Im scared for this space launch

Candle Power Forums

Help Support Candle Power:

Those guys are nuts, too.
Agree.
I don’t quite understand why some are always trying to reach some utopian perfection. As humans we’re quite imperfect. For an example; sports. Why do we need some computer to tell us if it’s a ball or strike? If it was offsides? Humans are playing the game, not robots. Why do we need AI taking over jobs or driving our vehicles. We’re taking the humanity out of life and it’s very strange to me. Somehow we’ve survived being imperfect for a long time. I’m not sure changing that is a good choice.
 
Won’t happen in our lifetime.

Never say never. The bean counters are the ones who will probably make it happen in the end. Road injuries/deaths are a HUGE drain on the economy, not to mention tragic on a human scale.

Getting back on topic, I've read we'll probably send unmanned craft to Mars before people to build a habitat. Unlike the moon where we only stayed a short time, orbital mechanics dictate we'll be on Mars for months.
 
Agree.
I don’t quite understand why some are always trying to reach some utopian perfection. As humans we’re quite imperfect. For an example; sports. Why do we need some computer to tell us if it’s a ball or strike? If it was offsides? Humans are playing the game, not robots. Why do we need AI taking over jobs or driving our vehicles. We’re taking the humanity out of life and it’s very strange to me. Somehow we’ve survived being imperfect for a long time. I’m not sure changing that is a good choice.
I'm on the fence about calling balls and strikes. I'm guessing a machine will do it more accurately than a human. Isn't that what we want in order to ensure a completely fair game?

AI taking over driving vehicles is 100% about safety. To me I don't understand how 40K to 50K annual road deaths in the US alone, several million worldwide, was even remotely acceptable. That's more than one 9/11 every single month! If that many people died in planes or trains the NTSB would shut them down until the problem could be fixed. Dying while getting from point A to point B is the most pointless, useless way to die. I might make an exception if you're doing something cutting edge like testing a new plane or going into space. At least there you're paving the way to make things safer for those who follow.

A lot of AI and robots taking over jobs is either about safety, or relieving people from repetitive work so they can focus on creativity. By increasing productivity we can shorten the work week. I enjoy many things, but there's literally nothing I would want to do 40+ hours a week, week in and week out. If people work one or two days a week on average then they'll have a much better work/life balance. For some not having to work at all is even better if they're self-starters. I think being a modern corporate drone like a robot tied to a clock is dehumanizing, not the opposite. I only knew one person who wouldn't quit their job in a heartbeat, and he was running his own company.

Like it or not automation is coming. Every expert has said that. What's the alternative? Keep people in make-work jobs which many of them hate anyway?
 
Last edited:
I'm on the fence about calling balls and strikes. I'm guessing a machine will do it more accurately than a human. Isn't that what we want in order to ensure a completely fair game?

AI taking over driving vehicles is 100% about safety. To me I don't understand how 40K to 50K annual road deaths in the US alone, several million worldwide, was even remotely acceptable. That's more than one 9/11 every single month! If that many people died in planes or trains the NTSB would shut them down until the problem could be fixed. Dying while getting from point A to point B is the most pointless, useless way to die. I might make an exception is you're doing something cutting edge like testing a new plane or going into space. At least there you're paving the way to make things safer for those who follow.

A lot of AI and robots taking over jobs is either about safety, or relieving people from repetitive work so they can focus on creativity. By increasing productivity we can shorten the work week. I enjoy many things, but there's literally nothing I would want to do 40+ hours a week, week in and week out. If people work one or two days a week on average then they'll have a much better work/life balance. For some not having to work at all is even better if they're self-starters. I think being a modern corporate drone like a robot tied to a clock is dehumanizing, not the opposite. I only knew one person who wouldn't quit their job in a heartbeat, and he was running his own company.

Like it or not automation is coming. Every expert has said that. What's the alternative? Keep people in make-work jobs which many of them hate anyway?
As far as baseball (or other sports) the players make many mistakes every game. Should we replace them with robots? Again, fairness is not a measuring stick it is a mindset. We know players cheat so the fairness thing doesn’t wash with me.

