Posts by Pete 2 3497 publicly visible posts • joined 10 Jun 2009
Uber eats, Netflix, remote working, hotel wifi, plus all the other trappings of "modern" tech existence.
It makes a chap wonder what the situation would be like if this virus had hit 10 or 15 years ago.
While many of those things were possible for a few, they would not have been widespread enough (or familiar enough) to scale to a national level.
> Ever taken a call from a customer unable to discern fact from paranoid fiction?
Many decades ago, when the concept of "technical support" was still in its infancy I was called in to see the boss's boss's boss. Me and another individual from the "industrial" (I worked with "commercial" customers) side of the biz.
We got a rollocking because one customer had filled in a survey (I didn't know we did surveys. If I had, I would have "helpfully" filled them out for my customers - just another example of going "above and beyond") and given a big fat ZERO for customer satisfaction. Somehow mentioning me and my colleague by name, hence the carpeting.
When put under the spotlight about this, I rightfully said that I have never heard of this customer. Nor were they one of "ours" - which my colleague agreed with. The B's B's B took all that in, then ignored it with the words "I don't want to know. Go out there and fix it"
From the customer's point of view this all worked out fine. They got the attention they were seeking. They got loads of free support and they got their own lack of technical competency sorted out. They also got my office phone number - although by an error of transcription (I'm sure I said " ... 2, 3" - but they wrote down 3, 2 ) that phone line was terminated at a junction box in a store room.
> they knew it was risky to recycle passwords or light variations on a theme
People reuse passwords on so many sites because it is of no consequence to them if those accounts get compromised.
For example, if you joined, or were forced to join, a website or forum because you ONE TIME wanted some information that was only available to members, it is quite reasonable to use abc123 as a universal password.
The same if you wanted support from a user forum. Join - ask question - get ignored as a noob - leave.
If the account gets hacked and your password is stolen, it's no big deal (you've probably forgotten about it anyway). There is no risk as nothing of value is being risked.
> Two of those will also be making use of United Launch Alliance's (ULA) upcoming Vulcan launch system
Both Blue Origin and Dynetics are critically dependent on the Vulcans. So really the race is between SpaceX and an outfit that have yet to lift anything off the ground.
The maiden flight is planned to take place in July 2021 [ wiki ]
> business process changes to allow for greater social distancing
Unless Amazon can magically increase the size of their warehouses and shipping facilities, increasing the distance between employees while keeping the acreage the same can only be solved by having fewer people in the same space.
> The Terbium team reckons that these guides, ...
> make up just under half (49 per cent) of all data transactions on the store (not including drugs or for-hire services like DDoS attacks)
So in reality, just a tiny fraction - when you exclude all the high profit stuff!
It also makes you wonder what is in these "how to" guides for online fraud.
I can imagine how the advertising goes:
Buy my book on how to commit online fraud. Only ₿1
and inside the book is just the sentence:
Create an advertisement for a book telling people how to commit online fraud.
> a waste product, such as the urine of the personnel who occupy the moon bases
Any Moon base will have to be a closed system. There will be no such thing as a waste product.
As for using materials (and resources) available on the Moon? The other thing there is plenty of is sunlight. How about tightly focusing that to fuse the regolith into solid blocks, or refining out the metals and building the structures with that.
Or do what The Jam suggested: Going Underground?
Huawei ..... oh yes: 5G rollout.
Way back in the beginning of time (i.e. December 2019, before Coronavirus took over every news channel) there was a bit of a kerfuffle about the UK using Chinese kit in its nascent 5G network. Doncha know? they might spy on us. Just imagine!
Three months later and governments all over the world are considering ways to do just that. To track their people and see who they come in close proximity to. It seems to me that this "spying" stuff is ideally suited to this application. And since the Chinese seem to be ahead of the game in mass surveillance, they could probably supply a turnkey system to fully integrate with their 5G systems. To do just what governments now consider a valid reason to "spy" on us all.
> buck arse nekkid while playing the bagpipes
From any distance, playing the bagpipes looks and sounds like a person molesting an animal - whether you are naked or not.
We all know that the British hold anything with 4 legs (but fewer than 6) in much higher regard than their own species. So be prepared for an armed response unit to descend on you for crimes against the animal kingdom.
> the police would have stopped those people in response to the system's result to check id
No. That is not how it works.
The facial recognition system scans thousands of faces.
It flags up a small number of potential matches.
Each of those potential matches is checked by an officer in the control room
If the officer agrees, a call goes out to stop the individual for an identity check.
