If there were no regulations (basically a form of self-regulation), then how exactly do you prove that you were not negligent?
This is a much easier problem for makers of self-driving systems because one of the primary functions of such systems is to gather large amounts of data about the precise locations of the vehicle and all other vehicles, pedestrians, cyclists and other objects around it, as well as status of any traffic signals, etc. So as long as the self-driving system stores at least a few minutes of data and retains it all in the event of an accident, the precise sequence of events and the decisions made by the self-driving vehicle can be reconstructed in tremendous detail -- orders of magnitude more accurately than any eyewitness testimony. This data can be evaluated, the root cause precisely determined and liability accurately assigned.
The evaluation process can even include replaying the data in simulation with slight adjustments to the behavior of every participant to see what scenarios could have changed the outcome. Even if the self-driving system was not at fault, if it turns out that a change in its behavior could have avoided the collision, the software can be updated and the incident simulated to verify that the change works correctly -- and then the updated software can be regression-tested against simulations of all of the other recorded accidents, as well as the full suite of tests developed during design and production. If the update is an improvement overall, it can then be pushed to all vehicles using the system, making all of them better drivers at once.
So the only regulation that might be needed is regulation requiring that self-driving vehicles create this record in a sturdy, tamper-resistant store, analogous to an airplane's "black box". At this stage in development, all self-driving vehicles already have something like this (modulo the tamper resistance, perhaps), because they're actively being refined. There's no reason ever to remove it so the regulation may not even be necessary.
One other bit of regulation that may or may not be useful (or required) is to demand that self-driving system makers share these accident logs with their competitors. Given that they probably won't all be using the same sensor systems, it's not clear how useful the logs would be, but perhaps they could extrapolate a physical model from the sensor logs and then compute the inputs that their sensors would have received from that physical model. If this works well enough, it could enable all self-driving vehicles to get better, and perhaps the NHTSA could use the method to produce a standard test suite that all manufacturers have to pass, though I'd worry about the risks of transformation fidelity and overtraining so that may not be as useful as it appears.
And the car owner, and the manufacturer of the car and of the software. There may be many problems with self-driving cars, but finding lawsuit-targets ain't one.
Sure, you can sue all of them, but if the self-driving system made a bad decision and caused an accident, the maker of that system is the only place courts could reasonably pin the liability.
Who do you sue when a driverless car runs you over?
The maker of the self-driving system. Nothing else would make sense. In this case the vehicles will be owned and operated by Waymo, so there's really no question. If a self-driving vehicle were privately-owned, unless the self-driving system maker can show that the owner of the vehicle did something wrong (didn't keep the software up to date, didn't maintain the mechanical functions of the car, actively interfered with the system during operation, causing the accident), then there's no entity other than that company who could be liable for any accident caused or contributed to by the self-driving system.
People always ask this question, but the answer is obvious.
If the work's getting done then you're doing the job you were hired for.
The mechanism doesn't matter.
However, if your script is doing the job and you're doing nothing, then your manager is not doing the job they were hired for. They should see what you've accomplished, give you a bonus or raise for your efficiency and assign you some more tasks to automate.
Version numbers should mean something.
Major number means a major revision or change on key components.
Minor is just for patches and fixes.
For libraries, I completely agree that versioning is crucial. In library version numbers:
The major number should change when the update is incompatible, i.e. apps built against version X won't run with version X + 1.
The minor number should change when the update is backward compatible. Apps built against version X.Y will run with version X.Y+n, but apps built against version X.Y may not run with version X.Y-n.
The sub-minor number should change with every release that is backward and forward compatible. Apps built against any version X.Y.n will run with any version X.Y.m, regardless of the relationship of n and m.
For applications, operating systems, and whatnot? Meh. Having an understandable linear ordering is nice (i.e. not 3.5 -> 3.5.1 -> 4.0 -> 2000 -> XP -> Vista -> 7 -> 8 -> 8.1 -> 10), but only so you can tell which one is newer. For stuff that is released annually or less often, I think year of release is the best option.
While I agree there's a need for a review/recommendation system that can't be gamed by bad actors, nothing about this one indicates it will be any better than the rest.
It sounds like in the long run the intention is to combat fake reviews by surfacing only reviews from people you have a social connection to. There are obvious problems with that, among them that it only works if people you're connected to (a) use the app and (b) have posted an opinion on something you're interested in, but it seems less prone to gaming.
