This is completely incorrect.

Microsoft do not concern themselves with what SMB versions Samba supports when considering maintenance. At all.

As it should be IMHO. We match current versions of Windows and only keep SMB1 around in an "off-by-default" state for customers who can't or won't update old Windows / DOS clients.

Elon promised us a Vactrain

When? Where?

He has talked about "Hyperloop" but I don't recall him ever promising to build it. I guess with Tesla, Boring Company, SpaceX, Neuralink, X/Twitter, and xAI, maybe he has enough going on at the moment.

I think the average canal would be a LOT longer than 10 miles.

I guess this is the disconnect between your thinking and mine.

I'm not imagining that the microinverters would be enough to drive the power all the way to the end of the canal... I'm just imagining wire runs some reasonable distance to a backbone power distribution network that would involve transformers.

As little as I know, I'm still pretty sure you can't have one transformer per panel. You will need to have a few strategically-placed transformers, and you just need the wire runs to make it from the panels to the transformers. The transformers would be hooked up to distribution lines. Again, since I'm not an expert, I'm imagining this basic power distribution backbone to resemble the power delivery lines system in my neighborhood. Presumably at 4 kV or higher.

Microinverters, a few thousand feet of cable, step-up transformers. Assemble together.

Alternatively: strings of panels, string inverters, a few thousand feet of cable, step-up transformers.

I am not any kind of electrical engineer but I think it's just common sense that if you want power to go literally miles you are going to need AC stepped up to higher voltages, for the very reasons you listed in your posts.

If I were put in charge of this project, the first thing I would do is find someone who is an expert in this stuff and discuss a sensible basic strategy. If you compute that it's infeasible to send power ten miles at 240 Volts, then... don't plan to do that.

I'm not an electrical engineer, but I think you aren't making any sense here. Everything you said argues for AC power.

DC power is only good for long-distance transmission at crazy high voltages. DC is also okay for low-power, short-distance tasks. For all medium to long distance power transmission at non-crazy voltages, AC is the best choice. It's the whole reason why AC won over DC for our power infrastructure.

https://blog.ucsusa.org/mark-specht/the-current-war-why-did-westinghouse-ac-beat-edison-dc/

Ya, because 12 foot diameter copper cables are cost effective.

Obvious hyperbole is obvious.

Enphase microinverters make 240 Volts. I looked it up, and with ordinary Romex wiring 240 Volts can be run 1000 feet with a voltage drop of under 6 Volts. It seems likely that the canal setup could have a distribution system that would work; it might involve step-up transformers, I wouldn't know because I'm not an expert on this stuff.

I presume you aren't literally claiming that 12 foot diameter is needed. Please feel free to use your knowledge of Ohm's Law to compute actual wiring sizes needed and show the problems with using 240 Volt microinverters.

Enphase microinverters double the price of your panels as they cost as much as the panels themselves, and are paired one to one.

I just checked prices and found that microinverters were around $180, and there are definitely some solar panels that cost more than $180. Also a few that cost less than that.

I checked prices on string inverters, and found some as low as $200 and some $1000 to $1500. I don't know enough to know whether the $200 one is cheap junk or the $1500 is overpriced or what.

But I stand by what I said: if the patents are finished, and companies can manufacture microinverters without the overhead of paying royalties, I expect mass-produced microinverters to be adopted as best practice generally.

There's a reason why transmission level power lines are over 100,000 volts.

According to answers on Quora, neighborhood power lines are from 4 kiloVolts to "tens of" kiloVolts. 100,000 kV would only be necessary for long-distance power transmission.

All modern panels have bypass diodes to stop a single panel, or group of cells in the panel, from dragging down the whole chain.

I hadn't heard of this. I looked it up:

https://www.electronics-tutorials.ws/diode/bypass-diodes.html

If I'm understanding correctly, low-performing panels are cut out of the circuit so they can't bring down the output from other panels, and also so that current can't flow from the high-performing panels and fry the low-performing panels. This sounds like a good idea, but doesn't sound as good as microinverters, which let all the panels contribute as much as they can make with nothing pulling anything else down.

If you have a better web site for explaining bypass diodes, please share it.

they'll have to use many small DC-to-AC inverters instead of a few long ones to manage the transmission losses.

When I got solar panels, I got a system with Enphase microinverters. There is one microinverter per panel, plus one microinverter per battery module. My home has dozens of microinverters.

IMHO it's clearly the way to go for a project like this. If a tree limb starts casting shade on one part of the canal, you don't want a whole "string" of panels to all have their output drop. See the section "Efficiency in partial shade" in the below link.

https://www.solarreviews.com/blog/pros-and-cons-of-string-inverter-vs-microinverter

Also, long runs of relatively low-voltage AC are safer than stringing together panels into a high-voltage DC run.

I'm not sure when the patents on microinverters will run out, but the first microinverters appeared in the 1990's so at least some of the patents must already be expired. IMHO microinverters really are a better technology, so once microinverters are free from patents (or at least free from the most important ones), I think they will become the standard for solar power.

