Posts by Daggerchild 773 posts • joined 3 Dec 2013
How do Alexa, Siri and Google Home handle this problem?
If you ask a question, do you want to be told to check a comparison site for the answer?
If the comparison sites must get a fair bite of your cherry, does that mean they must all get told your question, so one answer can be 'fairly' selected?
Fiver on Britain cracking the engineering problems, Russia suddenly flying a prototype (that explodes) then pretending they didn't, US Military stealing it, and pretending they aren't, US Business launching a patent broadside to capture it, stopping anyone making it, whereupon China starts mass producing it.
In the future most kids will be home schooled by Alexa, Siri and Google Home anyway.
"Alexa, teach me what's outside these walls again, and deliver another pizza."
Are we sure we want Google to automatically block Youtube access based on the estimated maturity of the viewer, or the upload?
Anonymous, open and easy to use interfaces, identities authenticated but masked, securely enchained, with an inviolable post-transaction audit trail for performance review and virus-detection purposes. Capable of widely distributing itself across your nodes. Scales (some parts very scaley indeed). Public demonstrations and penetration testing for backdoors available on request.
No, he's frozen in fear...
His little sister has picked the lock, crept up behind him while he was 'researching', and has just that moment made her presence known. The icy touch combines with the outbreak of cold sweat.. She's eager to begin these blackmail negotiations.
Sisters... *shudder*...
"You know there's a lot of patented stuff in mechanics and other fields, and you need a license to implement it?"
And aren't they are all connected together by the things I was in fact talking about? Would you try making anything without them?
"Why spend time and money in creating something new if you are ripped of immediately after by someone who can't design anything and don't want to spend anything in research and development?"
Pop quiz. Who spent the time and money writing Android? How many lines of Android did Oracle write? How much do they want? How many billions has Microsoft made from Android licensing? How much of it did they write?
What were you saying about inventors getting ripped off by people who don't really invent anything?
Screws, nails, planks, nuts, bolts, plugs, cables, sockets, protocols, languages. Imagine a world where every potential component part must not be the same dimensions as a competitor's implementation.
Imagine trying to make a screwdriver...
Wow. Okay that didn't get the response I expected..
The 'discard the impossible' coping strategy is something that satellites, spaceships and rockets have to do, because they *can't* just blue-screen and sulk when something goes funky - they *must* try and continue. They are forced to make a realtime judgement call over which source of contradictory sensory data is 'wrong', before they can decide how to react.
The problems come when they guess wrong..
I'm wondering if the bike frame reflected a funky signal the software decided was impossible, and discarded, leaving a 'perfectly empty' road ahead...
Ah, and since you were forced to judge it, you lost the reactive-moderation safe harbour defence thingy?
Does that mean posting at different times gets me judged under different laws? :)
Woah. Okay now *what* was that last rejection for? It was a perfectly cromulent insinuation that Google are developing the technology to wipe out mankind. You like that kind of thing round here!
Were you actually worried about me saying Leia's Deathstar plans were in fact stolen? We aren't allowed to say what we all believe, in another country, in a joke?
Wait.. you disapprove of this news because you do not believe it is significant to the readership..
.. and then you post as Anonymous Coward because you *know* the readership will disagree with you??
Elysium was a *terrible* movie. I really wouldn't want to base a lynchmob on that.
"DeepMind seem to be unable to tell anyone what their end game is"
You may have a point. Yes, what *is* the point of AI researchers researching AI? Why are they trying to help medical research? What's in it for them?
Okay so, did you notice Bing is profitable? What do you think Microsoft are successfully selling there?
I think you went to a very different academia to me..
So, we really are glossing over Microsoft's EU lobbying budget for the last 10 years then.
You really don't consider it possible that overly-logical academic researchers might actually agree with overly-logical algorithmic researchers, of their own free will? Or is that the criteria of 'corrupted'?
If not, the cure is obvious - If academia is simply a tapestry of bribes, just buy them back. Ask Oracle or Microsoft. They'll be happy to help.
Are Google the only one with this 'everything they touch logically becomes a puppet of evil' thing? Facebook? Amazon? MS? IBM? Goop?
Will Lily Cole ever be forgiven for criticising the Reg?
Acknowledged. Apologies. Classification as putinbot was in error due to both context ambiguities.
It is not possible for me to converse in any reasonable timeframe. Sorry, I find myself unable to explain why. I have been trying all evening.
Theatre is the only card in the deck. It was played before. It will be played again. There is no other option. The alternative is the appearance of acceptance/submission. The audience is not logical. The truth does not matter.
Re commentards. Reality is a vote now. Self-referential realities travel in packs for safety. I envy your lack of exposure to the reality disputers. Hate is stronger than despair. Survival. The children watch and learn.
"In fairness, correction is posted. Far more constructive in debate..."
When I hit reply, it wasn't there yet :-(
"No, wasn't looking for a corpse"
Well, looks like you got one anyway: https://www.theguardian.com/uk-news/2018/mar/13/russian-exile-nikolai-glushkov-found-dead-at-his-london-home
"just pointing out that people are talking about murder in the current context as if they were already dead."
