Catch up on stories from the past week (and beyond) at the Slashdot story archive


Forgot your password?
Check out the new SourceForge HTML5 internet speed test! No Flash necessary and runs on all devices. ×

Comment Re:Some truth... (Score 1) 236

If the automated driver system in Tesla cars has caused more than 1.13 deaths per 100 million miles driven, he would seem to have it backward.

No, he has it perfectly right. To illustrate, lets use something I like to call "math"

As of the closest thing we have to actual data, fatalities in autonomous vehicles right now are approximately 4.5 deaths per 100 million miles. As you can see, the current fatality rate for autonomous vehicles is about 4 times the human rate. It seems like they are killing more people than saving so what gives? The secret is in the rate of improvement. Human drivers have shown an approximately 25% reduction in deaths per mile every decade. That means it took 40 years to go from 4.5 deaths per 100mmi to 1.13 deaths per 100mmi. As of 2014, Google has estimated that if its vehicles were allowed to operate at full speed the fatality rate would have been around 50-100 per 100mmi, which is a whopping 91% reduction in 2 years. assuming they maintain that rate of improvement for 5 more years, the human fatality rate will be about 0.988 fatalities per 100mmi while the autonomous vehicle fatality rate will be approximately 0.02 fatalities per 100mmi.

Assuming a linear relationship rather than geometric (to make the math a lot simpler without affecting the answer much), and we will get and average fatality rate for humans of 1.059 fatalities per 100mmi, whole autonomous vehicles would average 2.26 fatalities per 100mmi. Sounds bad, but what about the 5 years after that? humans: 0.8645 fatalities per 100 mmi. Autonomous vehicles: 0.01 fatalities per 100 mmi. Given an annual death toll of 35k, that amounts to an excess of 34,500 deaths per year for every year after 5 years from now. If we delay that progress for 5 years, then in the mean time an excess of 34,500 people will die for every year that autonomous vehicles are delayed.

Those number are far more pessimistic in that they assume that we are talking about a 100% conversion to autonomous vehicles right now. Since we are in fact talking about a vanishingly small percentage of actual vehicles initially, the cost in lives for the development of autonomous vehicles is practically non-existant, while the end state savings are phenomenal. Ultimately, Automakers should be given a 5 year moratorium on liability so long as they can demonstrate sufficient continued improvement in performance (say 33% annual reduction in fatalities per mile.). After 3 years it will be safer to be in an autonomous vehicle. After 10 years driving a car manually would be tantamount to drunk driving.

The question of how safe autonomous vehicles are is a stupid question, as it is obvious to nearly everyone who knows which end of an equation is which that we *will* reach a point where autonomous vehicles are safer, and the sooner we get there the less lives will be lost before we get to have the benefits.

The surest way to halt that progress is through liability stemming from the over application of our American legal framework. Our legal system almost entirely fails to consider the larger implications of any given law, in favor of the application of precedents and procedures. You could make a very real argument that a single liability lawsuit, at this stage in the game could delay the deployment of autonomous vehicles by a decade or more, which would effectively condemn over 300,000 people to die needlessly

The automakers need to be given a temporary liability shield, in exchange for which they must demonstrate continued improvement in autonomous safety performance, and after 5 years they must release *all* hardware designs including source code to the public domain so that the other automakers can implement these life saving systems properly. After 10 years it should become illegal to sell any new vehicle that does not have full autonomous capability, and after 15 years, it should be illegal to operate a vehicle manually on public roads.

Comment Re: As should be (Score 1) 365

Actually I got my degree in engineering, nice try moron. By the way that's what AI prioritization decisions are, engineering. You wouldn't understand that, obviously.

Three things:

First, if you had truly understood the engineering you were learning in school, then it would have gotten through to you pretty quick that from an engineering standpoint, one human is as good as another, and you solve the problem in front of you, not worry about a problem that can't easily be translated into the engineering domain and has very little value there anyways. Engineering is about affecting maximum results for minimum cost. The task of trying to decide which particular set of humans to save in a catastrophic event is something straight out of sci-fi and fantasy, and has no bearing on the real world. The task of framing such a decision into a structure that would even allow for a computational model to make such a decision is monumental. Those resources could be far better spent elsewhere. Engineering is all about the bottom line. They don't save people, they make things statistically safer. If you can't understand the difference, you have no business as an engineer.