Taxis never cost more than Uber/Lyft where I live.
They also would never show up when called. Uber/Lyft made a huge difference here. I don't really know what the appeal was in bigger cities with a functional cab market, but in my little city (Wilmington, DE) the appeal was a ride within 20 minutes at 2am or 6am both. I saw someone wait over 2 hours for a cab that they scheduled where I work. They pretty much onky show up for airport runs, and you can get them at the train station, maybe a hotel if they have a good relationship with a driver.
Sounds like regulation forced taxi prices to an artificially-low level, resulting in an insufficient number of taxis. Uber/Lyft tune their prices to a more "natural" level, making them more available.
Personally, I'd like to see an Uber/Lyft competitor that doesn't set prices at all, but instead facilitates a real-time auction market between drivers and riders, so that prices are truly supply/demand driven -- and drivers are inarguably independent businesspeople.
Zero is not enough. The plan set out in the Paris Accord assumes that we'll go beyond zero and begin taking CO2 out of the atmosphere.
so where exactly along this long line of production do you expect to emit a negative value to overcome nature
It's not that we need to "overcome nature", it's that we need to overcome history. We've already put a lot of CO2 into the atmosphere and because the natural processes that remove it are very slow the only way to get the climate back into a normal [*] state and stop the warming effect (much less reverse the warming that has already occurred) is to lower the present CO2 levels faster than nature will do it for us. This assumes we don't use other geo-engineering techniques to reduce warming.
Unless we start photosynthesizing ourselves, we'll naturally emit.
Not unless we eat coal. Sure, when you eat an apple you end up combining a portion of the carbon in the apple with atmospheric oxygen, then emitting the result as CO2. But the carbon came from CO2 in the atmosphere, absorbed by the apple tree. Human (and other animal) CO2 emissions are part of a cyclical process that doesn't make a significant net change in atmospheric CO2. What does make a change is digging up long-sequestered hydrocarbons, burning them to release energy and dumping that carbon into the atmosphere.
[*] "Normal" is a funny word here. There is no such thing as a "normal" Earth climate. The planet's climate changes all the time. Usually -- but not always -- over very long timeframes relative to human perception, but still, it's always changing. All of recorded human history has occurred in a brief warm period of the Quaternary Ice Age. But the point is that we know and like it the way it has been, and in fact we want it to stay this way forever (or at least for a very long time). We don't want it to get hotter and force us to move massive chunks of our population and figure out how to change our food production approaches, which is what we're currently facing. We also don't want it to get much colder. Luckily, we have now proven that we can "engineer" a warming climate, so we can ensure that we never have to deal with colder temperatures. We did it by accident, but we could do it again on purpose if needed. Now we need to figure out how to stop warming, and perhaps even cool the planet a bit. Thus, we can halt the natural fluctuations of the planetary climate forever and lock it permanently in the state that we want it; the one we call "normal".
And the solution is always: exterminate capitalism immediately, implement worldwide socialism, the United States must pay the staggering cost of the whole thing by impoverishing its people, and we need global governance without any voting or input from the deplorables. The whole thing might be more believable if the "solution" wasn't always the same thing.
Basically nobody is proposing that "solution" beyond a few fringe wackos who think it's a good idea regardless of climate change but are happy to jump on this particular bandwagon. Certainly, market-driven approaches like carbon credit trading and carbon taxation are absolutely nothing of the sort, nor are voluntary international agreements like the Paris Accord.
Market-based capitalism is by far the best system we have for accurately optimizing economic activity. The powerful signalling mechanism that prices provide does a better job than anything else of eliminating waste in production and redirecting resources to the most useful sectors -- where "useful" is determined by the needs and desires of the population as a whole. Not that markets are perfectly efficient, but they're far better than anything else we've found. But markets don't consider externalities, so we need some way to make carbon emission a cost of production. Capital will then flow away from carbon-emitting processes and to carbon-neutral -- and eventually even carbon-absorbing -- processes. Moreover, markets price the future as well as the present, so it's not even necessary to impose the carbon costs right away to start seeing immediate benefits. If you enact a plan to impose carbon taxes (or similar) and convince markets that the plan is not going to be rescinded or relaxed, capital reallocation will start even before the costs hit, as industries and the moneymen who back them seek to minimize the shock of the change.
To address the climate change problem, we need to apply the power of capitalism, not exterminate it.
