The bitter lesson I've learned over the years is that the larger the team, the simpler the tool set and the less clever the implementation must be.
And it makes sense. You pretty much have to target the 10th percentile programmer in your organization or your going to have your systems ruined over time.
In a small organization, you can probably hire selectively enough that your 10th percentile programmer can handle Scala. In a large organization (like Twitter is now), not so much.
Obviously this was meant to be funny, but there's some underlying truth - the skills that are needed for success in our post-industrial society (things like conscientiousness and empathy (needed in many service jobs)) seem to be more prevalent in women.
It's been slowly becoming a woman's world for some time now - and yes, the patriarchy (supported by the men and women of most cultures) has been getting in the way of their natural ascendancy.
Luckily, as a society we're likely to be a little more cognizant of biases as we adjust to this realization. I doubt my sons will end up having to be twice as competent in order to be considered half as good as their female counterparts.
To be meaningful, you need at least two numbers: the number of crashes avoided because of software intervention and the number of crashes caused by driver inattention.
I think that two numbers would be deceptive because almost no-one is capable of acknowledging their inattention. If you found at that that 50% of accidents are caused by inattention, but the autopilot is a 20% *worse* driver than someone paying attention, you *know* that everyone would flee from AutoPilot it on the assumption they won't be part of the 50% failing to pay attention.
One of the primary problems is that humans (in general) are incapable of acknowledging the weaknesses that cause accidents, thus making it very hard to take measures that reduce the number of accidents. Splitting the numbers apart would contribute to this problem.
Sadly, I can be pretty close to 100% sure that some minimum wage flunky has no path whatsoever to channel feedback on an non-retail oriented issue to upper management.
Life's hard enough for retail employees.
Well, that, and with my luck, she'd find the one retail clerk who was a corporate-drone-hopeful that would defend the practice, and then I'm stuck bailing my wife out of jail *and* figuring out where to send flowers.
As someone who knows three people who were Windows-10'd against their will, telling them to waste x hundred hours of time trying to get compensation for the dozen hours (or $200) it took for them or someone else to undo the damage seems a little... counterproductive.
However, when we passed a Microsoft store advertising the Windows 10 upgrade, I did have to stop my wife (one of the victims) from barging in there and giving the staff a piece of her mind.
> Take that, along with some scary findings that the US's "PRODUCE MORE NOW! WITH LESS THAN YOU DID LAST QUARTER, OBEY OR BE FIRED! DO NOT DARE ASK FOR MORE THAN THE 1-WEEK OF VACATION ALLOTED TO YOU! CORPORATE HAS SPOKEN!" culture produces unbearable levels of stress, which increases risk for depression.
Well, there's the small problem that the US (and the rest of the industrialized world) has a standard of living many times the rest of the world. To sustain that, we need to be many times more productive than the rest of the world. And that is becoming harder and harder.
We're going crazy trying to stave off the Great Reckoning, when the developed world median income falls into place with the rest of the world. So, yeah, we're a little stressed. Who wouldn't be? But personally, I don't blame anybody but myself. Once we accept we can't stop water from flowing downhill, and accept a 75% drop in income, (and survive the stress from *that*), the pressure to produce 4x what everyone else in the world produces will be off.
And yet... somehow, I keep fighting that day, pushing it off for just one more year, mental health be damned.
> If you accept the premise that there's a problem* then isn't this exactly the right thing to do?
You are correct *if* there are no associated trade-offs.
But there are *always* trade-offs.
The only interesting question is whether the benefits (fewer children doing things that in some fraction of the cases have significant long-term consequences) is worth the costs (creation of infrastructure to real-time censor images, etc.)
It's not all that different from what asking what costs are acceptable for reducing the level of teenage drinking and driving, another activity that some children are prone to that can also occasionally have disastrous consequences.
Lotus 123 macros were my introduction to the fact that "tools that you don't have to be a programmer to use" actually means "tools that end up being *immensely* more complex than any computer program because we don't actually give you the constructs to program sanely".
To this day, I still find people who loudly claim they are non-programmers because programming is too tough who maintain 'macros' whose complexity put my 100,000 line programs to shame. If they'd been given real tools, they'd probably have created human level AI during their coffee breaks, except, of course, "real tools are for programmers" and that would be 'too difficult' for them to even contemplate using.