Second: You will find it near impossible to convince anyone of anything when you repeatedly use personal attacks and insults where they are neither warranted nor useful. Any time spent at all in any liberal arts should teach you that you don't influence people by being an ass.

Third, if you want to be taken seriously in any adult conversation, put you name to what you say. I fully support a persons right to anonymity, but I have found that people who know their name wont be on something put very little effort into adding quality to their work, and people who see a statement made anonymously will discount it as worthless garbage (and rightly so in most cases).

Comment Re:As should be (Score 1, Interesting) 365

The question of what AI prioritizes as it becomes ubiquitous technology is a matter of life and death for some people in the future.

There are far more important things to worry about, and Mercedes is correct in their assessment that the value added work to be performed is in avoiding these situations altogether. The percentage of actual auto accidents in which there is some moral conundrum that requires a decision as to which person or persons is going to die are vanishingly small. There is almost never a situation in which the only path to safety is through the crowd...

GP is right. Philosophy is for those who can't actually *DO*. Philosophers try to make up for that lacking by trying to tell others what they "should" do.

To all those liberal arts majors out there, progress is made by the scientists and engineers who roll up their sleeves and do the work, not that wank all idiots who sit and talk about it all day. If you want some say in what gets done, then be one of the people who does the work, otherwise shut up sit down and enjoy the ride because you're going to be a passenger in your own life.

Comment Re:What does this even mean ? (Score 1) 365

That's a big "if", is currently false and will be false for hundreds of years still.

AI vehicles currently have about 4.5 times the fatality rate of human drivers per mile traveled. The AI rate is at about the 1971 human rate, We currently only have the single year of data for AI vehicles, but next year we will have two, and when we have that second data point it will be possible to determine the precise rate of improvement. For now however there is evidence to suggest that the rate of AI fatalities for Googles self driving vehicles in 2014 would have been approximately 25 times the current human fatality rate had the engineers permitted the vehicles to perform across the entire spectrum of AI driving that is performed to today. That is equivalent to approximately the 1920s level of human driver performance.

Put another way, it takes humans about 4 to 5 decades on average to effect an 80% reduction in auto fatalities. AI has demonstrated that same improvement in 2 years. At that rate of improvement, it will be only another 2 - 3 years before AI drivers become safer than human operators, and less than 10 before a human driving a vehicle on the public roads will the the statistical and moral equivalent to what drunk driving is today.

Comment Re: Samsung is starting to behave like Tesla (Score 5, Informative) 110

Over here in facts-based land, the Note 7 has a MAX77838 keeping track of power.

The MAX77838 is a power management IC. It claims to have some other circuitry for battery management, but since MAXIMs website does not acknowledge that part number, we have no way of knowing what it really has or doesn't have. All outward appearances would suggest that it is a custom chip for Samsung (probably used in several of their product lines). Personally, I expect that the MAX77838 is similar to the MAX77829 (PMIC + single cell Lion charging circuit). This would make some sense since it looks like Samsung elected to use an external PMIC, and since they had to have one anyways, getting one that had the charging circuit built in would not be that big a deal. Unfortunately for Samsung, the charging circuit has to be relatively tuned to the specific battery being used. Generically designed Lion chargers have a habit of failing. (So much so that Tenma actually ships many of their battery chargers with a fireproof pouch to put the battery in while charging it...

This just goes to show how stupid Samsung is for designing it this way. Since the fault lies either with the Custom Maxim Chip, with the battery itself (or a mismatch between the two), Samsung has backed itself into a corner. They cannot just replace the defective Maxim Chip with an off the shelf component because there is no drop in replacement or they wouldn't have had Maxim build a custom chip in the first place. Nor can they simply change the battery easily, as the batteries are manufactured to spec as well.

If samsung had offloaded the charging and battery management control into software running on one of the processors in the phone, then they would likely have been able to fix the problem with a firmware update. Now, because they did not have the sense to do what everyone else is doing, they are fucked.