We absolutely need to work. We were hunters and gatherers before this modern age. We gain harmony when we are working by seeing the fruits of our labor. Many people, after retirement, return to the workforce because they miss the positive results of what it brings. It’s not dehumanizing it’s humanizing, it’s what we are.

Now, there are extremely dangerous jobs, like checking to see if a bomb is live or needs to be deactivated, that I am all for robots to do. Go for it.

Automation of the likes you are hoping for will destroy humanity. Be careful what you wish for.
 
As far as baseball (or other sports) the players make many mistakes every game. Should we replace them with robots? Again, fairness is not a measuring stick it is a mindset. We know players cheat so the fairness thing doesn’t wash with me.
I think maybe consistency is a better word for it. I want consistent officiating in any sport. It's why sports where time matters are timed by automation now, not people with stopwatches. I know players make mistakes and that's part of the game. The other side capitalizes on those mistakes. If we played baseball with robots every game would probably be a perfect game with no hits, no runs, no walks. Who would want to watch this?
We absolutely need to work. We were hunters and gatherers before this modern age. We gain harmony when we are working by seeing the fruits of our labor. Many people, after retirement, return to the workforce because they miss the positive results of what it brings. It’s not dehumanizing it’s humanizing, it’s what we are.
We need to do something. I agree with that. It doesn't need to be a full-time job. My late father seemed bored during retirement. He spent a lot of time watching TV or doing "hobbies" which weren't particularly engaging, mostly just collecting stuff. I could make the argument that many of the people bored during retirement are this way because decades of doing a mindless job on a set schedule destroyed their initiative. Indeed, I saw that happening to me after working only a few years. I said to myself I have to get away from this. I did by starting my own home business where I work my own schedule and probably never averaged over 15-20 hours a week. That's an average, mind you. I've had busy weeks where I put in 70 hours but also had weeks or even months completely off. I was never bored.

I think my father's job killed all his initiative over the decades. He used to do a lot of fun, creative stuff when I was a kid. That got to be less and less as he got older. I even offered to get him a computer but he had no interest in learning anything new. Really sad.
Now, there are extremely dangerous jobs, like checking to see if a bomb is live or needs to be deactivated, that I am all for robots to do. Go for it.
Collecting garbage is one of the statistically most dangerous jobs believe it or not. It also kills your back. Delivering food is another very dangerous job, especially in NYC. Long-distance trucking is dangerous also, more so to the people in cars who are at the mercy of tired, overworked truck drivers in charge of an 80,000 pound semi.

It's not just dangerous jobs which need to be replaced but also boring ones.
Automation of the likes you are hoping for will destroy humanity. Be careful what you wish for.
There I have to disagree. I think it'll usher in a new golden age of humanity where people can choose their own paths, free of the burden of needing to do something just to pay living expenses. I see what people come up with watching you-tube. All sorts of creative pursuits. Many of these people are only able to do this because they're free of needing to work. I think we'll become a society of artisans when automation takes over. And don't think there won't be a demand for human-produced goods even once the machines can make virtually everything we need. The slight imperfections of human-made furniture, for example, will have appeal to some over uniform stuff made by machine.

And of course we'll start exploring the universe. That will keep us occupied literally forever.
 
As an Amazon Associate we earn from qualifying purchases. Product prices and availability are accurate as of the date/time indicated and are subject to change.
@jtr1962 I can't help but notice in your idtechex link that the first couple of lines read "will match or exceed human safety by 2024"...aren't there still "ghosts" in some of the systems?
Of course there are. By and large AI driving already is safer than human driving, although there are still some problems to be worked out. The main problem is operating on urban streets. By its nature this is far more unpredictable than highway driving.* Also, the majority of AI vehicle crashes are due to a human driver doing something stupid or unpredictable. If all vehicles were AI this wouldn't happen.