I'd feel more shame for what my spreadsheet did to the sanity of the poor soul who came after me, but there was simply no way to sanely code 123.
> Honestly, I would have hoped that they would/could find another position for those 2 drivers instead of dumping them out on the street. *sigh*
They may have actually done so or, given the trucking industry was booming at the time, they may have found jobs the next day. The boss obviously wasn't losing any sleep over it. (He used to regale me with stories of the games the employee truckers and shippers would play before he instituted some minimal controls on how and what they were shipping. Mostly low-level stuff - hire the trucking firm whose president invited the factory shipper out to his cottage for a few days each summer, or pretend to pick up a non-existent shipment from a customer that was close to a trucker's girlfriend's house, that short of thing.)
But yes, hot and cold is exactly the right description. I was a bit of an approval junky, and my work was being taken seriously by *adults*! And then suddenly, I had cost *real* adults their jobs.
As I said elsewhere, the shame is not improving efficiency, it was that I didn't even consider the *possibility* humans could be affected at all. I guess it was a lesson to learn early...
The point of my story was not that automating jobs is bad (or good). My point is that it's easy to forget that there are real world implications to what we do, especially for the young or inexperienced.
My shame is not in automating those jobs away. My shame (although I excuse it by being a 17 year old adult approval junky at the time) was that I didn't even consider that my job *could* have real world implications.
If we're talking about how are code was used, I remember in high school (many moons ago) writing Turbo Pascal programs and Lotus 123 macros for a shipping department of a sizable company that hadn't yet computerized. I was brought in by the manager of the shipping department because he could hire a high-schooler when he couldn't get authorization to computerize from within the internal IT department (which was busy sinking the company with some massively expensive software controlling the manufacturing).
Anyway, I was very proud of allowing my boss to get all the data that he wanted, and he was very, very pleased that his department now had some means of seeing what was going on.
I distinctly remember when he called me in and thanked me. Due to my program, he'd had enough data to improve efficiency 25%!
I glowed.
Now he'd been able to let go 2 out of the 8 drivers they had.
> more as indicative of an amateur in the area than any real flag of a problem area.
Alright, but then the innovation of multithreading (and each innovation that increases complexity) has relegated more programmers to the status of amateur. And at the point at which you have a feature that's very difficult for the majority of programmers to handle, you have a dragon.
It's also a matter of scale. Sure, almost any program nowadays is likely to require half a dozen threads to do anything, and that's not an incredible burden. But as the push comes to make *every* aspect of the program multithreaded, it's pretty easy for the complexity to rise exponentially, rendering all but 0.001% of th programmers incapable of handling the program.
So, as more and more of us become 'amateurs' by dint of increasing complexity, the dragon is getting bigger and bigger.
UTC has one big win: co-ordination of an event between different time-zones.
*Every* other use of time is either neutral or heavily in favour of Time Zones. Since for the vast majority of humans, co-ordination of non-local events is a trivial amount of their references to time, Time-zones win hugely.
This aside from the obvious problems during travel. Set your watch once (if your phone doesn't do it for you) when you arrive at a new time zone? Or learn the scores of "usual times" for meals, business hours, etc. for the new location.
I doubt a Civilization's AI's real-world applicability to real-life diplomacy, but a truly successful Civilization AI (i.e. one that played at Master level exactly as a human would) would terrify the hell out of me. Being a master of Civ involves managing limited information and about a *thousand* degrees of freedom each move, if not more.
It makes Go look like a cakewalk (in terms of the size of the decision tree).
If AI's can do that, then probably 2/3rds of the intellectual-related jobs on the planet are toast.
(As for applicability to real-life diplomacy, you have to be able to evaluate "success". Since there's no agreement as to what that even means, there's no teaching an AI.)
To me this makes it crystal clear that *now* is the only time to make significant changes lest you suffer Python's fate of essentially forking the user base.
Making breaking changes over 22 years is a disaster. Making them over 2-3 is survivable.
I like that analogy a lot. There's a careful balance between inculcating into my children that hard work, study, etc. is necessary, but also letting them know that it's not a guarantee. You also need some luck, but less of it, the harder you work. Hopefully it means that they won't feel betrayed in areas where they "do all the right things" but don't get the appropriate reward.