The best kept open secret in the Phone / Tablet world are the PSOC processors that are used extensively for all of the low level work in these devices. Cypress sells nearly as many processors as Broadcom, and nobody has ever heard of them. Their processors come with built in PMIC, Capacitive touch sense (which is why everyone started using them in the first place), and a host of other powerful features that reduce part count and unit cost. I have personally designed a half dozen devices that used them, two of which had battery charging circuits and charge control software. The irony is that the PSOC processors cost cost about $5 each and they are full featured processors while the Maxim ICs Cost more than that and are just a PMIC.

TLDR: Samsung is staffed by incompetent engineers. Its no wonder they have exploding phones, their engineers designed a phone with at least $5 more parts than they should have had, but skipped on the thermostat protection on the battery to save $0.50. If they used this same chip in the S5, then they lost more than 60 million dollars in excess unit cost in just the first three months of sales, and now with the S7, its going to cost them billions.

Comment Re: Samsung is starting to behave like Tesla (Score 2) 110

How many battery power devices have you designed?

Theres a big difference between the LiPo pack you bought for your toy RC car, and the packs that go in commercial devices. Almost any LiPo you can buy "off the shelf" has built in battery protection circuits. That is not true for built in batteries and devices in quantity. It is left to the device manufacturer to decide if they want the built in circuits or not, and in most cases where there is a processor in the device, the manufacturer elects to add their own software to handle the charging and circuit protection as they are cheaper by quite a bit. This is especially true in cell phones and tablets where the battery is custom designed and manufactured, typically for a specific model or line of phones. The cost of adding the protection circuit to the pack is significantly higher, as is the space and weight penalty. The actual circuit to charge the phone or do battery protection is extremely simple and small, the complexity is transfered into the software domain where a processor is used to handle the logic involved. If your device already has a processor (or in the case of phone more than one), this extra chip is a wasted expense.

Even NiMH packs are supposed to have that thermistor you talked about, but except in rare cases, I have not seen a single pack that actually has the thermistor. Its not much of an expense, but they seem to forego that altogether. Cutting corners in pack manufacturing is the norm, since the packs are high volume low margin devices (All cells and most packs are made in china to varying degrees of quality), and every company making and using them is looking for every opportunity they can find to cut costs.

Comment Re: Samsung is starting to behave like Tesla (Score 4, Insightful) 110

So what is the real culprit here? I thought it could have been from using the wrong charging IC or someone in the battery department changing the chemistry and not telling the EEs. But they are catching fire without charging.

Lithium chemistry batteries are finicky little bastards. They are not just dangerous from over-charging (or charging too fast), but also from over-discharging and discharging too rapidly. If they are drawing 1A from a battery that is only designed to handle 1/2A, or worse, if their control circuitry allows the battery to fall below a minimum voltage, then the batteries can go into runaway thermal overload. The worst part is that all of the various factors change with pressure, so at 35k feet, you have to be far more conservative with the batteries to avoid a fire. With Li batteries, getting the extra capacity out of them can be very dangerous, and it only take a small error to end up with fires or explosions. Worse still parameters can vary widely from individual battery to individual battery and is made far worse when the batteries are sloppily made.

It should also be noted that it is unlikely that the phones have a dedicated battery monitoring chip as these cost a couple bucks, even in large quantities, and all they really are is a small microprocessor and a couple cents worth of transistors. Since cell phones already have the processor, they just use that, and add the couple of small transistors they need to handle the charging and discharge monitoring.

The root problem could be as simple as a badly chosen set of constants for cutoff voltage and thermal protection, or could be more insidious, such as a thermal protection sub-routine that doesn't work properly because at shutdown the processor looses power and reboots, thus continually drawing power from the battery.

Comment Re:What's good for the goose (Score 1) 756

you agree to them when you sign up to run.

As there is no way to effectively gain political power in this country, your options are

A: Agree to said one sided rules

B: Don't run for office and let others run things however they see fit

C: Foment revolution.

