Future models
As the internet begins to fill up with crap generated by current AI systems, how difficult will it be to curate good quality training data for future AI systems?
Crap in, crap out.
176 publicly visible posts • joined 7 Aug 2007
"Don't forget to email corrections@theregister.com if you spot anything wrong."
I tried that a few months ago and the article wasn't corrected. This one was funny, that is all.
It looks to me like you're using speech recognition to write? I can't see how that's a keyboard typing error.
Err, no.
The change is referring to the orbital period of the Didymos / Dimorphos system. They both orbit a common centre of mass with the same orbital period, which just reduced by 32 minutes.
The orbital velocity of Didymos (relative to their common centre of mass) will have changed by 1/104 of that of the change of orbital velocity of Dimorphos, because it's more massive and closer to the system's common centre of mass.
I'd be interested to know what the change of the orbit of the Didymos / Dimorphos system around the Sun is, but that wasn't the objective of the mission and will take a much longer observation time.
According to wikipedia, the two bodies were tidally locked. I imagine that lock has been broken (somewhat), and the eccentricity of their orbit also changed. This is surely worth investigating up close for an extended period of time, to learn something about how binary asteroid systems, and contact binaries, come to be.
"This means a future of abundance,"
Abundance of what? Jobs?
"implying that he hopes Tesla's robots might one day replace humans on production lines"
Oh, maybe not.
"A future where there is no poverty"
How will that work? Will the robots be dumped in a hole in the Congo to mine minerals for their own construction?
"make sure DART is on the right path and counting down the hours before it can attack"
Attack? Are there jewish space lasers on board?
"And then an hour prior to impact for the first time you'll see just one pixel in the field of view of our camera, or DRACO, and that's going to be Dimorphism. And so at that point, this is when we switch from guiding towards Didymos towards guiding to Dimorphos."
One hour seems rather last minute to correct course to target when going over 14,000mph. That's barely enough time for me to swerve to avoid pot-holes when driving.
Build large scale solar powered electrolysis plants to produce hydrogen gas to vent into the atmosphere. It will rise to the top of the atmosphere and provide drag on all orbiting objects on its journey to space. It'll bring all stuff down quicker, and maybe reduce sea levels to boot.
May need a big, well enforced no-smoking zone :)
I've never heard of cameras being used for this. Sounds unreliable, but I'm sure someone has patted themselves on the back for using high-tech.
"The new ones are much more focused so they can ignore vehicles making right turns."
You must be in different country to me.
Where I am nearly all light controlled junctions have induction loops in the ground, and a box with an opaque window pointed at each vehicle approach, and also at the pedestrian crossings if present.
Then somebody programs the controller badly and causes unnecessary traffic problems for the next several decades.
I believe the boxes here use RADAR because they are accurately speed sensitive, some pedestrian crossings are programmed to stop traffic if you drive toward them over the speed limit.
"This sort of thing is supposed to be address in V2X in a more cross industry wide basis. I would expect that the tech would eventually be mandatory in all phones / vehicles as an OS feature, not a 3rd party app. I think its based on broadcasting position and direction vectors."
So we can expect in the future for it to be mandatory to carry a beacon in public that broadcasts our position at all times, for our own safety, to protect us from self driving cars. I'm sure that won't be used for any other purpose.
"The camera based detectors which are on 95% of our intersections only pickup cyclists 25% of the time. "
They are not camera based, they are RADAR. If you wear a silly cycling hat then try sticking a metal corner-cube reflector on top, that ought to help.
It used to be possible to purchase shoes with steel toecaps in them, stopping and placing your foot on the corner of a traffic loop in the road would make your presence known to the lights. Modern safety footware no longer uses metal :( . Maybe your bicycle wheels are still made of metal?
"It will take less energy to heat CO2 to the same speeds as water, which makes for a more efficient system, which is the whole point of the article"
I thought the point of the article was that it was the lack of phase changes in the working fluid made it more efficient. To quote the article, "The Rankine cycle is inefficient and loses a lot of energy turning steam back into water", but I think it's a bit more complicated than that really.
And the available working temperature range is a bit better too, for future designs.