Or as I've told them:
Most of my success has depended on luck, but it's amazing how the harder I worked at something, the luckier I was.
Actually a fair number of the older PINPads take a crazily long time to generate ARQCs and validate ARPCs. I suspect whoever was supplying the HSM equivalents for the PINPads decided to go green and power them with an easily-tired gerbil rather than electricity.
Sure Chip + Sig will reduce card cloning, which is *by far* the biggest problem *at the moment*.
My worry is that once since crime migrates, and the fraudsters have got a lot of very smart engineers and programmers working for them now, once card cloning isn't a big business, will they migrate to something that isn't protected by Chip + Sig and we'll have this heartache all over again.
Certainly not helped by the fact that Visa and M/C are pushing merchants to do away with ARPCs and now they're even proposing to not include the amount in the ARQC data so they can do pre-insertion. Talk about reducing chip to the minimal possible security!
The choice is up to each individual card issuer (and they could vary this among card batches).
But indeed, almost all credit cards are chip and sig in the US. I'll admit the first time I encountered this, I failed the test cases I was running as obviously a major screw-up. Thought my boss was having me on when he told me that seriously most US issuers were going chip + sig.
Still, in the end, it prevents card cloning, which is where the losses were beginning to become industry threatening. (The US had no real intention of switching a few years ago - I guess they didn't realize they were going to become the magnet for card fraudsters for the entire world.)
Even at the weakest level, EMV adds one important security factor. You can't simply skim a chip card and make a new working chip card.
Without PIN, chip cards won't prevent the card from being individually stolen and used, but that's not where the industrial level losses were occurring. It had reached the point of being a major business for organized crime, and this will put a serious crimp in it. (When I was more involved in bank security a few years ago, you could find franchising skimmer opportunities on YouTube that were renewed every few minutes as they got taken down.)
As well, as one wealthy hold-out to chip, the US was attracting the attention of the world's high tech criminals. Since crime migrates to the weakest link, you don't want to be the slowest deer in the herd, which the US was rapidly becoming. (The US punitive legal system had kept the US from being a favored target when other countries had left their doors unlocked, but once there weren't any other wealthy countries with low hanging fruit, cyber crime was going exponential.
There'll be other forms of crime (crime migrates to different types of crime as well), but few that worked so well on the an industrial scale.
If we're gearing education towards the top 10%, then sure, emphasize creativity because they probably have the culture, and more importantly, parents who can backstop their "finding their career".
But if we're talking about everyone else, we're looking at those who need to get a job fresh out of school - the sort of service jobs whose primary requirement is mastery of a limited skill set (that does require training and a willingness to learn) and the conscientiousness required to show up and do an adequate job every day, no matter how you are feeling. You need the conscientiousness that allows you to serve customers respectfully and cheerfully no matter how you personally feel.
As we plunge into the great equalization, where Western salaries drop and world income's rise to meet at ~$10K, I don't think we'll have the time and/or money to support a lot of creativity. Our future (and the future of mankind in general) will look rather like China for a generation or two before we finally see a truly global increase in incomes.
(Unless the robots come - in that case, we all starve.)
So, I'm wondering if you've tried going to your boss and asking for more money?
I'd be very careful about offering that advice. There's an arm-load of studies showing that women are penalized far more heavily (by both men and women) for being perceived as selfish. And for many, many teams, demanding your due *is* construed as selfish. The team has limited resources, and by demanding more, you are prioritizing yourself over the team.
Far too often I've seen "She's a jerk." "Well, he's even more of a jerk." "Well, yes, but she should know better"
Men get away with it more easily because we're assumed to be selfish jerks. Women are expected to understand that difficulties that team faces and prioritize the team's needs over their own.
It's yet one more way that women *cannot* behave the same way as a man without paying a significant penalty. It's not just that they prefer to play by different rules, it's that they are also penalized when they don't.
Slashdot Mirror
> Yes, some, but let's not forget the millions of others who can easily prevent it by quitting the damn sugars.
Easily? I don't think that word means what you think it does. Almost by definition, if it was easy, they'd already be doing it.
Or by easy, did you mean "restructure society and re-engineer human dopamine reward systems" type of easy? :-)
Very good point indeed. You should have posted under a real name to make it more visible!
