Motoring Journalists
"Lord knows what motoring journalists make of it all but the ever-extending deadline..."
In the UK it'll be whatever their VAG/BMW/Merc paymasters tell them to make of it.
Up, up and away-a-ay in my beautiful, my beautiful balloooooon… Bye, then. I'm staying here. Nothing would persuade me to get into one of those hot-air death traps. Off you go, floating among the stars, singing your song while sailing along a silver sky or whatever. I've got work to do. OK, OK, I admit I'm scared of the …
I don't know about the UK, but in France it would rather be "whatever their GM masters tells them to". Round these parts, merkin gas-guzzlers that practically need an oil tanker on-call to run for a full daily commute are considered "clean", while my German-made bike which sips 3 litres per 100 km is grounded on pollution alert days.
Sip-related tea-note: as any civilised tea drinker knows, tea is to be brewed in water heated from 70 to 90 celsius (depending on the tea), but NEVER in boiling water. That kills all the aromas.
"... tech company decides to run an on-road test for its autonomous vehicle, it should enforce a Board of Directors' Walk To Work Week."
Good one!
There are very mixed signals about when self-driving is coming:
- Waymo building 20K Jag iPace into self-driving cars over the next 2-3 years.
- On the other side: A Tesla that can't spot a stationary firetruck on a freeway
I've been known to sail on the Solent, and not being anywhere near there I rent a boat. One of them had an autopilot system that had been disabled apparently on the orders of the insurance company, because too many people were setting the same course, so rather than having the whole sea to play in all the boats were in a narrow corridor and the crew weren't keeping proper watch.
I predict flocking behaviour will be an interesting failure mode for fully autonomous cars...
Unless you are a 'Round the World Soloist' in the middle of a respectable ocean, the general rule is that someone is on an active watch. Normally though when sailing/cruising on autopilot you are doing somewhere between 2 and 6 knots depending on whether you are motoring or sailing and the wind and where it is relative to your course. That gives you, at those speeds time to react to impedimenta even if you are reading a book.
With allegedly self driving cars the rule until they are genuinely autonomous should be that the driver/ commander/watchperson should be as engaged and alert as any other driver with their hands in actual contact with a means of control should one of the aforementioned impedimenta appear suddenly in front of them.
An alternative, and this is radical out of the box thinking here, is to limit their speed and have a man witha red flag walking in front of them.
In 2005 I had the happy task of travelling to the Solent to pick up a new boat. The night before we stayed in Milford on Sea and the landlord recited a recent story about Autopilot...
A crew were returning a large Sunseeker to the factory in Poole for some work; they had programmed the course into the nav system and unwisely used coordinates of buoys to set the route. Having exited the Solent, they went through the gap by Hurst Castle to take the ‘coastal route’ to Poole.
The next waypoint was a large yellow buoy off Milford and presumably they weren’t keeping watch properly as they slammed into this large buoy at Sunseeker speeds, promptly sinking the shiny new and very expensive boat.
(Fortunately it wasn’t far off the beach and the crew swam ashore safely.)
Switching on an autopilot in a plane or boat just means that the system takes some of the routine workload of the person operating the thing. It does not mean that the craft is working automatically, you still have to have an active watch. The same applies to cars, in fact more so because unlike a plane or ship the traffic flow is less regulated. So people who switch their autopilot software on in their Tesla and then think the car is going to drive itself are just 'driving without due care and attention'.
"I predict flocking behaviour will be an interesting failure mode for fully autonomous cars..."
Your prediction is already wrong. That was the fault falure mode of all GPS systems for the first 15 years of their existence.. just like your boat example above. We are WAY past that. Waymo already does realtime traffic sensing and makes offers to divert around jams while using their mapping system. It makes sense that they would put this logic, that they already posses, in cars to drive around jams.
Funny thing is that when they named it Autopilot, I thought that the name was absolutely clear , since on an airline flight, although the plane has an autopilot, there are also _two_ pilots to make sure there's always somebody to take over when the autopilot can't handle it, and everybody knows that there's always a pilot on the plane.
Naive, I know.
but there are many, many people in the USA who see it as Nirvana, the promised land. A good number of them are Tesla Fanbois but that is another issue.
Unless ALL vehicles are equipped with the same level of automation and these vehicles are working together, it is a disaster waiting to happen (in my worthless opinion)
I don't normally agree with the views expressed in the articles by Christian Wolmar but his book
"Driverless Cars: On a Road to Nowhere (Perspectives)" published last January is a good read and does provide an insight into the problems involved on the rock road to driverless utopia.
no matter how good the automation is that vehicles of the future will employ there is always the unpredictability of us Humans and don't forget Animals as well. Given the death recently in the USA, I don't think that driverless is going to be here before 2024 at the earliest.
Given the death recently in the USA, I don't think that driverless is going to be here before 2024.
If it was cause by Waymo or GM vehicles you would be right because they are the leaders. Unfortunately it was done by Uber which already has a shitty record and are known for their dodgy practices. So it likely to just be written off as something done by a company cutting corners trying to catch up to their rivals.
