If the AI was not able to pull over then the correct action would be for it to just stop.
Stopping in the middle of a freeway is a violation of the traffic rules in most places, and if there was an accident, Tesla would be to blame. Now the blame falls on the human driver.
when you create a thing and call it "autopilot" you have to figure out how to handle such a situation.
The thing called "autopilot" in a plane can only handle a very limited range of situations. If the real pilots are all incapacitated, it will keep flying until the fuel runs out, and then crash, while sounding the alarm.
It's perfectly anonymous to receive the coins in a fresh address. Spending them is a bit trickier, but if you're not in a hurry, you can do that a few years later, maybe from a open WiFi network somewhere in a parking lot during vacation abroad.
It's an order of magnitude gat between 'all of bitcoin" and 'one bank'.
Western Union isn't a bank, it's the world's leading business for international money transfer, which is a function that bitcoin can provide with less hassle, more security, and lower fees.
The problem is, most tasks that humans do don't require a full-scale human level AI.
Full scale human level AI is actually pretty bad at a lot of tasks. Ask a human to look at all the Google streetview images, and identify every single bit of text. They'll get bored and distracted after a few hours, and start making mistakes, like skipping entire images.
Google's AI platform can finish that task in less than a week with superior accuracy. Of course, it's going to make some hilarious mistakes once in a while, but on average, it's going to outperform any human.
The strict rules of Go don't help to figure out if a position in the middle of the game is winning or losing. Similar to driving. The rules of the road are clear. Figuring out if a dirt road is passable is hard.
If it's "Far from autonomous", they shouldn't be calling it autopilot.
Seems like an appropriate name, actually. A plane on autopilot also just keeps course and altitude, but doesn't fly around storms, or start an automatic landing procedure when fuel is running low.
No... it should pull over and stop as soon as it is possible to safely do so
Finding a safe place to pull over and stop may be beyond the capabilities of the autopilot system. The Tesla was not designed as a general purpose self driving car.
$500 worth of potable water/copper scrap/cotton cloth/wheat is still $500 worth of something with known and stable demand.
Paying the bills with crypto is a lot more convenient and secure than paying with pieces of copper tubing. The convenience and security is what creates value and demand.
Hearing the amount of negativity, I guess we're still in the early adopter phase. Sure, there's a certain amount of mania, but still very limited as a percentage of the population.
I'll settle for rising asymptotically to a stable level. And plenty of people are already using it as a currency. The network is doing about 3-4 transactions per second right now. Compare to Western Union which does 31 transactions per second, it's not a huge gap.
If the probe is small, the sail can also be small. Some proposals are for a probe the size of a pack of cigarettes, or even a postage stamp.
How many useful instruments, including a transmitter capable of reaching us, can you pack in a postage stamp ? And how are you going to do a close flyby of the earth-like planet we would be interested in ?
We announced the TPU last year and recently followed up with a detailed study of its performance and architecture. In short, we found that the TPU delivered 15–30X higher performance and 30–80X higher performance-per-watt than contemporary CPUs and GPUs.
I think that counts as a bit more than "gradual" speed improvement. And that was just the first generation. They've already made significant improvements in the second generation (45 TFLOPS/chip)
And the 3rd reason is quickly improving algorithms. Also, the GPUs are being replaced by TPUs, dedicated for neural processing at much higher speeds and lower power.
Depends on what you call "fundamental". There are plenty of new ideas tried out. The number of research papers written on AI in the last couple of years eclipses all the papers that were written before that.
Slashdot Mirror
If the AI was not able to pull over then the correct action would be for it to just stop.
Stopping in the middle of a freeway is a violation of the traffic rules in most places, and if there was an accident, Tesla would be to blame. Now the blame falls on the human driver.
when you create a thing and call it "autopilot" you have to figure out how to handle such a situation.
The thing called "autopilot" in a plane can only handle a very limited range of situations. If the real pilots are all incapacitated, it will keep flying until the fuel runs out, and then crash, while sounding the alarm.
Yes! How are slashdotters missing this?
I think some people enjoy missing things way too much.