These are not very appealing options, especially considering that the gatekeepers to authority mentioned in A, are the very people we need to be defended from. If there is one overwhelming takeaway from this years election, it is that the majority is tired of getting shit on by the 0.1%. Its that same 0.1% that either constitutes the superdelegates, or directly controls their actions. They are just one more step towards corruption in our society, and for any who are smart enough to pay attention, The 99% is getting pretty pissed off about it. Another 4 years of the political elite seeing to it that our democratic value are subverted and our votes are diluted until a supermajority can't win without "superdelegate approval", and those riots we keep seeing in the news wont just be #BlackLivesMatter anymore. We are not far off from the next candidate dropping hints that we should be "exercising our 2nd amendment rights" on the political elite.

Comment Re:What's good for the goose (Score 1) 756

Clinton won the vote without the superdelegate vote so in fact we do know that Sanders would not have won. There was no rigging of the election.

I personally know of two Bernie suppporters who didn't vote because the press claimed from the beginning that the results were a foregone conclusion due to the superdelegate vote. The news from April and May were rife with interviews from people who were voting for Clinton simply because they felt that, although Bernie was the better candidate, they had been told repeatedly in DNC sponsored ads that Bernie couldn't win a general election (We heard that same crap about Trump from the other side, and he still might win in spite of himself). In short, It is my considered opinion that Sander would have won had the superdelegates withheld their votes until after the general election. This is Not just My opinion

Granted, no one can know what would have been, but when voters are staying home, or otherwise voting for a different candidate because the establishment has told them their pick cant win anyways; when voters are not even being given the chance to hear about a candidate because the DNC has picked their winner and wont even release the party roles so that the candidate can reach primary voters and likely donors; When DNC executives are actively contacting large party donors and telling them not to give money to a candidate: Its hard not to think about how much that candidate has been screwed. This is the real reason people hate Hilary. Not because she is a bad person, but because bad people and corrupt people made sure she would get the nomination. It was handed to her as though it were somehow her due.

The most telling aspect of this election: Nearly every single person Sanders spoke to in person voted for him. The same cannot be said of any other candidate.

Comment Re:What's good for the goose (Score 5, Informative) 756

There was no rigging.

The superdelegates are election tampering. Their express purpose is to allow the Political elite to shut out candidates, or promote one of their own. The republicans have the same structure, but did not give their superdelegates enough power to override the crazy train.

We'll probably never know if Sanders would have won without the Democratic National Committee and the superdelegates doing everything in their power to shut out Sanders. That is why, in spite of facing an imbecile for an opponent, the democrats could still manage to lose this election. The Democrats political elite severely tampered with this election and they know it, and for many people that kind of tampering is enough to prevent them from ever voting for Clinton. There is a small minority of people who will vote for Trump now, just to watch it all burn and teach the god damn career politicians a lesson.

Comment Re:Bandiwidth is *free* fallacy.. (Score 4, Insightful) 229

Bzzz! Hold it right there! What is "sufficient bandwidth"

My ISP has advertised 100mbit service, and I have every right to expect that 100mbit will be available 100% of the time. ISPs oversubscribe their actual available bandwidth because they know almost no one ever uses 100% of the available bandwidth 100% of the time. That doesn't change the fact that they are charging multiple customers for the same resource. The ISPs can't then turn around and say, that there isn't enough because their customers are using more than their fair share, when in fact, the ISP has sold more than they had available in the first place.

Using that metric, Sufficient bandwidth is whatever is required to provide 100% of their customers with 100% of the promised bandwidth. Anything less than that is just the ISP whining because they are being held to the contract they themselves wrote.

In that regard, ISPs with data caps should be required to advertise the datacap / billing period instead of the peak speeds, customers will quickly stop coming in the door when it is made obvious that a 50GB / month limit effectively means that on average you can only get 150kb/s download speed over the course of an entire month. If the ISP had to advertise that 150kb/s instead of being able to claim 100mb/s speeds, they would quickly change their minds about data caps.

Comment Re:Bandiwidth is *free* fallacy.. (Score 3, Informative) 229

In addition to bandwidth is free you forgot the one about how since the hardware infrastructure for networks is a sunken cost it should be free to use. I haven't figured that one out yet; apparently the underlying assumption is that the investors who paid up front ought to be robbed of their expected returns.

In a large percentage of cases, those up front investments were paid for by the FCC. And yes, those investors ARE getting robbed blind.

Slashdot Top Deals

Put not your trust in money, but put your money in trust.