There going to be a autonomous taxi at the end of 2018, that Waymo goal. An in several cities across America by 2020. It likely the UK will also have various driverless schemes up an running by then as well, probably in terms of small buses shifting people from town centers to out of town car parks or train stations. An countries like Singapore, Dubai and Qatar will have will all have something similar being set up. Singapore wants self driving buses because they can't hire enough bus drivers.
I bet on Qatar having self driving cars and buses driving around world cup fans for 2022 world cup. Because they want to show off what a modern country they are. They might even have a few flying autonomous taxis flying around by then as well.
Rules:
1) Obey traffic laws
2) Don't hit people, animals or objects that will damage the car
3) If you must hit people, animals or objects that will damage the car, hit the brakes so that the crash speed is as low as possible
4) Try not to hit small animals and other objects that won't damage the car
Unpredictability of humans really is an overblown argument, at best. The technical challenge is reliably to recognize objects and motion. If they can't do that, we won't have autonomous vehicles, if they can, we will have autonomous vehicles.
Me, I scream like a girl if I even so much as glimpse a CorelDraw logo.
Try using Visio. You'll go running home to the comforts of Mother CorelDraw.
The only thing going for the hot air balloon is that it is, in a very practical sense, the only full autonomous vehicle in actual use around the world.
The balloon fraternity would take exception to your assertion that they're not piloted. Steering them's tough (but, with knowledge of the winds, possible), but going down is dead easy. Up is almost as easy, so long as on has one's Ronson handy.
Outside the civilised world, such as France, tea is made with room-temperature water from the nearest hot water tap fed by a boiler that broke down yesterday. Even then, having poured the water into a receptacle, the French tea-maker will wait an additional minimum period of time – somewhere between 30 minutes and a fortnight – before adding any leaves or a tea bag.
They do it on purpose you know, out of spite, something to do with having lost the Napoleonic wars.
It's also very hard to make proper tea in the USA. Their kettles switch off below the boiling point of water, so it's not hot enough to make tea properly. I once got stuck in the middle of Mojave desert doing some tests, but I'd come prepared and had taken a British kettle with me. That, plus a handy ex-US Army generator that could be set to output 240V, enabled me to sup on a decent brew in amongst the sand, scorpions, rattle snakes, and bemused American colleagues. Though as it was at 3000ft up, the water still wasn't quite hot enough.
Despite being sold a then-future (now-present) of autonomous cars, what we got instead was idle, distracted Uber fatties running down pedestrians. As members of the public, we are all fair game to the experimental whims of tech billionaires. One by one, we will be beta-tested to death to make another disruptor incrementally richer.
Autonomous cars is simply the latest tech band wagon. There's a lot of VCs and shareholders putting a ton of cash into groups working on self-drivers, but there's zero chance they'll get a return. The money men are getting burned by the tech evangelists, and when they realise that it'll become harder for more realistic propositions to get funded.
The thing to invest in is nuclear fusion. If that ever gets going (and so far the signs are good), that's the true killer app of the future. Unfortunately for the VCs, various governments have got the rights to that pretty much sown up (ITER).
Great post but I disagree about fusion reactors. I think that's farther off than truly safe autonomous vehicles, which is much farther off than predicted. I'd suggest pebble bed reactors as a reasonable alternative for the near future, or other such nuclear tech. I do think that money needs to put into fusion, but I believe the payoff on its viability it much farther off. Unless there's been a new breakthrough in fusion technology of which I am not aware and that is entirely possible.
I've been at a complete loss to understand why we have a Gadarene rush by manufacturers, legislators and technical-commentators-who-should-know-better to try to convince us that self-driving cars are just around the corner.
Until I remembered a quote (source forgotten) that described cars as one of the instances of real freedom available to ordinary people. Once, we're all in self-driving pods, that disappears.
See also, "smart" meters.
The potential value of autonomous driving is absolutely humungous.
If a company is successful with a reasonable affordable system it would be massively disruptive to vehicle sales.
The R&D is expensive.
The belt-and-braces systems that use a bunch of lidars are very expensive, which means that their market value is 0.
Add all of that together and you have a problem where a bunch of companies have to put a lot of money into autonomy but they aren't going to see any return until they deliver a usable product.
So, it's not surprising that they're all saying "Soon, soon".
>A growing number of seniors who will be too old to drive....
That's all of us, more or less. You don't become 'too old to drive' overnight, the process is gradual and you may not notice it until one day you end up stopped in traffic pointing the wrong way without any idea of how you got there. (That happened to a friend of my wife's.) I don't need a fully autonomous vehicle but anything that can reduce the risk of a crash due to a moment's inattention is worth having.
Inattention applies to other road users, though. That recent fatality may have been caused by a flaw in the software but it was almost certainly precipitated by someone crossing the road at night without paying attention. Unless we can figure out how to write software that automates precognition this type of accident is inevitable (....and for evidence to back me up on this look at the number of people who are flattened by trains in the US -- these things are huge, they have bright lights on the front and a horn that not only can be heard miles away but also is used frequently.)
"these things are huge, they have bright lights on the front and a horn that not only can be heard miles away but also is used frequently"
They're also deceptive in terms of depth perception. They move faster than they look and (since they're usually pulling tremendous weight) have tons of inertia meaning stopping is a matter of yards if not miles. Many times, the train incident is a case of misjudging closing distance.