* Urban driving is still a work in progress. Dealing with pedestrians and cyclists, both of whom are very unpredictable, is the primary problem. The solution will probably be some combination of better prediction and simply slowing to safe speeds (i.e. 10 mph or less) when passing pedestrians or cyclists.
**and to be flippant/facetious, the safest automated vehicle is the same as the safest human driven vehicle, one that isn't there.
True. IMO the best way to deal with this is to simply limit or ban vehicles in areas with many pedestrians or cyclists. It's the concept of inherent safety versus operator dependent safety. That really applies to many facets of safety, not just road safety. We try to set up factory processes so even if a human makes a mistake it can't result in disaster. Or modern planes simply won't let the pilot fly them into a mountain. Or modern cars with pedestrian avoidance systems. Even if we don't have full automation we can have some failsafes which override poor choices by a human operator.
 
Last edited:
Also, the majority of AI vehicle crashes are due to a human driver doing something stupid or unpredictable. If all vehicles were AI this wouldn't happen.
DOT. DOT. DOT. I couldn't help but spell out the dots this time. I will henceforth refer to you as Agent Smith...
 
DOT. DOT. DOT. I couldn't help but spell out the dots this time. I will henceforth refer to you as Agent Smith...

1489595617-1489592550-600px-matrixdeserteagle-4.jpg
 
Of course there are. By and large AI driving already is safer than human driving, although there are still some problems to be worked out. The main problem is operating on urban streets. By its nature this is far more unpredictable than highway driving.* Also, the majority of AI vehicle crashes are due to a human driver doing something stupid or unpredictable. If all vehicles were AI this wouldn't happen.

* Urban driving is still a work in progress. Dealing with pedestrians and cyclists, both of whom are very unpredictable, is the primary problem. The solution will probably be some combination of better prediction and simply slowing to safe speeds (i.e. 10 mph or less) when passing pedestrians or cyclists.

True. IMO the best way to deal with this is to simply limit or ban vehicles in areas with many pedestrians or cyclists. It's the concept of inherent safety versus operator dependent safety. That really applies to many facets of safety, not just road safety. We try to set up factory processes so even if a human makes a mistake it can't result in disaster. Or modern planes simply won't let the pilot fly them into a mountain. Or modern cars with pedestrian avoidance systems. Even if we don't have full automation we can have some failsafes which override poor choices by a human operator.
I’m just going to speak about your first sentence: “AI driving is already safer than human driving”. That is unprovable. Testing one AI driven machine in one town is not proof that if there were a lot more they would be safer everywhere. And let’s not forget about reliablity. What if the system crashes or makes just one mistake? You can’t make comments assuring something that doesn’t exist except in your mind. Life doesn’t work that way.
 
I’m just going to speak about your first sentence: “AI driving is already safer than human driving”. That is unprovable. Testing one AI driven machine in one town is not proof that if there were a lot more they would be safer everywhere. And let’s not forget about reliablity. What if the system crashes or makes just one mistake? You can’t make comments assuring something that doesn’t exist except in your mind. Life doesn’t work that way.
It's not only provable, but already proven beyond any doubt. In real cities and highways worldwide, there's been about 100 million miles, 10 million hours, of autonomous driving.
Compared to human drivers (matched environments):
  • ~70–85% fewer injury-causing crashes
  • ~50–60% fewer police-reportable crashes
  • ~90% fewer crashes involving pedestrians
 
You can’t make comments assuring something that doesn’t exist except in your mind. Life doesn’t work that way.
Sure, he can. He does it all the time. Haven't you noticed? I consider that attempting to reason with him should've been included in Dante's Inferno. 👿 It's definitely a concentric circle of torment. 😅
 
