Tesla and just about everyone else in the "autonomous" driving game is using an Expert System. This isn't AI, it is something dug up from the 1970s that sort of modeled what AI could do. Someday.
Well, someday isn't quite here yet. There is no underlying intelligence to these things. It is all based on rules and if you get to the bottom of the list of rules, the car has no idea what to do. This is freaking dangerous.
A true "AI" would have some default precepts, like "don't crash" and "don't hit people". These precepts or basic concepts do not exist in these vehicles. So when something unexpected comes up, there is no rule and no action. So when the roadway presents a situation that isn't recognized, the car will do something unpredictable - or it will do nothing and just keep on going.
Why can't the cars simply be programmed to brake gently to a stop if there is a condition they do not understand? That would require defining what "understand" means. An Expert System has no real intelligence, it is just a series of rules that say when you recognize the conditions of X then do Y. So a basic requirement of this is to load up the computer with every possible scenario that can happen on the road. Anyone that has driven a car for more than a week will tell you that is impossible well, maybe possible but it would take a huge amount of work.
Google had a car that drove into the side of a bus. The response was to add 3600 new rules to the package. Does that illustrate the scope of the problem?
Having a fleet of autonomous vehicles, such as what Uber is thinking of, is foolish and dangerous. Driving on the road with one of these present will present and unlimited capacity for chaos because if something unexpected (or unprogrammed) happens, the car will do something unexpected. And that could be dangerous to everyone around.

Well, uh, I am pretty sure you are wrong.

People are social animals and not technophiles. Most of the behaviors associated with technophiles are distinctly anti-social and this comes from a variety of sources. For example, for every one source in popular media that extolls the wonderfulness of a technophile, there are 10 that make the technophile out to be the stereotypical nerd and an object of derision.

Who wakes up in the morning with the thought in their head "I want to be less likeable and more an object of derision." Nope, sorry, technical abilities and skills are a passing fad.

What is going to happen, clearly, is that the technical requirements - especially very picky detail-oriented ones - are going to all but disappear. Complex passwords, password vaults, using lots and lots of different passwords, etc. are all going to disappear. The reason is simple, it would take someone that would be easily classed as a "nerd" to manage such a life and people intensely do not want to be nerds nor thought of as nerds.

What we will do is to simplify the whole problem. How about a bracelet that reads biometric data (to insure it is worn by its owner) and is your "passphrase". If you take it off, it has to be reset so it cannot be stolen. It probably should not use any protocol that can be intercepted. It can be an extremely complicated device but the key is that it is simple to use - you just stick the bracelet in a loop that reads it and it works for everything.

OK, maybe we have implantable chips instead. Or a badge that runs off body heat. Something. It can be a pretty expensive device because the alternative is getting left out of everything or having stuff stolen. It should clearly be something that cannot be stolen and reused, and obviously it should be something that is very, very difficult to lose.

OK, think up some nerd-proof device that the average Joe can use, never lose and isn't subject to being stolen and being reused. I'm sure it is possible and if the interface to read it is cheap enough we can start with putting one on every PC and every ATM. Then use it for credit card data and get rid of all the cards. Because it IS in fact your identity it can also replace a driver license and likely replace a passport as well.

Trust me, nobody is going to want to walk up to an ATM and look up the proper 17-digit PIN on their phone password keeper. Moreover, they will not do it. Something else will happen.

There is a product called Vindex which is specifically designed for "indexing" video. It intelligently collects up "events" and displays a gallery view (bunch of thumbnails) representing these events. By "intelligently" I mean that it isn't just looking at raw differences between frames but does considerably more math to eliminate false positives.

It can also process video at up to 60x normal playing speed, which means analysis of a night's worth of video takes a few minutes. Compared to just playing at high speed you don't miss very short events which are way too easy to miss with high speed playback.

No, it isn't open source. There is a free evaluation available at www.infinadyne.com so it is easy to check out. There is a new version in the works which will add video enhancement.

What you are describing is a situation where "demographic" data doesn't get sold. The origin of this in large part insurance companies because they had an interest in collecting as much data as they could for actuarial purposes. If you can notice that people that own motorcycles tend not to live as long as people that own rickshaws you are going to charge more for life insurance if the person owns a motorcycle, right?

It got noticed that there was some marketing use for this data as well. Maybe if you can see that a more affluent zip code has a higher percentage of DLink wireless routers whereas a less affluent zip code has more Belkin wireless routers this information can be useful to all wireless router companies for targeting advertising both online and offline. So suddenly there is a value in collecting the MAC address, zip code and enough information to identify the manufacturer about wireless routers nationwide and making this information available.

Sorry, but I think you are maybe 100 years too late to stop the sale of important demographic data. And at least 50 years too late to get people to understand that their information has value and should not be available on the open market.

Why would you think the company has any obligation to keep information that they own in their records secret?

Understand it is the company's payroll records, not the employee's.

The facts about utilization of healthcare aren't owned by the employee, either. Now, what the reason for the utilization are likely covered by HIPPA. I would say this is especially true if the company is self-insured as most large companies are today.