7 brain dead responses.
THE MONEY. a driverless truck can drive 24/7, requires no breaks, no sleep, no logs. There are 3.5 million professional truck drivers in the U.S. Walmart, king of the cheap shits, pay drivers $73000 a year.
The DOT says that the average driver makes $40,000. Simply accepting the lower number: In JUST TRUCKS ALONE, thats about $140 BILLION per year. Has this question become ridiculously stupid to answer yet?
There are many reasons... traffic remediation, safety, etc, , but the #1 is ALWAYS money.
"In headlines."
I wondered if you'd slipped that one in deliberately to see if anyone noticed.
Here's one to make your toes curl: living in High Wycombe we used to get balloons floating over from a site used by a ballooning club somewhere further up the Hughenden valley. One day I looked up to see a balloonist who had dispensed with the basket. The pilot was sitting on a piece of board - essentially a swing seat. Nothing else. Just a swing seat.
I've flown one of those - including over the centre of Reading once at 1500 ft - and it was surprisingly safe and easy to fly - I was a bit lonely up there.
Bigger balloons with baskets took longer to deflate (on landing) and it was harder to pull the vent at the top, so they would drag along the ground for longer in high winds - not that we flew in what most people would count as high winds.
I never felt unsafe flying balloons, the basket edge is quite high. I have felt more uneasy on high ground based objects e.g. buildings.
It's an enjoyable (IMHO) but daft form of transport, though not quite as daft as a gas balloon where you have a lot less control of height.
Once pedestrians learn that they can cross a busy road by walking in front of driverless vehicles, forcing them to emergency stop ... at least it will redress the balance between cars and pedestrians, very much in the pedestrians' favour!
Rich and poor, as well, if they rob the passengers, or if they don't have anything worth stealing, hold them for ransom. While wearing masks of course - those cars have 20 cameras.
Perhaps the algorithms will be tweaked to run over pedestrians wearing masks.
This, however, may not go over well with certain minorities.
"Once pedestrians learn that they can cross a busy road by walking in front of driverless vehicles, forcing them to emergency stop ... at least it will redress the balance between cars and pedestrians, very much in the pedestrians' favour!"
Evidently you haven't read the news recently.
I feel I have to point out that autonomous cars are "on the roads" in 2018. I'm quite freaked at how accurate that prediction was.
True, they're not on many roads, but they're out there.
I don't know if you noticed, but last week Waymo announced they were buying 20,000 of the things from Jaguar (link). That's not a test run, that's an operational fleet - that will be on the road by 2020.
That seems pretty close to me.
Round where I live the wetware drivers do not have such a great record at avoiding road furniture other wetware drivers, pedestrians flora and fauna either. Either roundabouts are invisible to them or they earn special points for flattening anything and everything on them. Some also treat speed limits as a sort of minimum target to be beaten at all costs to other drivers.
Having said all of that I to suspect that we are some years away from a really autonomous vehicle, though a 'are you sure mode' of guidance might be a step forward.
As for the UBER disaster, as it is subject to investigation and probably legal action I should be guarded in what I say, but I find it troubling that neither the test pilot nor the vehicle registered the presence of an object in its vicinity. The alarms on my newish car go off only too readily and they are only warning the wetware in the driving position.
All I know is that Waymo is launching a self driving taxi at the end of the year
I guess the first three words of that sentence are correct at least, since you've regurgitated nothing but this "fact" 3 times in these comments already.
Look, just because Waymo says it's launching a self-driving taxi this year doesn't mean that it will, nor that it'll be autonomous and not require a safety driver, nor that it'll be able to operate on more than a limited subset of "good" roads, nor that it'll turn out to be a roaring success rather than a glimpse of just how far true autonomous cars have to go before they're a practical reality.
If they weren't confident they wouldn't have pledge to buy 20,000 ipace cars would they?. Already there cars only have a engineer or backup driver in the back seat as part of their pilot program, so far zero major or even minor accidents, have been reported and the press is all over even minor accidents that involve autonomous cars.
An the reason I repeated it three times is because I reply to three comments, not everyone reads the whole comment section.
Only they haven't, have they? They've announced (at a car show) a long-term collaboration, the first car to be in use by the end of this year for testing, first public passenger use by 2020, and an eventual order for "up to" 20,000 vehicles. No $billions put on the table to back it up.
It's the good roads caveat that's going to be a big issue where I am: what with construction, bizarre AD 17NN intersections, and a not-quite-one-to-one correspondence of road to Google Maps, these cars will have meltdowns on a regular basis. Phoenix, Los Angeles, even New York maybe, but not New England.
… when you're greeted by a cracker like "SOMETHING WHICKER THIS WAY COMES". Top notch, have one of these --->
As for self driving cars, the hype is reaching Everest-sized bullshit mountain proportions. I remain unconvinced about any claims of true Level 4 (let alone Level 5) autonomy on real public roads within the next 5 years. So many edge cases and unexpected events that could cause carnage.
Thank you for this excellent article.
Driver less cars only exist in wide open Californian roads, all the rest is click bait generated by car companies whose current line up of motor cars is not worth the fees for normal advertising.
