I thought Fakespot was an interesting premise for using LLMs in a vaguely legitimate way, but I did get the sensation that by using it, it would eventually come out years later that it was doing something nefarious, similar to Honey.
My assumption was that it would either teach fake reviews to get much better to be harder to identify...or it'd get sold off to some company that had a vested interest in keeping fake reviews around. I suppose this second possibility could still be the case! :)
I was always suspicious of Fakespot's ratings and had very little ways to properly A/B test. I would see products on Amazon - both with 4.5 stars - but one would have a Fakespot score of 1 star and the other would be 4 stars.
After that Honey expose I started to wonder if Fakespot was simply trying to steal the last click referrer and using its "we filter fake reviews" as a cover/lure to get people to use it.
I can see why it needs to be shut down: I think that Amazon has almost certainly worked to limit the usability and usefulness of such an app.
Plus, sellers have more tools than ever at their disposal to generate "real" reviews, including plain and simple bribes to real customers to make them, completely unenforceable by Amazon.
No, you actually want to reverse-engineer the PCB so that you have schematics. That allows you to more easily replace components with cheaper alternatives and/or modify the firmware to continue working on your unintended hardware. (Fake products typically replace the on-board components with cheaper variants, like using lower-rated caps, to squeeze out a tiny bit more profit margin. Like really in the $1 per 10k units range.) Plus PCB x-ray is a common thing that also reputable manufacturers do for quality control or security assurance (e.g. Is there a HW backdoor in this mainboard?) so you can just book that as an affordable service. I remember Siemens recently bought a PCB RE provider, but I can't find the news article anymore.
To give you some price ideas:
10x10cm 4-layer PCB x-ray and RE: $200
STM32 firmware dumping: $700
EEPROM dumping: $300
The STM32 is the most expensive because you need to decap them to get the hardware encryption key. But it's still A LOT cheaper than building your own firmware from scratch.
The only stuff I tend to bookmark are things I want to use as a keyword search, everything else is a tab in a certain window. There's not enough of them that I have to search to find, maybe a fifty in total?
There is the rare bookmark in my toolbar tabs for undetermined future stuff.
It's a rare stretch of American interstate highway where the majority are sticking to the speed limit. On different roads the ratio of speeders to "speed limit doers" varies, but interstates have fairly consistent driving conditions and the safe speed for those conditions is consistently higher than the posted limit.
Usually traffic on American interstates is staying at the speed limit if: there is a fair amount of congestion, everybody is stuck behind somebody doing the limit, or there is rain/fog/snow/etc.
>and the safe speed for those conditions is consistently higher than the posted limit.
A cursory look at traffic safety statistics would seem to dispute this statement. The US consistently ranks quite high in traffic fatalities, which seems to indicate that you couldn't safely drive faster than the current speed limit in most circumstances. Of course you'd have to control for a lot of other factors for a definite conclusion, but conversely, I'd like to see an actual source for your theory.
Enforcing a speed limit is actually not that hard if you want to. You can of course put stationary or mobile speed cameras in place. But even better may be systems I've seen in some places in southern Europe: Cameras at the beginning and end of a stretch of highway that compare time stamps of your passing and calculate the average speed of your car in that stretch. If you've passed the stretch quicker than should be possible according to the speed limit you'll be fined. It's quite low tech, cameras are prevalent on the highway systems of many countries, anyway, and it solves the problem of people adhering only to the speed limit if they know they are being observed by law enforcement.
There are a number of interstates (and interstate like US highways) near me that have posted speed limits of 70MPH but any normal day you'll find traffic going 55mph or even less.
When everyone is driving beyond the speed limit, the ones actually obeying the speed limit are the dangerous drivers. It is unfortunate that speed limits in the US have not corresponded to how people actually drive since 1973.
How does this make sense? If I'm driving the speed limit and someone else is crashing into me from behind while speeding, they are the dangerous driver. No matter how many other people also ignore the speed limit.
Also, increasing the speed limit does nothing to make traffic safer. That doesn't make any sense at all, as increased speed is correlated very well with increased accident rates and severity of traffic-related injuries:
It should be obvious that driving substantially slower than everyone else would make you a danger to others. Try driving at the speed limit on a high way in southern NY, especially in the left lane. You will have many near accidents and the reality is that you would be the dangerous driver for not keeping pace with everyone else.
Increasing the speed limit to the 85th percentile so that you do not have the few people who actually obey it posing a hazard to others does make things safer.
Getting cars off the road sooner by reducing travel time, decreases the number of cars on the road. This increases the distance between cars and accidents only when the distance betweeen a vehicle and something else reaches 0. Forcing people to drive slower therefore causes collisions by bringing cars closer to one another.
The severity is a separate matter from whether there is a collision. As for severity, people drive much faster in Germany where there is no high way speed limit for much of the autobahn yet their autobahn network has half the fatalities that U.S. highways have. The safety data from 2012 shows this:
As for your link, it talks about pedestrian safety. As per the data there, pedestrians are unsafe on highways no matter what the speed limits are. There are also no pedestrians on highways. There is no point to setting highway speed limits based on studies showing the danger to non-existent pedestrians.
Did you post the first link that seemed to agree with your position as part of some fallacious appeal to authority because logic failed to agree with your preconceived notions that you were never equipped to defend? I suspect that is exactly what you just did.
>Increasing the speed limit to the 85th percentile so that you do not have the few people who actually obey it posing a hazard to others does make things safer.
[citation needed]
Seriously. That is a pretty bold claim that you should be able support with actual studies, if true. You have several things working against your hypothesis:
- While it may be true that driving at the legal speed limit might slightly increase the risk of accidents when many other drivers drive faster than the speed limit, a general increase of the speed limit might severly increase risk for everyone.