The bitter lesson I've learned over the years is that the larger the team, the simpler the tool set and the less clever the implementation must be.
And it makes sense. You pretty much have to target the 10th percentile programmer in your organization or your going to have your systems ruined over time.
In a small organization, you can probably hire selectively enough that your 10th percentile programmer can handle Scala. In a large organization (like Twitter is now), not so much.
Obviously this was meant to be funny, but there's some underlying truth - the skills that are needed for success in our post-industrial society (things like conscientiousness and empathy (needed in many service jobs)) seem to be more prevalent in women.
It's been slowly becoming a woman's world for some time now - and yes, the patriarchy (supported by the men and women of most cultures) has been getting in the way of their natural ascendancy.
Luckily, as a society we're likely to be a little more cognizant of biases as we adjust to this realization. I doubt my sons will end up having to be twice as competent in order to be considered half as good as their female counterparts.
To be meaningful, you need at least two numbers: the number of crashes avoided because of software intervention and the number of crashes caused by driver inattention.
I think that two numbers would be deceptive because almost no-one is capable of acknowledging their inattention. If you found at that that 50% of accidents are caused by inattention, but the autopilot is a 20% *worse* driver than someone paying attention, you *know* that everyone would flee from AutoPilot it on the assumption they won't be part of the 50% failing to pay attention.
One of the primary problems is that humans (in general) are incapable of acknowledging the weaknesses that cause accidents, thus making it very hard to take measures that reduce the number of accidents. Splitting the numbers apart would contribute to this problem.
Sadly, I can be pretty close to 100% sure that some minimum wage flunky has no path whatsoever to channel feedback on an non-retail oriented issue to upper management.
Life's hard enough for retail employees.
Well, that, and with my luck, she'd find the one retail clerk who was a corporate-drone-hopeful that would defend the practice, and then I'm stuck bailing my wife out of jail *and* figuring out where to send flowers.
As someone who knows three people who were Windows-10'd against their will, telling them to waste x hundred hours of time trying to get compensation for the dozen hours (or $200) it took for them or someone else to undo the damage seems a little... counterproductive.
However, when we passed a Microsoft store advertising the Windows 10 upgrade, I did have to stop my wife (one of the victims) from barging in there and giving the staff a piece of her mind.
> Take that, along with some scary findings that the US's "PRODUCE MORE NOW! WITH LESS THAN YOU DID LAST QUARTER, OBEY OR BE FIRED! DO NOT DARE ASK FOR MORE THAN THE 1-WEEK OF VACATION ALLOTED TO YOU! CORPORATE HAS SPOKEN!" culture produces unbearable levels of stress, which increases risk for depression.
Well, there's the small problem that the US (and the rest of the industrialized world) has a standard of living many times the rest of the world. To sustain that, we need to be many times more productive than the rest of the world. And that is becoming harder and harder.
We're going crazy trying to stave off the Great Reckoning, when the developed world median income falls into place with the rest of the world. So, yeah, we're a little stressed. Who wouldn't be? But personally, I don't blame anybody but myself. Once we accept we can't stop water from flowing downhill, and accept a 75% drop in income, (and survive the stress from *that*), the pressure to produce 4x what everyone else in the world produces will be off.
And yet... somehow, I keep fighting that day, pushing it off for just one more year, mental health be damned.
> If you accept the premise that there's a problem* then isn't this exactly the right thing to do?
You are correct *if* there are no associated trade-offs.
But there are *always* trade-offs.
The only interesting question is whether the benefits (fewer children doing things that in some fraction of the cases have significant long-term consequences) is worth the costs (creation of infrastructure to real-time censor images, etc.)
It's not all that different from what asking what costs are acceptable for reducing the level of teenage drinking and driving, another activity that some children are prone to that can also occasionally have disastrous consequences.
Dear God, you are so right.
Lotus 123 macros were my introduction to the fact that "tools that you don't have to be a programmer to use" actually means "tools that end up being *immensely* more complex than any computer program because we don't actually give you the constructs to program sanely".