"are these REAL scientists?"
Probably. They're trying to translate it to what they think an ordinary american can understand, I see this sort of nonsense elsewhere.
I've never understood why it's degrees anything. Was Mr. Fahrenheit's thermometer a dial type using a Helix protractor re-purposed for the scale? Can we convert it to radians?
The grid connection is a publicity stunt.
The generator and its control systems are separate from the thermal cycle and the turbine, the two are connected by a rotating shaft running at a nearly constant speed. It's the CO2 based cycle they're trying to prove here, the generator part at grid scale was proven decades ago.
I was curious as to where the heat came from for this test, reading the press release confirmed my suspicion that it was an electrical heater, no doubt connected to the same grid.
I think more test rigs should recover energy somehow to use (if only to power the office lights!), but 10KW is pretty small beer.
"Plastic removed by the group is sent for recycling"
That irresponsible, I'll end up as trash in the environment again. It should be destroyed by incineration in a properly constructed and run plant to generate energy.
"Thus, if we repeat this 100,000 kg haul 1,000 times – the Great Pacific Garbage Patch will be gone,"
Utter bollocks. It's density will reduce, so the total will diminish approximately exponentially.
"A long chain of floating buoys suspended between two boats is dragged slowly through the water, collecting trash and funneling it into a collection area at the rear. A 3-meter deep net wall runs the length of the system and collects plastics floating below the surface"
How much sea-life did it collect?
"And with more chipmakers bringing 176-layer QLC SSDs to market"
How reliable is this cost-cutting technology? Out of 8 ssd's I own, 2 are TLC, and both return corrupted data, which is worse than outright failure.
(1*samsung,1*WD)
Can I buy some more MLC drives anywhere? Or do I have to buy a large pool of QLC turkeys and run 2* ZFS-Z3 to get reliable data storage?
Question about layers that I couldn't find the answer to: How many layers per cell?
I often think some better write-ups would help. I read this article and don't know if it came from Khujirt crater or Dampier crater, or why they think it seems to have travelled from one to the other. I read about this meteorite a few days ago in an article written by an academic. They made no mention of it originating from multiple craters, but explained that using machine learning they found 94,000,000 craters, eliminated around 80,000, and were left with 19 candidates. Whilst I'm sure someone knows what they're doing, I'm stiff baffled as to what that actually was.
Either way there's a trap here, using such a low precision with a wide dynamic range means that when adding multiple numbers, small ones may not contribute to the total because they are individually too insignificant. If there are many such small numbers then you'll get a very wrong answer.
' Chinese foreign ministry spokesman Zhao Lijian responded by calling the remarks "a smear campaign against China's normal and reasonable outer space endeavors."'
That doesn't look like a denial from the chinese, whatever the reality of Bill Nelson's words.
Not that they could defend the moon from invasion by american missions, anyway.
Sound like they'll expect staff to just stand bolt upright all day, not moving at all. That won't be better for peoples bodies than doing the same twisting and bending motions all day.
What would be better was if staff were able to move in lots of different ways, lifting to different heights and working in different positions.
So they already have "new and emerging UAS that possess the ability to receive directed laser energy" according to the RFI. If that were the case then you'd expect them to specify what wavelength range they're looking for, but they don't. It just asks for a big-ass laser. Of 100kW, which conveniently is the minimum power expected to be usable as a weapon.
I think they'd like to beam energy at other people's drones so as to help them travel quickly to the ground.
"People missing the point of alternative cost in that you'll be spending what, 20 minutes on charging alone every 100-200 miles, then if there is a queue multiply that by number of cars in the queue."
Just as there's currently a small industry reverse engineering ECU's to re-map them for punters, I'm sure people will figure out how to connect a petrol-powered generator in the boot to charge the battery while the car is running...
"The average electric vehicle has 2,000 chips in one car"
I'm sure the semiconductor industry's trend of integrating more functions into chips will continue, helped by the market maturing and volumes increasing which will make specialist chips for EV's worth making. That ought to reduce the chip count in time, but to what I don't know.