So they haven't attempted to codify it, just established a panel that applies some general guidelines. The only real rule is that the individual must have transitioned (the meaning of which isn't really defined and I strongly suspect has shifted and will continue to shift over time) two years prior to issuance of a gender recognition certificate.
We can expect all of this to continue being further muddied as additional notions about gender arise, particularly the expanding list of non-binary categories.
I think that this apparently "statistical" nature is a function of something very fundamental that we don't yet understand, and some day, somebody will find the reality with a slap to the forehead and a loud "D'OH!"
I see no reason to expect this.
Basically, what bothers you is that our intuition, which is formed from macroscopic observations made through our physical senses which were optimized via evolution to enable our survival in a world of macroscopic objects, doesn't map well onto the world of subatomic particles. But why in the world would you expect it to?
No, I do not think laws such as this would fall under the measures to eliminate discrimination” category, nor that they would do any good.
That's a nice statement of what you believe. Can you elaborate on your rationale? Without an explanation of the basis for your belief, your statement is merely a vote, not an argument to sway others' opinions. This is a discussion, not a poll.
No we're saying that the best candidates *should* be chosen...
But there's no objective "best" in this case. There are highly-desirable attributes which can be objectively identified and measured, but it's simply impossible to create an objective stack ranking of all available candidates, male and female. Rather, you have a large pool of candidates, all highly qualified and all roughly equal in ability. Each candidate has particular pros and cons, but it's rarely the case that you can clearly identify that any one of them is strictly better than any other.
In that sort of context, people fall back on personal relationships to make choices, even when they're doing their level best to be objective, because there is no objective way to choose. Which means that in a male-dominated system, men have a systemic advantage, because the male dominance means that men have more of the relevant personal relationships.
It's important to recognize that it's possible for a system to be biased even when none of the participants is biased and even when there are no explicitly-biased rules in place. How do you remove the bias from such a system? Perhaps the only way in some cases is to inject an explicitly counter-biased rule. Applying intentional discrimination to counter unintentional discrimination is paradoxical, but I don't see any other workable approaches on the table.
Note that I strongly doubt the effectiveness of reverse discrimination applied on larger scales where the numbers mean that it will be necessary to elevate less-qualified applicants. That results in stigmatization and often exacerbates the problem. In this case, though, the numbers are small and I think the sort of women who will be chosen for these boards will be more than capable of holding their own, so I don't think we need to worry about this action harming them or creating a perception that female boardmembers are mere tokens.
Sometimes it really is for the public good. The key is to put substantial due process around it, and codify extensive protections against abuse, and guarantee adequate compensation for the person whose property is taken. I think the US generally does a decent job with its handling of eminent domain. Not so much with asset forfeiture.
Does it really work like that in California? Most places it's not just enough to declare you are female, you have to actually live as a female. Shave your beard, wear women's clothing, change your name etc.
Cite? This may be practice, but I'd be really surprised to see it as codified law anywhere, and I'd love to read the statute if it is. I can imagine someone trying to draft such a law getting wrapped around the axle in all the corner cases and details. It seems extremely difficult to craft.
Nowhere does the freedom of speech mean that you're entitled to whatever platform you like and however you like it
Certainly not. But deplatforming doesn't mean that no one is obligated to give you a platform, it means mobs rising up to forcibly reverse your invitation to speak, merely because some of them might be offended.
And yes, the fewer Nazis or Racists out there, the better.
I certainly agree. But I disagree that violent suppression of their political speech is either necessary or effective.
So if they have a 90 day announcement period for more stock sales, they would need to announce by the end of October. Having yourself locked down at that time is basically a death sentence.
Is there any indication that the suit against Musk would "lock down" Tesla?
Yeah, the problem isn't the physical building tech (on the moon). It is just building stuff. On the moon. Did I mention it was on the moon? Space nutters.
I use a formula for all of my password websites (but it's actually very different to the one above, I don't think anyone could ever reverse engineer my password to figure out my formula
If I saw your password for two sites I'd know generally what you're doing, and what the invariant portion is. What's left is a 4-6 character alphanumeric password; maximum entropy 31 bits. You really don't gain a great deal with the invariant string, other than fooling password strength meters, and maybe yourself.
Slashdot Mirror
If there were no regulations (basically a form of self-regulation), then how exactly do you prove that you were not negligent?