Also lefties love driver less cars, since their objective in life is to restrain humanity as much they possibly can. Taking away the steering wheel is just another item on their jailing list.
Thanks for pointing that out AC. Have an upvote, @naive. You seem like a good old curmudgeonly Register poster. Keep up the good work.
@Steve Davies 3 - I know lefties who oppose driverless cars, and conservatives who don't. There's a word for them: outliers.
Let's see them work their around Metro Manila during afternoon rush. Put it this way, it's SO crowded that most roads become parking lots, SO unruly that it's considered necessary to ghost drive to get anywhere, and SO overpopulated that not even the commuter rail way provides enough relief (stations routinely have lines going all the way back to the street).
@Charles 9
I'm surprised people can navigate metro Manila in cars, soo much fun, the only place I've been where the horn is used for "I'm here please move" and the junctions off the expressway are just an ever expanding mass of vehicles. Prior to experiencing that I drove in Turkey which when you stop at traffic lights it turns into an episode of wacky races. There is no way they are using self driving cars in either of those places.
I’ve never been to Manila. I have been to Cuidad Mexico. There is no way, no way at all, that Manila can be worse than Mexico City. In particular, there is no way that driverless cars can survive in Mexico, not with the way that the jaguar knights driving those VW taxis drive. Taxis in Mexico City are the fastest things on four wheels, despite the awesome traffice, because Mexico City taxi drivers aim for where there had better be a hole when their made-in-Mexico German-designed terror device arrives. And that VW will be trailing Cherenkov radiation... okay, I exaggerate, it’s only a sonic boom. I understand that they’re retiring the VWs. No doubt they’d be replaced by F1 racecars, except that those would be too slow. Anyone attempting to operate a driverless vehicle in Mexico City would find themselves in serious trouble, shortly before being rammed by one of the 140,000 taxis. Probably by more than one.
Mexico sounds fun, it's on my bucket list. In manilla you don't only have to content with cars, there's the carts of doom (as I call them) motorised tricycle carrying 400 people (exaggeration but sometimes you do count to be sure), jeepnees (old american army trucks converted to buses), taxis converted to run on gas with the container mounted on the boot (they will also happily stop off to fill up while working and smoke at the same time). It's that rammed depending on your licence plate number you can't drive on certain days.
Some fun facts.
https://faq.ph/facts-about-traffic-in-metro-manila-that-you-should-know/
I don't think self driving cars are going to the 2nd/3rd world anytime soon.
I wonder how they will fare on Montreal (Quebec) roads.
Our roads are so full of potholes that only drunks drive in a straight line.
If you manage to survive winter/spring without a pothole induced tire blowout, you are lucky. Last winter I hammered out two steel rims on my cars.
Will driverless cars avoid potholes or drive straight through?
@David 164
Nope they are being tested on London roads, https://www.youtube.com/watch?v=cfRqNAhAe6c
In case you missed it, that video shows a car which had to have a driver who was fully engaged in watching the road, and twice in the course of the video (10 minutes?) he had to take manual control.
Not very autonomous, is it?
As members of the public, we are all fair game to the experimental whims of tech billionaires. One by one, we will be beta-tested to death to make another disruptor incrementally richer.
I was under the impression that Dabbsy's column was supposed to be satire, not on-the-nail factual reporting. The only thing missing was "and the powerful another step closer to omniscient and as hard to shift as a floating turd."
"...horrors the Brits perpetrate when making/serving coffee. ;-)"
Coffee bag in a mug. Add boiling water. Risk blistered fingers by dunking it about a bit until it's cooled enough to drink. Drink slowly, leaving bag in mug to achieve a turkish level of strength by the time you get to the end.
What's wrong with that?
"What's wrong with that?"
Starting with what the Brits consider "suitable" for putting into those bags? *shudder*
(Disclosure: I have worked for a dutch coffee company which produced on order for the UK (beans, pre-ground and instant), and still have a fair feel for the blends. The UK blends are ...different.. from the rest of Civilised Europe. And damned if you tried to show them there's BETTER stuff for the same price range...)
Those ARE ashes... Reminds me of my honeymoon (2009) in Vegas and we decided to splurge on a $170 bottle of Real Deal (tm) French Champagne... It was fucking disgusting, you could almost taste Pierre's feet*... ok you could smell Pierre's feet too. 0/10 would not recommend.
* went off on a tangent concerning Pierre, the most popular and sexy grape stomper in his village. Women from miles around come to town to admire Pierre's callused hairy moneymakers... They swoon when he flexes his toes...
...The only thing going for the hot air balloon is that it is, in a very practical sense, the only full autonomous vehicle in actual use around the world....
For a technology journalist you're not very good with technology.
A balloon - hot-air or gas, is steered by the pilot. He steers by raising or lowering the balloon, so that it passes through different air currents moving in different directions. Easy to google if you didn't know it already. https://science.howstuffworks.com/transport/flight/modern/hot-air-balloon2.htm
A horse is a good example of an autonomous vehicle - I think someone mentioned this below....
"He steers by raising or lowering the balloon, so that it passes through different air currents moving in different directions."