The only time a call is made is if the person (or persons) who have re-checked who the computer flagged agree that there is a match. 5 times out of the 6 times, the officer (a person) made the same error of positively matching that the computer did.
> The AI tech flagged eight as being possible matches; seven turned out to be false positives, five of whom were actually stopped by the cops and two
> dismissed as obvious errors. The remaining person turned out to be a true positive, and was intercepted by the British plod.
> That's an inaccuracy rate of 87.5 per cent.
The "hits" that the facial recognition system makes are then referred to real, live, people for verification. If the cops get to the point of stopping someone (to ask them for identification, not to arrest them) then it is because an officer has agreed: yes, the face flagged up is actually someone we want to talk to.
At no point did "the computer" arrest anybody.
So a better argument would be that the system flagged 8 people. 6 were passed by police officers, one was genuine and the other 5 were "misses".
So the computer got 7 of 8 wrong, but the officers got 5 of 6 wrong (83%). That shows that the computers are almost as good as the police at identifying wanted individuals. And a damn sight faster - cheaper, too.
Only The Guardian would try to twist such a good (comparable) success rate as meaning facial recognition was a failure.
I am told there is a principle in designing parks. You build the park, see where people want to walk, then build the paths along those lines. If a designer decides that a path will go from A to B, but people prefer to take a different route, they will.
The same applies to security. It is just plain dumb for an infosec person to say "this is what users should do" and then build security around that. A better approach is to see what users actually want and then make those available in a secure manner.
> the purpose of that training set is (was?) to allow a categoriser to identify things
Yes, we agree. And as long as those same things appear in other parts of the training set, it doesn't matter if they are missed out in one or two single frames. The AI will still "learn" them from the other frames where they ARE labelled.
The article neither says nor suggests that entire classes of object (all traffic lights or all prams, for example) are systematically omitted. Missing from the entire set. Just that the occasional thing that is labelled in most frames has not been labelled in a few.
> Thousands of vehicles, hundreds of pedestrians, and dozens of cyclists were not labelled.
It seems to me that there are two sorts of mistake. There is labelling something incorrectly, such as labelling an elephant as a bicycle. And there is a different sort of error which is not labelling something at all.
In the first case, mislabelling obviously leads to errors in the system that is trained on the data. But what about the second?
If something is not labelled, then it doesn't form part of the training. It is as if the unlabelled object doesn't appear in the image. ML systems are not trained to identify every single thing in an image: every leaf on a tree, every piece of litter. So things that are not labelled will simply be ignored: not forming any sort of association.
So the question is: does it matter if some images don't have some things labelled?
> Should the UK find something better on which to spend the BS billions, and make use of the Galileo PRS instead
So instead of shooting dozens of satellites into space (presumably not on Ariane vehicles). wouldn't a better system than another GPS be a comprehensive network of road signs?
Answer: SpaceX.
It seems to me that NASA can only continue to get government funding for these remarkable feats / token gestures, if nobody else beats them to it. Once some other outfit achieves NASA's goal, they have no more reason to continue trying for it. And the politicians have no reason to spend more money on it, either.
Just like happened between NASA and the Soviet space programme.
> gear to test an automated malting procedure to compare malt produced in space with that on Earth for genetic and structural changes.
After all it's long dry trip to Mars. You've gotta know that when you get there, the microbrewery you sent ahead will perform as required.
But what is a good name for a Martian beer?
> Machine learning requires a large number of examples
And luckily many media outlets already have a large number of examples.
I wonder if anyone would notice if all the back editions of The Daily Mail or The Guardian (being two of the more predictable newspapers) was fed into this things learning algo. Then each evening the respective editor could just type in a few keywords and out pops the next day's edition.
It could be that the only difference is the lack of grammatical errors and spelling mistakes.
November 9 was the fifty-second anniversary of the first Saturn V launch.
It would seem that in the USA, state-sponsored rocketry has made no progress in that half century and may even be going backwards. That launch was a success!
In perspective, the 50 years from 1919 to 1969 saw commercial aircraft development go from wooden bi-planes to the 747.
> Smart TVs do not in fact, demand an internet connection before they work.
For a very limited definition of "work" that opinion can be forgiven. But when a "smart" TV is sold on the basis of the additional features that being "smart" provides it is another matter. Apply the same principles to any other gadget that calls itself "smart". A "smart" phone, for example is only as good as the apps it will run.