This is a much easier problem for makers of self-driving systems because one of the primary functions of such systems is to gather large amounts of data about the precise locations of the vehicle and all other vehicles, pedestrians, cyclists and other objects around it, as well as status of any traffic signals, etc. So as long as the self-driving system stores at least a few minutes of data and retains it all in the event of an accident, the precise sequence of events and the decisions made by the self-driving vehicle can be reconstructed in tremendous detail -- orders of magnitude more accurately than any eyewitness testimony. This data can be evaluated, the root cause precisely determined and liability accurately assigned.
The evaluation process can even include replaying the data in simulation with slight adjustments to the behavior of every participant to see what scenarios could have changed the outcome. Even if the self-driving system was not at fault, if it turns out that a change in its behavior could have avoided the collision, the software can be updated and the incident simulated to verify that the change works correctly -- and then the updated software can be regression-tested against simulations of all of the other recorded accidents, as well as the full suite of tests developed during design and production. If the update is an improvement overall, it can then be pushed to all vehicles using the system, making all of them better drivers at once.
So the only regulation that might be needed is regulation requiring that self-driving vehicles create this record in a sturdy, tamper-resistant store, analogous to an airplane's "black box". At this stage in development, all self-driving vehicles already have something like this (modulo the tamper resistance, perhaps), because they're actively being refined. There's no reason ever to remove it so the regulation may not even be necessary.
One other bit of regulation that may or may not be useful (or required) is to demand that self-driving system makers share these accident logs with their competitors. Given that they probably won't all be using the same sensor systems, it's not clear how useful the logs would be, but perhaps they could extrapolate a physical model from the sensor logs and then compute the inputs that their sensors would have received from that physical model. If this works well enough, it could enable all self-driving vehicles to get better, and perhaps the NHTSA could use the method to produce a standard test suite that all manufacturers have to pass, though I'd worry about the risks of transformation fidelity and overtraining so that may not be as useful as it appears.
And the car owner, and the manufacturer of the car and of the software. There may be many problems with self-driving cars, but finding lawsuit-targets ain't one.
Sure, you can sue all of them, but if the self-driving system made a bad decision and caused an accident, the maker of that system is the only place courts could reasonably pin the liability.
Who do you sue when a driverless car runs you over?
The maker of the self-driving system. Nothing else would make sense. In this case the vehicles will be owned and operated by Waymo, so there's really no question. If a self-driving vehicle were privately-owned, unless the self-driving system maker can show that the owner of the vehicle did something wrong (didn't keep the software up to date, didn't maintain the mechanical functions of the car, actively interfered with the system during operation, causing the accident), then there's no entity other than that company who could be liable for any accident caused or contributed to by the self-driving system.
People always ask this question, but the answer is obvious.
If the work's getting done then you're doing the job you were hired for.
The mechanism doesn't matter.
However, if your script is doing the job and you're doing nothing, then your manager is not doing the job they were hired for. They should see what you've accomplished, give you a bonus or raise for your efficiency and assign you some more tasks to automate.
Version numbers should mean something. Major number means a major revision or change on key components. Minor is just for patches and fixes.
For libraries, I completely agree that versioning is crucial. In library version numbers:
For applications, operating systems, and whatnot? Meh. Having an understandable linear ordering is nice (i.e. not 3.5 -> 3.5.1 -> 4.0 -> 2000 -> XP -> Vista -> 7 -> 8 -> 8.1 -> 10), but only so you can tell which one is newer. For stuff that is released annually or less often, I think year of release is the best option.
While I agree there's a need for a review/recommendation system that can't be gamed by bad actors, nothing about this one indicates it will be any better than the rest.
It sounds like in the long run the intention is to combat fake reviews by surfacing only reviews from people you have a social connection to. There are obvious problems with that, among them that it only works if people you're connected to (a) use the app and (b) have posted an opinion on something you're interested in, but it seems less prone to gaming.
covfefe
Is this system President Trump's backup plan, in case Twitter ever gives him the boot?
Taxis never cost more than Uber/Lyft where I live.
They also would never show up when called. Uber/Lyft made a huge difference here. I don't really know what the appeal was in bigger cities with a functional cab market, but in my little city (Wilmington, DE) the appeal was a ride within 20 minutes at 2am or 6am both. I saw someone wait over 2 hours for a cab that they scheduled where I work. They pretty much onky show up for airport runs, and you can get them at the train station, maybe a hotel if they have a good relationship with a driver.