It's perfectly anonymous to receive the coins in a fresh address. Spending them is a bit trickier, but if you're not in a hurry, you can do that a few years later, maybe from a open WiFi network somewhere in a parking lot during vacation abroad.
But none of these currencies will ever replace government-backed currencies. Because they can't assure stable value.
So, how's that working for the Venezuelan Bolivar ?
It's an order of magnitude gat between 'all of bitcoin" and 'one bank'.
Western Union isn't a bank, it's the world's leading business for international money transfer, which is a function that bitcoin can provide with less hassle, more security, and lower fees.
Is it tedious and unhelpful to point that out for EVERY article on AI tech?
It's especially tedious and unhelpful if the article did not actually make a mistake with using the term AI.
The problem is, most tasks that humans do don't require a full-scale human level AI.
Full scale human level AI is actually pretty bad at a lot of tasks. Ask a human to look at all the Google streetview images, and identify every single bit of text. They'll get bored and distracted after a few hours, and start making mistakes, like skipping entire images.
Google's AI platform can finish that task in less than a week with superior accuracy. Of course, it's going to make some hilarious mistakes once in a while, but on average, it's going to outperform any human.
This new HTTP hype is just another cycle that will go away once the VCs grab a few dollars.
Fixed that for you.
The strict rules of Go don't help to figure out if a position in the middle of the game is winning or losing. Similar to driving. The rules of the road are clear. Figuring out if a dirt road is passable is hard.
4 out of 5 doctors approve: https://i.redd.it/wfbt456v0wqy...
over 20 years ago it was hotter in Phoenix, 122 degrees.
Yes, also related to global warming. Local temperatures are a noisy signal on top of a gradual rising trend.
If it's "Far from autonomous", they shouldn't be calling it autopilot.
Seems like an appropriate name, actually. A plane on autopilot also just keeps course and altitude, but doesn't fly around storms, or start an automatic landing procedure when fuel is running low.
That would be the proper thing to do for a full autonomous car, not for car with glorified lane assist.
No... it should pull over and stop as soon as it is possible to safely do so
Finding a safe place to pull over and stop may be beyond the capabilities of the autopilot system. The Tesla was not designed as a general purpose self driving car.
$500 worth of potable water/copper scrap/cotton cloth/wheat is still $500 worth of something with known and stable demand.
Paying the bills with crypto is a lot more convenient and secure than paying with pieces of copper tubing. The convenience and security is what creates value and demand.
in 1990 it was 122 deg F in Phoenix.
Which is pretty much a confirmation of the article.
I guess we're in the mania phase.
Hearing the amount of negativity, I guess we're still in the early adopter phase. Sure, there's a certain amount of mania, but still very limited as a percentage of the population.
I'll settle for rising asymptotically to a stable level. And plenty of people are already using it as a currency. The network is doing about 3-4 transactions per second right now. Compare to Western Union which does 31 transactions per second, it's not a huge gap.
we'll be doing something differently by then
Unless we'll be actively removing CO2 from the atmosphere, it will still be there in 80 years.
To send one ton to Alpha Centauri in 40 years you need a 3.6 km sail and a 65 GW laser.
Plus a huge lens.
If the probe is small, the sail can also be small. Some proposals are for a probe the size of a pack of cigarettes, or even a postage stamp.
How many useful instruments, including a transmitter capable of reaching us, can you pack in a postage stamp ? And how are you going to do a close flyby of the earth-like planet we would be interested in ?
We announced the TPU last year and recently followed up with a detailed study of its performance and architecture. In short, we found that the TPU delivered 15–30X higher performance and 30–80X higher performance-per-watt than contemporary CPUs and GPUs.
I think that counts as a bit more than "gradual" speed improvement. And that was just the first generation. They've already made significant improvements in the second generation (45 TFLOPS/chip)
And the 3rd reason is quickly improving algorithms. Also, the GPUs are being replaced by TPUs, dedicated for neural processing at much higher speeds and lower power.
Depends on what you call "fundamental". There are plenty of new ideas tried out. The number of research papers written on AI in the last couple of years eclipses all the papers that were written before that.