A case in point. We bought a top o' the range "smart" TV. Much ado was made about the apps it ran and the benefits they gave the lucky user: Youtube, iPlayer, etc. After about 6 months some of those stopped working - simply because whatever was at the other end of that app stopped providing service. Some time later the TV updated its software and more stuff stopped working. Now, 3 years on, it is nothing more useful than a large screen and a remote control - what with there not being any 4K OTA transmissions. And the internal tuner not being capable of receiving them, if the future brings any
> will switch off the "smart" portion of its Insignia-branded smart home gadgets this coming Wednesday, rendering them just plain old dumb gear.
There is a view that the coming recession will see a lot of unicorns becoming extinct. Investors will tire of throwing $$$ billions more on the bonfires and then they will rapidly run out of cash.
And when that happens, many "smart" devices: including "smart" TVs and streaming boxes, will be of very little use.
If we learn anything from that, it will be a reinforcement of the idea that unless you have complete control over your stuff, you really don't own it at all.
> The BE-7, of course, has yet to actually leave the test stand. Bezos told the audience that to date, the company had managed 13 minutes of test time, including a three-minute continuous firing.
There is a big difference between a test firing on Earth, in 1G of gravity and a nice, accommodating environment and hoping that the same engine will even light up after several days of near absolute zero, vacuum and after the stresses and strains of launch.
> they looked at more than 72,000 C++ code snippets in 1,325 Stack Overflow posts and found 69 vulnerable snippets
You are whinging about 70 errors in 70 thousand samples?
If Microsoft managed such a low rate, Windows would have had the invulnerability of Fort Knox in 1989 (even before SO existed).
Hell, if any professional programmer made so few mistakes, software development would almost be considered a respectable way to earn a living.
> And launching from say 30000/10000 ft/m does eliminate some atmospheric drag
It's not the drag so much as the atmospheric pressure. The shape of the nozzle (where the exhaust comes out) plays a large part in the efficiency of the engine. And that shape is determined by the surrounding air pressure. What is best at ground level is different from what work at higher altitudes.
And since most of the travel will be through low-pressure air, higher up, the benefits of a more efflicient nozzle become significant.
In addition the amount of fuel that is burned just to get the rest of the stack up to 30,000 ft. is a saving too.
> NASA's initial OPOC order, costing $2.7bn, is for three Orions as soon as possible to cover Artemis missions III, IV and V. The next three will be ordered in 2022 for Artemis missions VI, VII and VIII at a projected cost of $1.9bn.
So shouldn't the first of these be Artemis IV (a new hope), then V and VI
It is noted that these capsules will have a much shorter life (though probably just as infrequent a roll-out) as the films did.
> classify tweets into four different categories: normal, aggressor, spam, and bully
Given that half of americans can't tell when a Brit is calling them an idiot, I don't hold out much hope for this.
> I must radically change my course module at UCL almost every year to keep up with the rate of change
Which gives the impression that those people who took the course a year or two ago have knowledge that is now obsolete.
Leaving aside the ethics of charging a student for a course that has such a short shelf-life (possibly shorter than the length of their degree course), isn't it a bit of a waste of time to teach stuff that will be out of date so quickly?
If a person did buy an (extremely expensive) electric car now, how long would it be until it became obsolete due to the advent of AVs?
If they were to become practical and available in 10-15 years, wouldn't it be better to hang on until then, rather than spend a lot more than a conventional ICE car costs now. Especially if that electric pride and joy had virtually no resale value when you wanted the benefits of an AV, just like everybody else.
If I was designing an electric car now, I'd have an eye on the upgradeabilty of it. So that at some future date it could be retrofitted with all the sensors and smarts to make it driverless. And even better, with a battery that didn't need recharging before you got to the end of your street!
Not only do many entries seem somewhat "unusual", but the order looks pretty dubious. I get the feeling that very few entries got more than a single person singing their praises. Further, that whoever it was at Lonely Planet that compiled the list has never actually been to the UK. Probably couldn't even find it on a map.
Yes. A quick search "for academic reasons" honest! yields a street price of $400-$3000 per ounce. Converting roughly from obsolete units starts at about $16k per kilo. Or USD 4.4million for this haul. Even at top whack (12 times more expensive), it's nowhere near this estimate unless the Aussie dollar is having a really tough time.
You have to wonder whether there will be people in the Sydney Constabulary asking their dealers for their money back.
If there was ever a piece of religious advice that is pertinent to the internet age it would be from the Acts of the Apostles (also made popular by Johnny Cash)
> Your point is obscure - what are you talking about?
One possible point is that $40 will buy you a 2GB + 8GB eMMC Orange Pi 3 (plus lots of extra on-board goodies).
if that platform had any working / usable software ir would be eating the RPi 3 (or 4)'s lunch every day of the week.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