Sounds like regulation forced taxi prices to an artificially-low level, resulting in an insufficient number of taxis. Uber/Lyft tune their prices to a more "natural" level, making them more available.
Personally, I'd like to see an Uber/Lyft competitor that doesn't set prices at all, but instead facilitates a real-time auction market between drivers and riders, so that prices are truly supply/demand driven -- and drivers are inarguably independent businesspeople.
I still wonder what happened to uber. Their rates have gone up, not down, yet drivers seem to be making substantially less than just a few years ago.
They figured out how to squeeze more out of their drivers while paying them less.
This answer begs the question. How did they squeeze more out of the drivers while paying them less?
Why do they need to hit zero?
Zero is not enough. The plan set out in the Paris Accord assumes that we'll go beyond zero and begin taking CO2 out of the atmosphere.
so where exactly along this long line of production do you expect to emit a negative value to overcome nature
It's not that we need to "overcome nature", it's that we need to overcome history. We've already put a lot of CO2 into the atmosphere and because the natural processes that remove it are very slow the only way to get the climate back into a normal [*] state and stop the warming effect (much less reverse the warming that has already occurred) is to lower the present CO2 levels faster than nature will do it for us. This assumes we don't use other geo-engineering techniques to reduce warming.
Unless we start photosynthesizing ourselves, we'll naturally emit.
Not unless we eat coal. Sure, when you eat an apple you end up combining a portion of the carbon in the apple with atmospheric oxygen, then emitting the result as CO2. But the carbon came from CO2 in the atmosphere, absorbed by the apple tree. Human (and other animal) CO2 emissions are part of a cyclical process that doesn't make a significant net change in atmospheric CO2. What does make a change is digging up long-sequestered hydrocarbons, burning them to release energy and dumping that carbon into the atmosphere.
[*] "Normal" is a funny word here. There is no such thing as a "normal" Earth climate. The planet's climate changes all the time. Usually -- but not always -- over very long timeframes relative to human perception, but still, it's always changing. All of recorded human history has occurred in a brief warm period of the Quaternary Ice Age. But the point is that we know and like it the way it has been, and in fact we want it to stay this way forever (or at least for a very long time). We don't want it to get hotter and force us to move massive chunks of our population and figure out how to change our food production approaches, which is what we're currently facing. We also don't want it to get much colder. Luckily, we have now proven that we can "engineer" a warming climate, so we can ensure that we never have to deal with colder temperatures. We did it by accident, but we could do it again on purpose if needed. Now we need to figure out how to stop warming, and perhaps even cool the planet a bit. Thus, we can halt the natural fluctuations of the planetary climate forever and lock it permanently in the state that we want it; the one we call "normal".
And the solution is always: exterminate capitalism immediately, implement worldwide socialism, the United States must pay the staggering cost of the whole thing by impoverishing its people, and we need global governance without any voting or input from the deplorables. The whole thing might be more believable if the "solution" wasn't always the same thing.
Basically nobody is proposing that "solution" beyond a few fringe wackos who think it's a good idea regardless of climate change but are happy to jump on this particular bandwagon. Certainly, market-driven approaches like carbon credit trading and carbon taxation are absolutely nothing of the sort, nor are voluntary international agreements like the Paris Accord.
Market-based capitalism is by far the best system we have for accurately optimizing economic activity. The powerful signalling mechanism that prices provide does a better job than anything else of eliminating waste in production and redirecting resources to the most useful sectors -- where "useful" is determined by the needs and desires of the population as a whole. Not that markets are perfectly efficient, but they're far better than anything else we've found. But markets don't consider externalities, so we need some way to make carbon emission a cost of production. Capital will then flow away from carbon-emitting processes and to carbon-neutral -- and eventually even carbon-absorbing -- processes. Moreover, markets price the future as well as the present, so it's not even necessary to impose the carbon costs right away to start seeing immediate benefits. If you enact a plan to impose carbon taxes (or similar) and convince markets that the plan is not going to be rescinded or relaxed, capital reallocation will start even before the costs hit, as industries and the moneymen who back them seek to minimize the shock of the change.
To address the climate change problem, we need to apply the power of capitalism, not exterminate it.
Here's the situation in the UK: https://en.wikipedia.org/wiki/...
So they haven't attempted to codify it, just established a panel that applies some general guidelines. The only real rule is that the individual must have transitioned (the meaning of which isn't really defined and I strongly suspect has shifted and will continue to shift over time) two years prior to issuance of a gender recognition certificate.