It still needs an air current going in the direction the pilot wants to go, or at least sufficient variety of them to sum to that. If there aren't any the pilot can steer but not where he wants to go.
Scientists are discussing the decrease in strength of the magnetic field and a possible flip in its poles, which, I think, takes a very long time. Will that have much affect on computers. Then there's the potential for an EMP or the sun throwing out an massive solar flare at a point when the earth's magnetic field is very low how will that impact computers? Let's face it that's not just cars being affected. Water pumping stations, power plants, farm equipment, etc etc
I, really, don't like to be a Luddite but as technology progresses I can't help but believe that what should be prerequisite safety precautions are not being taken because profits.
>I, really, don't like to be a Luddite but as technology progresses
One day the Earth will be consumed by the Sun as it expands into a red giant and by then I'll be long gone and so will you.
Knock off the weed and don't think too deeply about things that haven't happened, just go and enjoy the short time you have inhabiting this planet.
Thanks, and I do :)
Thought I'd point out a potential issue that is largely avoidable. I work in the computer industry but still think that there is an over reliance on computers to run infrastructure. The problem as I see it isn't that computers run infrastructure but that the computers that do aren't sufficiently hardened against potential threats either physical or existential :)
They will learn as these cars takes nice picture of their number plates and even of the individual faces and send it straight to the police for them to be fined!
Already Essex police are fining people base on dash cam footage sent in by the public, I suspect they are rubbing their hands at the thought of all those automated cars sending footage of their human rivals breaking the law, their fine revenue will go up ten fold, at least until humans learn to leave autonomous cars alone.
>Already Essex police are fining people base on dash cam footage sent in by the public
Perhaps, but what is the correct procedure when someone pulls in front of you and you no longer have a "safe distance" between you and the car in front. Ah, yes, the legally correct thing to do is to pull back until you have a safe (braking) distance.
Now who do you think will end up at the back of every queue?
What are some other rules... keep left unless over-taking? I'm sure that will go well for autonomous cars, stuck between lorries.
I fail to see the problem being solved with this tech. We'll still have the same number of cars on the road at rush-hour. It would be far cheaper and have a far more certain outcome to buy a few more trains. Maybe electrify city centre roads so cars could run without fume and without expensive batteries. Maybe limit autonomous driving ambitions to auto-braking ad-hoc road-trains on the motorway - essentially using sensors to remove the "reaction time" component to squeeze cars close together. Urban and suburban uses? Probably not for a very long time (tech-wise).
"We'll still have the same number of cars on the road at rush-hour."
And you'll ALWAYS have the same number of cars on the road at rush hour because no other transportation tech can operate door-to-door for most practical distances. Unless your alternative can match the convenience of being able to cover both the first and last mile, your best bet is to manage the crush as best as possible, and one way is to remove control of the vehicle from capricious, error-prone humans.
"Unless your alternative can match the convenience of being able to cover both the first and last mile, your best bet is to manage the crush as best as possible, and one way is to remove control of the vehicle from capricious, error-prone humans."
Another way is to stop insisting that most of the population has to turn up at work at 9AM, in person.
As of now, it's rather hard to erect a building, sell a six-pack, or comfort the sick by remote control, last I checked. And it's hard to keep schools scheduled anything other than rigid blocks due to the need to organize their transportation.
Put another way, schedules are the least-worst solution for keeping an interlocked civilization from crushing itself into chaos.
That's an understatement. It took me two seconds to find this: https://www.youtube.com/watch?v=GJrncrveDLc - and those are only the scariest of the last year or so.
It's amazing no one wants to ban them.
If you are in search of civilised transportation, Alistair, may I recommend something along the lines of the Windermere Steam Launch "Dolly" or similar.
These craft are often fitted with a Windermere Kettle in which live steam from the boiler is used to generate properly boiling water in seconds.
"possibly the most expensive kettle ever to roll off a production line."
Quoth Dabbsy:
My immediate conclusion is that Koreans are as bad as the French at making tea. As anyone in the civilised world knows, tea must be prepared using boiling water
As anyone in the civilized world knows, tea must be prepared using water approximately 25 degrees below boiling. One does not boil civilized tea. (I.e., green tea. Thus saith the civilizations that invented the stuff, including the Koreans.)
Only poseurs pretend wilted oxidized tea is better.
25 degrees below boiling from which scale, C or F?
I can tell you that 75 Celsius is going to give you a cup of warm off colour water, I drink my tea at almost 60 degrees according to my laser thermometer. Standard 'British' black tea needs boiling water, all the other stuff conforms to Alf Garnett's foreign muck or poncy like Earl Grey.
Coffee on the other hand is destroyed by boiling water and will likely be come bitter, 97C is apparently the max for coffee but it depends on the bean, the blend, the roast and how it is ground. Strangely, there are courses you can go on in the UK for coffee making but I have never heard of a course for tea making.
Presumably because tea making should be inherent in British genes.
From what I can read, it all depends on the tea. Basically, the more oxidized the tea, the hotter you need the water to be to make it ideal. From what I've read, you need some boiling before you pour, but not too much or you lose too much oxygen in the water; plus you want the water to stop boiling once you pour so that it hits the sweet spot (for unoxidized teas like green and white, that's about 80 C, semi-oxidized like oolong is best at 90 C, and black teas which are fully oxidized should be left just below the boiling point of 100 C). The best answer basically is, "It's Complicated. Your mileage may vary."
