available" make.

Once starlink gets their secondary orbit satellites up, where the low orbit satellites will be able to "hop" communications between them to reach further away ground stations, service will technically be "available" but at significantly higher latency than ideal.

There's been no updates on when that will happen either.

Or extract the carbon from her body and then take it to dinner and a movie.
It may end up with more of a mess, but some say that is the spice of life.

Our new found ability to be able to listen to an album in order, in NO way limits your ability to click shuffle.

Those monsters! Won't someone please think of the children's easy bake ovens?!?