Unfortunately, what the Interstate Highway System gave us was unregulated trucking, which in turn pushed the railroads out of the freight business almost completely. With no need for rail freight except in special cases, we have ripped up much of the rail network and built suburbs over it. Most of the passenger rail lines were sold off to the freight companies as well.

People are now thinking it might be nice to take the train somewhere, only to find out there are no rails left. And putting new rails in would require demolishing lots and lots of houses and streets.

Was all of this planned?

A nationwide wireless Internet service might be a good thing. It might provide some competition to Google who will certainly offer such a thing if allowed to - and use it for ad revenue and collecting demographic data to sell. I do not like the idea that all Internet access will be controlled by the government and Google in a few years, but that is exactly where we are heading.

Easy. Instead of using microwave transmissions, just use lasers. 300 million tiny lasers on a satellite to follow and track each receiver. Well, maybe more like 1200 million - your car, your phone, your MP3 player and your home computer. Might as well just go for 2 billion of them, just to have a little extra capacity.

Of course these would have to be more than a couple of watts each to punch through the clouds and such.

Now that I would like to see... a satellite with an output of maybe 20 gigawatts.

What is the function of an ISP in such a model? Wouldn't they simply be a hub connecting users to the Internet at large?

I think the function of the ISP as a supplier of email services is pretty much dead. Certainly, the ISP as a provider of local content is dead. The ISP running a "portal" is dead.

My ISP provides a connection, and that is it. They do not provide search services, local weather, indexing of web sites, or anything else. They are certainly being bullied by content providers to add equipment to provide local caching of high value content and removing that would be a boon to everyone.

The ISP today is also the one maintaining the physical plant, be it DSL links or the cable TV/Internet system. If the "last mile" is taken over by some state, county or local government they are going to have to either take that on or pay someone to do it. Every ISP that I have ever encountered that outsourced maintenance of the physical plant to someone else failed spectacularly - because the maintenance company didn't care a hoot about customer service to customers that were not theirs. The brief history of "open" DSL connectivity where the ILEC was providing the wires and someone else was the ISP shows this clearly - the ILEC did not maintain the lines, partly because in most areas the regulated fees for access didn't pay enough for the ILEC to do any maintenance.

Do you think your municipality could or should take on a new engineering department to maintain the "last mile" connections?

How about customer service? You know, the folks that when you call tell you to unplug the modem and plug it back in. Outsourced or brought into a municipality?

How long before such an arrangement results in the State Bureau of Internet Connections because waste and duplication in every municipality makes it an easy target for some state representative?

There were no CD recorders in 1990. Around 1993 the first CD recorder was introduced but it cost around $50,000. 1994 saw "consumer" CD recorders for as little as $1000. It took until 1997 until they dropped to around $350 for home CD recording to take off.

I worked for a company in 1994 and 1995 that was involved in making CD-R discs for people because we had one of the $50,000 machines.

Olympus LS10 or LS11. 48Khz stereo PCM with line inputs in a handheld device. I strongly recommend using 32GB SDHC cards as these files get pretty large.

Yes, digital recording exists today, but it isn't a consumer product.

CD-RW and DVD+/-RW discs are not "photosensitive" in any regard. They are written to by heating spots on the disc with a laser - heat, not light. The material in a rewritable disc is a metallic alloy which has two stable states - amorphous and crystalline. By heating and cooling the material in different ways (slow or fast) the material changes state.

I will agree that CD-RW and generally all RW forms of DVD media are not long-term storage. It is unclear exactly what is happening in the disc but what it appears is some of the higher-energy state crystalline material spontanously reverts to a lower energy amorphous state. When enough of the written to spots do this a sector is unreadable. There are around 350,000 sectors on a CD-RW so this process can take quite a while to render a disc completely unreadable, but it can happen.

One problem with more recent discs is the alloy layer has been made thinner and thinner to be more responsive to laser energy at higher writing speeds. I have CD-RW (2x) discs from 1997 that are perfectly readable. I have seen discs from 2004 that were completely unreadable six months after being written. The quality of the disc, the thickness of the alloy layer and the storage conditions can all play a role in how long a rewritable disc will remain readable.

In 2010 I was able to recover files from a number of DVD-RW discs that were recorded in 2003, so long term storage is possible. It needed a better-than-average drive to read the discs, but they were readable.

There is some indication that strong UV will affect RW media although this is not any sort of photosensitivity but instead it is an energy transfer - the UV putting enough energy into the alloy layer to change crystalline spots into amorphous. Sunlight is going to take a long, long time to do this. I have heard of people "recovering" RW media with a EPROM UV light, however.

You are depending on the people to vote intelligently. This is like expecting a child to intelligently walk away from the feast of candies and cakes. Everyone in the US is looking for the candies and cakes and how that affects other people, people working and paying bills, is of no consequence.

The "official" government unemployment rate is 7.9%. That considers only people actively looking for work right now. If you count the people delivering pizzas with college degrees and the people that just gave up looking it is a far, far larger number and one that is difficult to pin down. Some estimates put it as high as 30% of the population. That means there could be as many as 100 million people out there counting on handouts, candies and cakes, from the government.