PS. Might also have to take into consideration the condition of the water being used, as hard water is notorious for not getting better with heat.
... ah, the working title that David Niven rejected for his memoirs...
I was in Toulouse for a couple of months last year and have to say that Carrefour was one of the very best 'Supermarkets' I have ever used.
From High-end produce at one end - including a sushi-bar - to lower end, 'ordinary' food and a huge household items area, it was an amazing place to see and use.
And they had a decent choice of fresh milk, probably solely for the British visitors - my Spanish 'in-laws' still use UHT despite having a fridge and a supermarket that sells fresh. I guess it is what you get used to.
Personally, can't stand UHT and I have the merest dash of milk in tea these days.
"Made me google UltraHighTemperature milk... That stuff, even by written word, is pretty nasty. Nope nope nope."
Well, in hot environments where refrigeration isn't a given, the only alternative to UHT milk is often powdered milk. Frankly, forced into a choice between the two, I'll take the UHT milk, though I think there will be some who see that and reply, "Stop the world, I wanna get off."
"Well, in hot environments where refrigeration isn't a given, the only alternative to UHT milk is often powdered milk."
I lived in tropical Darwin for ten years, where most food had to be trucked in from other places. You get used to UHT milk, and I'm usually very fussy about my milk.
I've lived in Guam and the Philippines, both tropical and bereft of dairy cows, meaning if you want real full cream milk, you're going to need it shipped in from the likes of Australia. Heck, I recall that most of the milk in Guam until the late 80's (there was only one dairy company there: Foremost) was filled (skim milk ships better, and they just used substitute fats once there). Only later did they figure out a way to safely transport and recombine the milk fats, allowing for full cream milk to be sold regularly. And even today, most milk advertised in the Philippines is powdered for convenience and shelf life (UHT keeps but is pricey and bulky; they do sell evaporated and condensed milks as well, but these tend to be used for culinary purposes instead--too expensive to reconstitute for drinking).
Finally, a tech writer who isn't a moron.
Self-driving cars don't care who they kill or main. The humans who produce self-driving cars don't care who their cars kill or maim. Computers cannot think. Computers can't anticipate in the same way that humans can. We are going to witness a fiasco caused by hubris and greed.
Oh, yeah, hydrogen cars.
See the kit involved in transporting slightly more than 200kg of transport fuel hydrogen to a refueling station, and the mayhem resulting when ineradicable human nature encounters a fuel that is even trickier and less forgiving than gasoline:
https://www.ntsb.gov/investigations/AccidentReports/Pages/HMD18FR001-preliminary.aspx
The sort of autonomy that people are believing is just around the corner is bollocks. Google's cars operate on a hyper-accurately surveyed set of streets. Other cars can operate on well maintained and marked motorways. Their aren't cars you can hop into and punch in any destination at random (within reason) and the car can take you there. For full level 5 autonomy, there will likely be the need for a revamp of roadway infrastructure. GPS/GLONASS/SatNav isn't accurate enough to guide a car since it's best commercial setting is +- 3 meters. With differential GPS it's possible to narrow that down to 2cm, but that takes permanently installed base stations to work. Still, that's not entirely impossible for larger cities. Smaller cities and villages would have to make too large of an investment to make it worthwhile and there is still the probability of system failure.
Real autonomy may be available for major city centers in another 5-10 years. It might be an option for consumer purchased cars but pricey and not useful except in those cities or along major motorways.
I agree that autonomous cars will be taken advantage of by peds, cyclists and kids playing games. Connected vehicles are wide open to remote attacks that could strangle traffic in a region.
"GPS/GLONASS/SatNav isn't accurate enough to guide a car since it's best commercial setting is +- 3 meters. "
That's not really a problem because the cars own sensors will be looking at the fine detail of actually driving correctly. My relatively old satnav with updated current maps is very good at knowing which road I'm on and which direction I'm heading but I don't and can't rely on the satnav to keep my position on the road. I have to do the fine control. But as you say, it's not going to happen anyway for a long, long time yet, so I'm agreeing with you :-)
JBnb noted, "My relatively old satnav with updated current maps is very good at knowing which road I'm on.."
That's a software trick. Your old GPS software assumes that you're on the road, and it can use turns (and other clues) to improve the position solution. All because you're staying on the road. That software trick doesn't extend to having GPS taking control of the steering wheel, as you've correctly noted.
GPS doesn't work for lane keeping. I hope that most people know that by now.
Do you think it needs GPS to drive itself? No it doesn't, otherwise the car would stop in tunnels. Also there are always road work, the car does not rely on a pre-existing path only.
That is why you have a huge amount of instruments on the car. Just check a video on how Waymo maps the road in real time in order to drive.
"That is why you have a huge amount of instruments on the car. Just check a video on how Waymo maps the road in real time in order to drive."