We can expect all of this to continue being further muddied as additional notions about gender arise, particularly the expanding list of non-binary categories.
I repeat, cite?
I think that this apparently "statistical" nature is a function of something very fundamental that we don't yet understand, and some day, somebody will find the reality with a slap to the forehead and a loud "D'OH!"
I see no reason to expect this.
Basically, what bothers you is that our intuition, which is formed from macroscopic observations made through our physical senses which were optimized via evolution to enable our survival in a world of macroscopic objects, doesn't map well onto the world of subatomic particles. But why in the world would you expect it to?
No, I do not think laws such as this would fall under the measures to eliminate discrimination” category, nor that they would do any good.
That's a nice statement of what you believe. Can you elaborate on your rationale? Without an explanation of the basis for your belief, your statement is merely a vote, not an argument to sway others' opinions. This is a discussion, not a poll.
No we're saying that the best candidates *should* be chosen...
But there's no objective "best" in this case. There are highly-desirable attributes which can be objectively identified and measured, but it's simply impossible to create an objective stack ranking of all available candidates, male and female. Rather, you have a large pool of candidates, all highly qualified and all roughly equal in ability. Each candidate has particular pros and cons, but it's rarely the case that you can clearly identify that any one of them is strictly better than any other.
In that sort of context, people fall back on personal relationships to make choices, even when they're doing their level best to be objective, because there is no objective way to choose. Which means that in a male-dominated system, men have a systemic advantage, because the male dominance means that men have more of the relevant personal relationships.
It's important to recognize that it's possible for a system to be biased even when none of the participants is biased and even when there are no explicitly-biased rules in place. How do you remove the bias from such a system? Perhaps the only way in some cases is to inject an explicitly counter-biased rule. Applying intentional discrimination to counter unintentional discrimination is paradoxical, but I don't see any other workable approaches on the table.
Note that I strongly doubt the effectiveness of reverse discrimination applied on larger scales where the numbers mean that it will be necessary to elevate less-qualified applicants. That results in stigmatization and often exacerbates the problem. In this case, though, the numbers are small and I think the sort of women who will be chosen for these boards will be more than capable of holding their own, so I don't think we need to worry about this action harming them or creating a perception that female boardmembers are mere tokens.
Sometimes it really is for the public good. The key is to put substantial due process around it, and codify extensive protections against abuse, and guarantee adequate compensation for the person whose property is taken. I think the US generally does a decent job with its handling of eminent domain. Not so much with asset forfeiture.
Does it really work like that in California? Most places it's not just enough to declare you are female, you have to actually live as a female. Shave your beard, wear women's clothing, change your name etc.
Cite? This may be practice, but I'd be really surprised to see it as codified law anywhere, and I'd love to read the statute if it is. I can imagine someone trying to draft such a law getting wrapped around the axle in all the corner cases and details. It seems extremely difficult to craft.
Nowhere does the freedom of speech mean that you're entitled to whatever platform you like and however you like it
Certainly not. But deplatforming doesn't mean that no one is obligated to give you a platform, it means mobs rising up to forcibly reverse your invitation to speak, merely because some of them might be offended.
And yes, the fewer Nazis or Racists out there, the better.
I certainly agree. But I disagree that violent suppression of their political speech is either necessary or effective.
I support deplatforming all extremists. Nazis and commies alike.
Deplatforming is a euphimism for silencing. I support freedom of speech.
And "indoor".
It was indoor even after it blew the door off of the enclosure.
So if they have a 90 day announcement period for more stock sales, they would need to announce by the end of October. Having yourself locked down at that time is basically a death sentence.
Is there any indication that the suit against Musk would "lock down" Tesla?
Yeah, the problem isn't the physical building tech (on the moon). It is just building stuff. On the moon. Did I mention it was on the moon? Space nutters.
Calm down. No one is asking you to do it.
If it's simple enough to do in your head, it's simple to reverse engineer.
I use a formula for all of my password websites (but it's actually very different to the one above, I don't think anyone could ever reverse engineer my password to figure out my formula
If I saw your password for two sites I'd know generally what you're doing, and what the invariant portion is. What's left is a 4-6 character alphanumeric password; maximum entropy 31 bits. You really don't gain a great deal with the invariant string, other than fooling password strength meters, and maybe yourself.