Nobody is going to get elected without promising big handouts.

The people that have a job aren't interested in much other than keeping their job and getting their debt paid down. Less than 60% of the people in the US voted in the last election - probably because they were not interested in the size of the handout going to other people.

Those minstrels got paid - and generally paid pretty good. They got to eat while many of their fellows were starving on the little plot of land they were working for the Earl or whatever. And they wandered from town to town because no town could afford to keep them very long.

Make no mistake, these folks were living pretty high compared to the rest of society in those times.

Sure, you can make your own music for yourself. Don't plan on selling it, though, because everyone else can make their own music also. Or, you can listen to other people's music for free - just pay the membership fee for the service and you have your choice. Of course, not even the streaming service is very profitable, much less the artists - there is no money in it anywhere.

Music for the last hundred years or so has been driven by promotion. You hear about it because people are paying to make sure you hear about it. There are (were?) magazines dedicated to music promotion. AM and FM radio have been driven because of promotion. Free concerts have existed because of promotion - where the artist gets paid but nobody paid any admission. This is all coming to an end and the end of the road is no more promotion - you hear about what you hear about and you don't hear about anything very much.

Maybe it is a more eglitarian form of entertainment, but it means the end of things like a common cultural reference. A band is never going to escape their locality, which might be geographic or it might be a very narrowly focused group of people, or both. It means that you can never talk to someone that you just met about a band you both have heard of.

Why would you purchase something that you can select to stream to a mobile device any time you want? Seems silly to me.

What streaming has done is given the power of the radio "request" to everyone and they all get their requests instantly. No need to buy anything, just make your selections and listen.

Oh, and if you want an MP3 file (for some unknown reason), that's what BitTorrent is for.

So who is getting paid here? Well, the streaming service is getting something, either a membership fee or ads. Both are a pittance because nobody is going to pay a high membership fee and the ads aren't generating lots of sales so they aren't very valuable. With that, the streaming service can pay the artists something - something like $0.0042 a play or about $500 a year.

There isn't any money in it. And there isn't anything that can be done to somehow push more money in or take more money out.

The problem with the "up and coming" band just getting by with touring is that there is no "coming". They might get enough exposure to net a better grade bar or two but nobody is paying for promotion. They are probably lucky if they can afford to stop off at Office Depot to run off some flyers to pass around. What the record label is for is paying for promotion - and making money by backing a few successful and a few more unsuccessful bands. They have seriously fallen down on that trying to control their risk, just like the rest of businesses today.

No risk = no reward. But that formula hasn't been taught very well in MBA class.

What "publishing" in general was 50 years ago was taking on 10 things, be it books, bands or whatever and promoting the heck out of them With reasonably good selection up front you had something like 7 flops, two moderate successes and one pretty good performer - which altogether paid for the promotion of all ten with some profit left over. The problem is the MBAs came in and decided they could make more money by getting rid of the seven flops without considering how you do that. So we have the entire spectrum of "publishers" trying to find the three successes without encountering any flops at all. Lots of really smart (and successful) people figured out a long time ago that you can't do that reliably and this lesson is being relearned every day. Unfortunately, MBA schools taught that you just had to be smart enough to find the three successes and all would be good. We are experiencing what happens when this is being applied across the spectrum of publishing - books, movies, music, software, magazines, etc.

The Dish Hopper is somewhat of a joke, It is a way to convert satellite broadcast into streaming - you see, it isn't a DVR at all but a device that requests something be saved for you at Dish Network HQ. Then, later you can have it streamed to you over the Internet. They claim the device is limited to 2000 hours, but this would appear to be an entirely arbitrary number. Since your "saved" content is likely shared with everyone else, why would there be any limit at all?

Do you really think that they are saving a unique copy of Two Broke Girls rather than simply having one that everyone shares?

Unfortunately, it is going to suffer the same fate as all streaming - congestion. We are starting to see streaming degrading because of Internet congestion now and it is only going to get worse as time goes on. Having a faster link from the "head end" to the home isn't going to fix it as long as we have a node configuration where a node feeds a neighborhood - both FIOS and every cable and DSL system utilizes this sort of configuration.

Cox in Phoenix is trying to be forward looking and reducing the number of homes per node from 1000 to 500 and that may help somewhat. But with higher and higher bandwidth expectations (see Netflix recent announcement), once we move into a point where streaming is being done by a large number of households it would have to be more like 100 homes per node - and that isn't going to happen without major restructuring. Major, as in when we moved cable from RF to digital distribution.

Most other cable networks are at 1000 homes per node and maybe 1Gb feed to each node. That means if homes are hoping for 10Mb/sec streaming only a 1 in 10 is going to get it. When we get past 1 in 10 streaming, that is about the end for streaming as a distribution technique.

So how long could the Hopper possibly last? Maybe three years. Maybe. Converting from satellite broadcast to streaming is a silly thing for Dish to be doing as there is no impending collapse of satellite distribution. Sort of like Netflix dropping,or thinking about and quickly forgetting about dropping DVD distribution.

I have three Roku boxes and an Apple TV box. I expect them all to be paperweights in 1-2 years.