So how does Waymo map an UNMARKED road (or one that's covered by a whiteout)? Or one that's incorrectly marked due to things like construction? Or ghost drivers? Or cycles weaving between cars and so on? These are REAL real-world conditions I'm talking about here.
"So how does Waymo map an UNMARKED road (or one that's covered by a whiteout)"
I don't think there has been snow in San Francisco that I can recall. Waymo maps roads by using a car that is outfitted with a mad number of different sensors and they make many passes in every lane and in both directions to create the maps the cars use. Sensing other cars, peds, cyclists, etc is a different system that handles variable targets, errrrr, obstructions, but the internal map is keeping track of the car's position with a high degree of accuracy. I'm sure they also have GPS in their system, but it would be a low valued first approximation of location. Waymo isn't allowing the cars to drive themselves on roads that they haven't previously surveyed.
I don't do the software, just the hardware for aerospace stuff. Navigation is a very tough nut and I highly respect my colleagues that can write that code.
Which is why I'm concerned about things in a more-chaotic setting where even the direction can change from day to day. Roads under constant construction where the markings are missing...or even wrong. Movable barriers, improvised paths, and so on. Sometimes, GPS may be the ONLY source of location you have because none of the other landmarks are present, and imagine a concrete canyon where almost NO radio signals work.
That's one reason I mentioned the Far East, because traffic discipline in a less-civilized Far East country is almost nonexistent. Same for uniform road standards. As a colleague of mine put it, if you can negotiate a trip through Metropolitan Manila in rush hour without incident, you can navigate through nigh anything.
Technology thats come into fashion over the last few years are things called CORS.. Continually operating Reference Stations, which are GPS base stations .. continually operating, and sending their correction messages out via phone and internet. Its subscription based (that I am aware of), a bit slower and less precise than your own dedicated base station, but saves you from having to set up your base in urban areas where anything shiny and unattended has a lifespan under an hour.
These CORS networks are becoming more prevalent though still not everywhere, the biggest problem with satellite based driving would be obstructed views of the sky.
"These CORS networks are becoming more prevalent though still not everywhere, the biggest problem with satellite based driving would be obstructed views of the sky."
Yet another monthly subscription to make a product work would just make me scream. If there were very accurate maps that the car was accessing for it's entire route, when the GPS signal was not available, an IMU would be sufficient for a short period along with RADAR, Lidar, etc. I've done that for rocket landers with award winning results over a span of 150m and 3 minutes. Since 3 minutes was the max safe flight time for the amount of fuel it could carry, that's all we could really test for.
Don't ride in a self-driving car that is powered by deep learning. A deep neural network is essentially a rule-based expert system. They all have a fundamental flaw: they fail catastrophically if they encounter a situation that they have never seen before. For example, it may not recognize a pedestrian in a Chewbacca suit because it has never been trained to recognize one. The human brain does not have to be trained for every possible situation, an impossible task. We have common sense. This is why we are still orders of magnitude safer drivers than autonomous cars. The California DMV disengagement data on self-driving cars proves this.
Transportation agencies everywhere should ban all self-driving cars on public streets until they can prove they have common sense.
"or you just program them to avoid all objects in the road that are bipedal."
That's a good start. Now what about women pushing bikes across the road? Or pushchairs/prams with babies in them? Animals? How do we teach it to avoid a bottle dropped or blown into the road? Most humans would likely spot if it was plastic and not worry too much, or glass and try to avoid it. When will it be good enough to distinguish between potential hazards and large numbers of autumn leaves either blowing in the wind or getting caught in a small (or large) moving pile in a wind vortex? As I posted yesterday, this is really really hard stuff to program for. I'd be very interested to see each of the leading AV makers test their cars on a windy Autumn day down a tree lined avenue. Not to mention in snow where all the lane markings are covered over.
If the issues are solved, why are Teslas still crashing when they should be employing ten-year-old tech and be smart enough to avoid obstacles a human eye can and would be able to detect well ahead of time? Can this ten-year-old tech be able to tell between stuff it can drive through and stuff it should avoid...even when they look the same (like an old woman versus a carjacker)?
>> When will it be good enough to distinguish between potential hazards and large numbers of autumn leaves either blowing in the wind or getting caught in a small (or large) moving pile in a wind vortex?
A couple of months ago, I was following a self-driving car down a street when it swerved to avoid a low pile of leaves on the roadway, low enough that I had no problem with my decision to simply drive over it.
David 164 suggested, "...or you just program them to avoid all objects in the road that are bipedal."
Thereby running into four-legged cattle. Or two little kids holding hands and skipping across the road.
Volvo's boffins had to visit Australia because they failed to anticipate the existence of kangaroos. They seem to be thick; the boffins.
The entire approach that you've suggested is a recipe for disaster.
...when it decided that crashing straight into a concrete barrier was a perfectly reasonable course of action.
http://www.bbc.co.uk/news/world-us-canada-43604440
Apparently the basic Accident Avoidance systems are not working in these circumstances.
AI boffins and their managers are clearly exhibiting a Confidence/Competence Ratio that dangerously exceeds unity.
Autonomous Vehicle fanboys had better mentally and emotionally prepare themselves for the Great Disappointment of 2018/2019.