To this day, I still find people who loudly claim they are non-programmers because programming is too tough who maintain 'macros' whose complexity put my 100,000 line programs to shame. If they'd been given real tools, they'd probably have created human level AI during their coffee breaks, except, of course, "real tools are for programmers" and that would be 'too difficult' for them to even contemplate using.
I'd feel more shame for what my spreadsheet did to the sanity of the poor soul who came after me, but there was simply no way to sanely code 123.
> Honestly, I would have hoped that they would/could find another position for those 2 drivers instead of dumping them out on the street. *sigh*
They may have actually done so or, given the trucking industry was booming at the time, they may have found jobs the next day. The boss obviously wasn't losing any sleep over it. (He used to regale me with stories of the games the employee truckers and shippers would play before he instituted some minimal controls on how and what they were shipping. Mostly low-level stuff - hire the trucking firm whose president invited the factory shipper out to his cottage for a few days each summer, or pretend to pick up a non-existent shipment from a customer that was close to a trucker's girlfriend's house, that short of thing.)
But yes, hot and cold is exactly the right description. I was a bit of an approval junky, and my work was being taken seriously by *adults*! And then suddenly, I had cost *real* adults their jobs.
As I said elsewhere, the shame is not improving efficiency, it was that I didn't even consider the *possibility* humans could be affected at all. I guess it was a lesson to learn early...
The point of my story was not that automating jobs is bad (or good). My point is that it's easy to forget that there are real world implications to what we do, especially for the young or inexperienced.
My shame is not in automating those jobs away. My shame (although I excuse it by being a 17 year old adult approval junky at the time) was that I didn't even consider that my job *could* have real world implications.
If we're talking about how are code was used, I remember in high school (many moons ago) writing Turbo Pascal programs and Lotus 123 macros for a shipping department of a sizable company that hadn't yet computerized. I was brought in by the manager of the shipping department because he could hire a high-schooler when he couldn't get authorization to computerize from within the internal IT department (which was busy sinking the company with some massively expensive software controlling the manufacturing).
Anyway, I was very proud of allowing my boss to get all the data that he wanted, and he was very, very pleased that his department now had some means of seeing what was going on.
I distinctly remember when he called me in and thanked me. Due to my program, he'd had enough data to improve efficiency 25%!
I glowed.
Now he'd been able to let go 2 out of the 8 drivers they had.
I stood there speechless.
There were real people underneath those numbers.
> more as indicative of an amateur in the area than any real flag of a problem area.
Alright, but then the innovation of multithreading (and each innovation that increases complexity) has relegated more programmers to the status of amateur. And at the point at which you have a feature that's very difficult for the majority of programmers to handle, you have a dragon.
It's also a matter of scale. Sure, almost any program nowadays is likely to require half a dozen threads to do anything, and that's not an incredible burden. But as the push comes to make *every* aspect of the program multithreaded, it's pretty easy for the complexity to rise exponentially, rendering all but 0.001% of th programmers incapable of handling the program.
So, as more and more of us become 'amateurs' by dint of increasing complexity, the dragon is getting bigger and bigger.
UTC has one big win: co-ordination of an event between different time-zones.
*Every* other use of time is either neutral or heavily in favour of Time Zones. Since for the vast majority of humans, co-ordination of non-local events is a trivial amount of their references to time, Time-zones win hugely.
This aside from the obvious problems during travel. Set your watch once (if your phone doesn't do it for you) when you arrive at a new time zone? Or learn the scores of "usual times" for meals, business hours, etc. for the new location.
Agreed, this seemed deliberately click-baity to me (and it worked!)
I've never heard the term naked black hole before.
I doubt a Civilization's AI's real-world applicability to real-life diplomacy, but a truly successful Civilization AI (i.e. one that played at Master level exactly as a human would) would terrify the hell out of me. Being a master of Civ involves managing limited information and about a *thousand* degrees of freedom each move, if not more.
It makes Go look like a cakewalk (in terms of the size of the decision tree).
If AI's can do that, then probably 2/3rds of the intellectual-related jobs on the planet are toast.
(As for applicability to real-life diplomacy, you have to be able to evaluate "success". Since there's no agreement as to what that even means, there's no teaching an AI.)
To me this makes it crystal clear that *now* is the only time to make significant changes lest you suffer Python's fate of essentially forking the user base.
Making breaking changes over 22 years is a disaster. Making them over 2-3 is survivable.