I’m just going to speak about your first sentence: “AI driving is already safer than human driving”. That is unprovable. Testing one AI driven machine in one town is not proof that if there were a lot more they would be safer everywhere.
This tells me we just need to do a lot more widespread testing. If it has fewer incidents per million miles after testing in multiple scenarios then it's definitely safer. AI driving is also a moving target. Every time a crash happens you analyze why it happens and reprogram every other vehicle with AI so that particular scenario never happens again. Human driving will always remain comically dangerous. For starters we simply refuse to train people at the level needed to truly operate much more safely (think something like getting a pilot's license which requires thousands of hours of study and regular recertification). For another thing the majority of people simply lack the coordination, spatial awareness, intelligence, or proper attitude to safely operate a multi-ton high-speed vehicle. Most people shouldn't operate anything heavier or faster than an e-bike.
And let’s not forget about reliablity. What if the system crashes or makes just one mistake? You can’t make comments assuring something that doesn’t exist except in your mind. Life doesn’t work that way.
There are working commercial systems. Texas is operating driverless trucks on highways now. If the safety record was worse than human drivers they wouldn't be able to get insurance.

What if the system crashes or makes just one mistake? The answer to the first part is modern embedded systems are much more reliable than the general purpose PCs which often crash. We trade some flexibility for greater reliability. Modern human driven cars are now often drive-by-wire. We never would have implemented this if the systems weren't extremely reliable.

The answer to the second part is in my first paragraph above. If there's a mistake, engineers reprogram all the AI vehicles so that mistake never occurs again.

I've been involved with livable streets and traffic safety as a hobby now for well over a decade. The human is the weakest link in the system by far. If you were a victim of traffic violence like myself you might be more willing to embrace things which will end or reduce it. To me the best answer would be simply to remove large, heavy, fast vehicles from close proximity to pedestrians and cyclists but this is nowhere near being remotely practical. Maybe in 50 or 100 years but not now. So the alternative is to remove the human from the equation since we seem reluctant to have realistic driver licensing standards.
 
It's not only provable, but already proven beyond any doubt. In real cities and highways worldwide, there's been about 100 million miles, 10 million hours, of autonomous driving.
Compared to human drivers (matched environments):
  • ~70–85% fewer injury-causing crashes
  • ~50–60% fewer police-reportable crashes
  • ~90% fewer crashes involving pedestrians
Please reread my comment. Fill an entire city with only AI vehicles and then I’ll listen.
 
It's not only provable, but already proven beyond any doubt. In real cities and highways worldwide, there's been about 100 million miles, 10 million hours, of autonomous driving.
Compared to human drivers (matched environments):
  • ~70–85% fewer injury-causing crashes
  • ~50–60% fewer police-reportable crashes
  • ~90% fewer crashes involving pedestrians
I wasn't aware it had gotten this good. Not perfect but unlike human driving AI driving will only get better with time.
 
Please reread my comment. Fill an entire city with only AI vehicles and then I’ll listen.
In that scenario the safety stats would be even better. The primary reason AI vehicles have crashed is because a human driver does something so stupid the AI didn't predict it. In a nutshell I think we should program AI cars the way I ride my bike. The general algorithm would be assume everyone on the road will do the worst possible thing at the worst possible time. Then leave yourself a potential out for every such scenario. Computers can do this far better than any human.

I've noted my prediction algorithm has holes in it over the last decade for two reasons. One, drivers have gotten even more stupid and unpredictable. Two, the acceleration capabilities of modern cars can get them into trouble far more easily. Back when I first started riding in the late 1970s I think the average vehicle took well over 10 seconds to hit 60 mph. In city driving they weren't all that fast off the line, either. Maybe 3.5 to 5 seconds to reach 30 mph. Even if a driver was completely nuts, the time it took for their vehicle to accelerate and cause me danger was well within my ability to predict and react. Now some cars can get to 60 mph in under 3 seconds. Many can reach 30 mph in only about 1.5 to 2.5 seconds. This has exceeded my ability to predict and react. An AI driver wouldn't have such limitations. A solution might be to GPS limit cars to rather sedate acceleration rates on local urban streets, as well as to cap them at the speed limit. I've proposed something like 3 mph/sec. This would make things much safer.
 
Back
Top