There are a bunch of videos on YouTube by people in Teslas showing that their cars aim straight for that barrier too. There is also a video showing the aftermath of the accident before emergency services got there and close the road. It's a very frightening mess. It has to be questioned why the software steers the car towards a stationary object at highway speeds. Does it not see it? Even if it thinks it's something other than a barrier, it should detect that it isn't moving.
I'm new to this website and I don't know if the author is joking or is he really serious?
Autonomous Vehicles are already much safer than humans (look at the leaders such as Waymo, not the worst like Tesla and Uber), if you look at the impressive stats you will see that it is the future. You did not present any argument on your side except personal feelings.
Also insulting French as not being civilized for not knowing how to make tea is like insulting the Americans and British for not knowing how to make coffee or food. Every country their speciality.
Until 100 years or so ago most vehicles were semi-autonomous, they had a quasi-intelligent propulsion system known as a 'horse'. Horses were able to do simple obstacle avoidance, they could follow a track and follow other vehicles in a procession with minimal input from their driver. They still needed a driver, though, because not only were they incapable of dealing with all situations (including complex navigation problems) but they also would occasionally suffer from bugs that caused them to bolt uncontrollably.
Our computer systems still aren't quite up to the intelligence of a horse so I don't know why people think they're capable of managing a vehicle unaided in all situations.
Too right. My horses never had a problem returning home. I could nap the whole way back if I didn't have to keep them from trotting or galloping all the way. They were very good at playing follow the leader too. If the person in front knew where they were going, the rest of the horses would just get in line.
I agree that the hyper for self-driving cars is totally unfounded, but I don't agree that they're unlikely to become a real thing, given enough time and development. Like the paperless office, it's coming. My office is pretty darn close to paperless, and as the greybeards retire it will creep along until nearly all offices are near-paperless.
>> I agree that autonomous cars will be taken advantage of by peds, cyclists and kids playing games.
I live in Sillycon Valley, where I see numerous "autonomous" cars on the roads every day. But I have yet to see one interact with human beings the way many (not all) drivers do -- make eye contact, make gestures to indicate intent (hopefully not rude), and even cede the right of way.
I find the need for highly mapped roads a key shortcoming of driverless vehicles. It's not the roads that need to be mapped -- it's all the humans, both in vehicles and outside of them that need to be understood.
Additional thought: If teenagers eat dishwasher detergent packs and snort condoms, how soon before they try playing "chicken" with autonomous vehicles?
... many years ago,.... like 1991 ish, I did a contract for a subsidiary of a Swedish paper pulping company. They'd diversified into making windows and doors, because the guy from IBM, not thinking about who he was talking to, dropped the phrase 'paperless office' when he was flogging them a System34 some years earlier. So, they'd commissioned a usage report, and discovered that the 'paperless office' at that point, was a total myth, and paper consumption had in fact been increasing. Computers just meant folks could print off more crap.
3. Paper backups of reports are easy to store and can be read years later without needing to have a very old version of software running on a very old computer using a years past expired operating system.
Something to think about if you are using a subscription based accounting package. What happens if you are audited 5 years down the road and can't access that old data. Do you really want to leave it up to the tax authority to determine whether you are compliant or not?
Don't you generate your reports using a format designed for the purpose such as PDF? I know for a fact 20-year-old PDF files can still be read today, and enough of the format spec is open enough that the capability should remain for as long as the format remain relevant (and given its design, it'll probably remain relevant much as the JPG format remains relevant).
This TED talk is from 2015
"How Google Driverless Cars See the World"
https://www.youtube.com/watch?v=tiwVMrTLUWg
It's 15 minutes, and gets interesting about half way through.
Really makes the point that on busy multi-lane roads and junctions, being aware of, and anticipating the movements of, everyone else is key.
How Many People Have Died From Falling Down Stairs?
https://ellisinjurylaw.com › Blog
Dec 13, 2017 - According to the most recent data on stairway accidents in the United States, an estimated 12,000 people die from falling down steps each year.
OK, so if I was an elevator maker, and I showed statistics that our elevators only kill 6,000 people per year, I could say we were twice as safe as stairs and everyone would be satisfied?
My last hot air balloon trip was over Damascus, but the Syrian civil war hadn't started then. In a way, I'd trust the basket weavers more than someone else's code in an autonomous car. I'm not at all risk-averse (I drink too much, I still smoke and I enjoy safaris in Zimbabwe) but I also still enjoy driving (even a diesel vehicle like my 2007 Land Rover Defender),
I really don't think we're ready for level 5 cars yet.
Just my opinion.
Glad I'm not the only realist out here.
I work on embedded control systems and totally agree that fully autonomous all-road vehicles are currently a myth and likely to remain so until someone comes up with a game-changing take on transducer and AI technology, as I've previously commented about over El Reg articles & New Scientist gushing.
What disturbs me most is that many of my colleagues - most of them reasonalby competent engineers - take this guff seriously. We don't even have autonomous street-level trains. We certainly don't have technology that can navigate an unmapped side street or country road while guaranteeing not to crash into unexpected tree branches, badgers or cyclists.
In the fifties pundits were predicting we'd all be zooming around in personal flying cars by 1980. The current bullsquirt over self-driving cards is no different.