Well, C was created in 1972, so that did give them some 20+ years before stability...
I like that analogy a lot. There's a careful balance between inculcating into my children that hard work, study, etc. is necessary, but also letting them know that it's not a guarantee. You also need some luck, but less of it, the harder you work. Hopefully it means that they won't feel betrayed in areas where they "do all the right things" but don't get the appropriate reward.
Or as I've told them:
Most of my success has depended on luck, but it's amazing how the harder I worked at something, the luckier I was.
Actually a fair number of the older PINPads take a crazily long time to generate ARQCs and validate ARPCs. I suspect whoever was supplying the HSM equivalents for the PINPads decided to go green and power them with an easily-tired gerbil rather than electricity.
Sure Chip + Sig will reduce card cloning, which is *by far* the biggest problem *at the moment*.
My worry is that once since crime migrates, and the fraudsters have got a lot of very smart engineers and programmers working for them now, once card cloning isn't a big business, will they migrate to something that isn't protected by Chip + Sig and we'll have this heartache all over again.
Certainly not helped by the fact that Visa and M/C are pushing merchants to do away with ARPCs and now they're even proposing to not include the amount in the ARQC data so they can do pre-insertion. Talk about reducing chip to the minimal possible security!
The choice is up to each individual card issuer (and they could vary this among card batches).
But indeed, almost all credit cards are chip and sig in the US. I'll admit the first time I encountered this, I failed the test cases I was running as obviously a major screw-up. Thought my boss was having me on when he told me that seriously most US issuers were going chip + sig.
Still, in the end, it prevents card cloning, which is where the losses were beginning to become industry threatening. (The US had no real intention of switching a few years ago - I guess they didn't realize they were going to become the magnet for card fraudsters for the entire world.)
Even at the weakest level, EMV adds one important security factor. You can't simply skim a chip card and make a new working chip card.
Without PIN, chip cards won't prevent the card from being individually stolen and used, but that's not where the industrial level losses were occurring. It had reached the point of being a major business for organized crime, and this will put a serious crimp in it. (When I was more involved in bank security a few years ago, you could find franchising skimmer opportunities on YouTube that were renewed every few minutes as they got taken down.)
As well, as one wealthy hold-out to chip, the US was attracting the attention of the world's high tech criminals. Since crime migrates to the weakest link, you don't want to be the slowest deer in the herd, which the US was rapidly becoming. (The US punitive legal system had kept the US from being a favored target when other countries had left their doors unlocked, but once there weren't any other wealthy countries with low hanging fruit, cyber crime was going exponential.
There'll be other forms of crime (crime migrates to different types of crime as well), but few that worked so well on the an industrial scale.
If we're gearing education towards the top 10%, then sure, emphasize creativity because they probably have the culture, and more importantly, parents who can backstop their "finding their career".
But if we're talking about everyone else, we're looking at those who need to get a job fresh out of school - the sort of service jobs whose primary requirement is mastery of a limited skill set (that does require training and a willingness to learn) and the conscientiousness required to show up and do an adequate job every day, no matter how you are feeling. You need the conscientiousness that allows you to serve customers respectfully and cheerfully no matter how you personally feel.
As we plunge into the great equalization, where Western salaries drop and world income's rise to meet at ~$10K, I don't think we'll have the time and/or money to support a lot of creativity. Our future (and the future of mankind in general) will look rather like China for a generation or two before we finally see a truly global increase in incomes.
(Unless the robots come - in that case, we all starve.)
So, I'm wondering if you've tried going to your boss and asking for more money?
I'd be very careful about offering that advice. There's an arm-load of studies showing that women are penalized far more heavily (by both men and women) for being perceived as selfish. And for many, many teams, demanding your due *is* construed as selfish. The team has limited resources, and by demanding more, you are prioritizing yourself over the team.
Far too often I've seen "She's a jerk." "Well, he's even more of a jerk." "Well, yes, but she should know better"
Men get away with it more easily because we're assumed to be selfish jerks. Women are expected to understand that difficulties that team faces and prioritize the team's needs over their own.
It's yet one more way that women *cannot* behave the same way as a man without paying a significant penalty. It's not just that they prefer to play by different rules, it's that they are also penalized when they don't.