Just to add, for this reason I'm only helping out with the C coding, not leading the effort :-). The person in charge knows his stuff !

It's a fair cop :-). I've only started FIPS work in earnest since I got laid-off from Google and joined CIQ, so I'm not cognizant of all the correct terms.

Yeah, it's important for customers though.

I agree with nearly everything you said. I would have tried to word things a bit more politely, but you aren't wrong.

In fairness though there is one point to discuss:

Citizens have taken to sabotaging them in protest.

I saw a TikTok video explaining how to sabotage them (step zero, steal a traffic cone; step one, put the cone on the hood of the Waymo car; hey presto, the car thinks it can't drive at all). That video gave a reason for doing it, and that reason was that self-driving cars are bad because everyone should be taking mass transit all the time.

So in fairness, some of the people sabotaging the cars don't care if the cars are working correctly or not. They believe that nobody should be allowed to have personal transportation, and everyone should be forced into mass transit. This is an axiom with some people.

Even having something like RADAR as a "don't crash into the big object in front of the car" type failsafe would be a good idea.

Tesla used to have a blog posting up called "Seeing the World in RADAR" that talked about the problems of seeing with RADAR. For some reason they removed this blog posting so I can't give you a link.

This blog posting documented some of the weird issues with RADAR. For example, a discarded aluminum soda can can appear to be a giant hazard right in front of the car, because the shaped metal dome in the base can reflect the RADAR beam back perfectly if the angles are just right.

Elon Musk has said many times that RADAR was a problem because the cameras would give one input and RADAR might give a different input, and it was a problem to reconcile the two. (For example, the cameras don't see a giant hazard, and the RADAR sees a discarded soda can.) He decided that it would be best to go 100% vision-based, and I think history will record that as a reasonable decision.

On Twitter, Elon Musk has said that only high-definition RADAR would be useful. The old RADAR that Tesla used to put on their cars was not HD.

The new "Hardware 4" sensor suite does appear to include an HD RADAR unit.

https://www.teslarati.com/tesla-hardware-4-vehicle-teardown-reveals-radar-module/

I'd willing to bet most accident reports contain something like the phrase "I didn't see them"

By the way, one neat thing about Tesla's self-driving is that its cameras are mounted on the exterior of the car. I sometimes don't see an approaching car or pedestrian if they are behind the "pillar" of the car. But the FSD hardware includes cameras mounted on the pillar, on the outside of the car, so there literally isn't a blind spot.

Tesla already has issues with their cars rear-ending others since they removed the front radar, as they simply can't recognize certain types of vehicle.

I'm gonna need you to provide references.

There have been documented cases of Tesla cars, driving themselves, hitting things and killing someone. To my knowledge, most or all of these cases involved very old versions of Autopilot, possibly even "Hardware 1" (which only had a single front-facing camera). And all the ones I know about involved the car hitting a non-moving object like a fire truck or a semi-truck trailer.

The old self-driving had problems telling whether an object in front of the car was actually in the road or not. If the car was heading uphill and there was a sign above the road, the sign could look like it was in the road so the car had to assume that not everything it saw that appeared to be in the road was actually there. (It had to, because panic braking for no real reason is also an unsafe thing to do.)

Now, Tesla self-driving decomposes the problem into two parts: first, a neural net takes the 8 camera inputs (360 degree view around the car) and builds a 3D vector space model; then, another neural net uses the 3D model to make driving decisions.

Clearly a 3D model, if correctly constructed, means the car can tell whether something is actually in the road in front of the car or not. And the car shouldn't be hitting anything, whether it can tell what kind of vehicle it is, or not.

I drive a Tesla that has the Full Self-Driving beta. At this point, I trust it not to hit anything; I don't trust it to correctly follow all the rules of the road or perfectly navigate the roads I drive on. I'm definitely not claiming it's perfect. But again, I trust it not to kill me or others.

Everyone else greatly simplified this task with lidar and radar.

Everyone else has also greatly simplified this task using high-resolution maps, with the result that their cars can't drive anywhere that isn't covered by an HD map. Only Tesla is trying to solve the full, general problem.

Are they planning to have an AI driver that is trained on the rules of the road?

Yes, I'm pretty sure that's what they are doing. When my car drives me on the freeway, it will often use the passing lane to pass other cars, but then it gets out of the passing lane. It actually shows me the message "changing lanes to get out of passing lane" (quote from memory and may not be exact). This must be a trained-in behavior.

I don't know if FSD beta was given a rule just for Washington state, but in Washington it's actually in the state law that the left lane is the "passing lane" and you are supposed to get out of it if you aren't actually passing.

But the carpool lane doesn't count as a passing lane; you can hang out in a carpool lane. And, my car doesn't try to leave the carpool lane.

FSD isn't perfect but I can see progress.

As Elon Musk said: "Tesla makes the impossible late."

I did enjoy visiting a store called "Book Off" that had used and remaindered books and videos. I got a DVD of a Pixar movie in Japanese for like 600 yen.

I will check out Do It Yourself!, thank you for the suggestion.