If I buy an EV (I have nowhere to charge one) then maybe I could keep some salt and vinegar in the car and eat some when I'm hungry :)
"states that a billion is 1 million million"
The long form of these numbers is logical =>
1 billion is a contraction of bi-million, a million million.
1 trillion is a contraction of tri-million, a million million million.
n-million=10^(6n), simple.
Instead people use the nonsensical billion=10^(3*3), because it makes things sound bigger than they are.
The other thing that's stupid is using the ratio of the circumference to the diameter. Using the ratio of the circumference to the radius makes much more sense mathematically, and would remove oddball factors of 2 from many equations.
Humans are a life form.
The only purpose of life is to exploit and consume as much resources as possible to reproduce and expand population.
All life forms must do this, else they die out.
Humans have technology, and something they call intelligence.
Technology vastly expands humans ability to consume stuff from our environment, enabling humans to expand their population.
When conditions are right, many life forms can rapidly consume resources, expanding their population, until the resources run out. Think of swarms of locusts, rats/mice.
human intelligence and technology has so far mostly prevented resources from running out for them, so they keep on expanding their population and consuming stuff from the environment, because that's what evolution has made them to do. humans are a swarm on the planet, but with little sign of resources running out to limit us.
humans cannot avoid trashing the planet.
Stopping global warming won't stop us trashing the planet. We'll still be over-farming land, digging disgustingly big and dirty mines, destroying nature to make way for our activities, polluting the place.
This damage is mostly cumulative, we consume much faster than nature can replenish.
The only way there won't be a major ecological catastrophe is for humans to limit their population, which they cannot do themselves. Although china tried that for a bit, did that work out well?
"Watson Order, which is being rolled out at a number of drive-thrus at fast food chain McDonald's to automate order taking"
That's a bit of a come-down for technology that got so much hype. Is it going to clumsily phrase everything it says as a question like on jeopardy!?
I just hope I don't end up talking to it when phoning a customer service line.
Yes, but probably not on a Mars mission.
https://news.mit.edu/2022/solar-panels-dust-magnets-0311
It involves moving an electrode over the panel, with the correct voltage applied between that and a transparent electrode covering the front of the panel. It requires relative humidity of 30% or greater, which would be the killer for this.
As a side note, I despair at the quality of much scientific news writing. The system described in the article above has nothing to do with magnets.
"The sheet-speakers supposedly require only 100 milliwatts of power per square meter of speaker area"
Er, how much power it consumes depends on how loud you want it, surely.
What size device are they testing? Was testing done at 100mW input power?
12 inches is quite close to make measurements, I thought 1 metre was standard.
They claim it goes down to 100Hz, they don't quote a figure for how loud, but extrapolating the 6dB/octave between 1KHz and 10KHz down to 100Hz suggests it outputs 46dB, so you'll need quite a lot of it to get any bass.
"sender and the recipient of such a message DO NOT KNOW ANYTHING ABOUT THE ENCRYPTION KEY."
That's not true. The sender needs the key to encrypt the message, and the intended recipient needs the key to decrypt it. This is still true even if the encryption is done on pen and paper and the resulting gobbledygook is typed into the device, which I think is what you're getting at.
Diffie Hellman key exchange allows 2 parties to agree on a common encryption key by only communicating information that doesn't reveal anything about the key.
Well it looks like intel don't understand bitcoin. Making mining machines more efficient will lead to an increase in the hashing capacity of the network, so the difficulty will increase, so more hashes will need to be calculated.
The money stream feeding the miners is provided by the price of bitcoin. If a miner has the choice of computing the same hash rate for a lower cost, or a higher hash rate for the same cost, They're likely to want to increase the hash rate to increase their share of the mined coins. Especially considering that others will be making the same decision.
The energy usage of bitcoin will only reduce if the price collapses, they switch away from proof of work, or maybe if energy prices go up a lot. Energy is far too cheap currently.
"From the blog post"? I expected that to be one of the primary pieces of information in the article I was commenting on. It wouldn't have helped anyway, given that I'm neither a computer programmer or a windows user.
The article mentioned chromium, but I try to avoid toxic shiny things.