- Many people do not drive above the speed limit, because they have a "higher normal", but because they feel entitled to "drive faster". I.e. they claim that they are "better drivers", they have a superior need to arrive faster, etc. Those people will just adapt their behavior to the new, higher speed limit and again drive above that, making the entire exercise pointless and dangerous.
- A potential and less risky alternative would be improved enforcement of the current speed limit.
There are probably plenty of other arguments that actual experts in this field would bring up.
>Getting cars off the road sooner by reducing travel time, decreases the number of cars on the road. This increases the distance between cars and accidents only when the distance betweeen a vehicle and something else reaches 0.
That makes no sense at all. It has been proven in both theoretical and practical tests that driving "all out" does not decrease travel time drastically in almost all circumstances. But it substantially increase the risk of accidents. So any minuscule decrease in car density will be far outweighed by increased accident risk per kilometer driven.
>Forcing people to drive slower therefore causes collisions by bringing cars closer to one another.
It's actually the other way around: Forcing people to drive slower drastically reduced the risk of collisions because people are slower and have more time to react. It also reduces the incidence of traffic jams (because sharp braking prevalent with speeding drivers is a main contributor to traffic jams), which are in turn a major factor in collisions.
>As for severity, people drive much faster in Germany where there is no high way speed limit for much of the autobahn yet their autobahn network has half the fatalities that U.S. highways have.
You'll have to control for other factors, of course. Cars in the US are much larger and heavier than in Germany, for example. In statistics, you want to compare apples to apples, so you control for vehicle weight when trying to make observations about the impact of speed on accident severity.
>As for your link, it talks about pedestrian safety.
True. But the same holds true for vehicle collisions. It's really basic physics. All other things being equal, faster cars have more energy. More energy = more severe accident outcomes.
>Did you post the first link that seemed to agree with your position as part of some fallacious appeal to authority because logic failed to agree with your preconceived notions that you were never equipped to defend?
No. There are plenty of other sources that support my views, i.e.
> Seriously. That is a pretty bold claim that you should be able support with actual studies, if true
The 85th percentile rule/principle has been understood for decades. Just search for information on it. You will find tons of results. Calling it a bold claim is like claiming asymptotic complexity is a bold claim. It is something that is well known, just not to you.
> It has been proven in both theoretical and practical tests that driving "all out" does not decrease travel time drastically in almost all circumstances.
Those tests do not seem relevant to highways, where it is easy to measure differences in travel time between driving at the speed of traffic and driving at the speed limit. When traffic is at 70mph and the speed limit is 55mph, keeping pace with traffic results in a 21% reduction in highway travel time. How things go when someone is ‘driving "all out"’ is not relevant here.
> Forcing people to drive slower drastically reduced the risk of collisions because people are slower and have more time to react.
A highway is not a regular road where it is stop and go based on lights. The purpose of a highway is to have a free flow of traffic such that you do not need to be continuously reacting to others. You do need to maintain a certain distance between you and the car to react in emergencies, but these are supposed to be exceptional and plenty of collisions occur when changing lanes, which would be lessened with fewer cars on the road. Cases where everyone needs to stop would also be lessened.
> You'll have to control for other factors, of course. Cars in the US are much larger and heavier than in Germany, for example. In statistics, you want to compare apples to apples, so you control for vehicle weight when trying to make observations about the impact of speed on accident severity.
Those same vehicles are legal to drive in Germany as far as I know. There is a possibility that they are popular in the U.S. because of the speed limits such that they would be less popular if the highways did not have speed limits. After all, their acceleration, braking and fuel economy are terrible. They would only be worse at autobahn speeds. The knowledge that it is legal to drive at higher speeds tends to encourage people purchasing vehicles to purchase ones that can handle higher speeds well. We could see vehicles more similar to those driven in Germany become popular if there were no speed limits and then things would naturally become apples to apples.
> True. But the same holds true for vehicle collisions. It's really basic physics. All other things being equal, faster cars have more energy. More energy = more severe accident outcomes.
There is no law of physics that dictates that such things cannot be done with greater safety than we currently have. Germany is a fantastic example of this. Germany permits speeds that would be considered hazardous by the thinking behind motor vehicle rules in the US, yet is substantially safer.
> No. There are plenty of other sources that support my views, i.e.
If I tell you what is wrong with your sources one last time, I hope you will stop posting links under the misguided hope that some random thing superficially agrees with your claims sticks. Your first link involved studies in a country where speed limits should obey the 85th percentile. The findings are not relevant to the U.S. where the 85th percentile is ignored. Even without knowing about the 85th percentile, it is obvious the applicability to other countries would depend on how similar the process for establishing the speed limit is. Your second link is behind a paywall and cannot be scrutinized, but the German autobahn likely contradicts it. Problems only occur when the distance between a vehicle and another object reaches 0. If that is avoided, the speed does not matter.
That said, it is impossible to prevent future Darwin Award recipients from earning their awards. If you insist on trying to stop them from earning rewards from motor vehicles, you might as well push for a complete ban on motor vehicles. That is the only thing that would eliminate motor vehicle fatalities.
I cannot speak for other states, but NY requires a driving test and has police regularly enforce traffic violations too. None of this supports the idea of setting highway speed limits below the 85th percentile or even having high way speed limits. It instead suggests that we should copy the Germans.
As for gruesome accidents, there will always be Darwin Award recipients. Trying to prevent them from earning their rewards is a foolhardy task. I believe the maxim is that the moment you make something idiot proof, the world makes a better idiot.
reply