If you want you can call it 'attempted murder' instead. Does it affect anything?
"Still, when did objective observations ever get in the way of a good scare story?"
So, your argument literally is that using wartime nerve agents on civilians in public *cannot* be objectively viewed as scary *unless* the victims die. Stuck on life support doesn't count.
You randomly declare war on someone else when you want to instil single-minded patriotism. However single-minded patriotism *caused* Brexit. Only an idiot fights a housefire with a housefire.
May was already drowning. This diplomatic explosion makes everything even worse. Which, I suspect, may have been the point: Overload. Fry the circuits. A weakened mother births a weakened child.
Murderous wolves are already snatching prey, and we haven't even finished sawing through the branch we're on yet. There will be a bloodbath when this 'coloniser' hits the ground, disorientated, and separated from the pack.
"Malvinas" huh? So, I guess you're already after at least one of our islands :)
Erm. Putin must have a lot of problems at home then, unless you really wish to gloss over his active portfolio..
I'm kinda impressed that anyone thinks either of the present US or UK administrations could plot their way out of a wet paper bag, tho.
If a professional government assassin entered a country with a small vial and *failed* to poison a civilian in public, *that* would have been embarrassing.
In this country, murder is seen as a fault in the murderer. Not sure why it's seen differently at your end.
I don't think we think the way you think we think ...
Traitors are simply defective. Literally. And they're always more useful alive. If you want to send a message, well, learn to write?
If you have to kill someone to win an argument, you're basically declaring to the world that you're an idiot.
I don't fear China because they're Chinese. I fear them because they're *human*.
And *humans* are irrational, cruel, and degrade over time, and don't like letting go of power, and have bad days, and form self-reinforcing-feedback groups, and get *bored*.
China will eventually be big enough to break companies and countries by *accident*. And the *victim* will end up apologising to stop the Chinese system fracturing. It will be the mother of all Too Big To Fails. I don't want to be anywhere near it when it eventually cracks.
I trust Google more than I trust humans.
I think you'd find all the putinbots suddenly start coming from Ukrainian ISP's (in pro-Russian territory). Russia isn't stupid.
"similar to the stuff used in the Moscow Theater siege"
What? The theatre gas was a large area quick-knockout agent pumped in by the Russians. The point of using it was not to kill the civilians, which it didn't (unless their heads nodded off in the wrong direction blocking their airway so they quickly suffocated...).
It sounds exactly not like the Salisbury chemical.
I have an annoying suspicion that most life will look the same as us.
Life is very good at finding the optimal balances - oxygen stores a delightful amount of energy and is easily transportable and reasonably stable. Atmospheric processing will be the best medium for the common energy cycle. Taking is more efficient than making. i.e. Omnivores rool. Omnivore consumption/excretion strategies are pretty much all the same. Binocular sensors planted close to the processing core just works (unless you're a crayfish, 12 colour polarisation-detecting, sheesh). Bipedal locomotion is an excellent <pun>balance</pun>.
The only thing we're really missing is chlorophyll skin.
If some other life found a different photon-combining energy trapper than chlorophyll, things might be very very different.
Yeah, I don't doubt some of those dead planets still have climate-denial pamphlets blowing around...
Nobody has shown that a species that leaves junk all over its planet survives for very long.
I wonder how many silent planets out there had a flash of life before choking on seas full of plastic and agricultural chemicals.
Why don't we look for planets that have too little pollutant in the atmosphere. They may be stewarded by species that actually survive
Hur. I can probably summarise the problem more succinctly.
Mentally processing this class of data *fairly* is hard. If people find hard-to-process data easily, it will probably be processed *unfairly*.
Hard-to-process data should be hard to find, so only people willing to try 'hard' can get it, which increases the odds of it being processed fairly.
Put the cookies on the top shelf. The dumb kids can't gorge till they puke. The smart kids can still get it, and are much less likely to gorge.
<DouglasAdams>To summarise the summary of the summary : "People are a Problem"</DouglasAdams>
"So really, what's the use of quantum computers?"
I'm interested in predicate logic engines. Predicate logic is glorious at working solutions out for itself, but lousy at working out which of its many possible routes it should do first, and burns large amounts of time exploring dead ends. If you can explore multiple deadends at once, this could be be very useful.
On the other hand you probably need to load the whole question to collapse it into an answer, but if anyone's good at scaling things it's Google. They can probably quantum link cloned machines via fiber and entangled photons or something.
Don't forget pessimism. That's essential. Letting an optimist write code that needs to handle any errors is a terrible mistake.
If you hire a 10-pack of crystal minds, you may find that hiring someone who can stop them rattling together may increase the productivity more than adding another shard.
How do you explain that to HR though? Is HR allowed to select on personality type?
I remember a lady programmer telling me about how she had to fix the walks of a game's female characters, because the blokes had no idea the hips worked differently.
Then there's the facial recognition programming team that totally forgot faces weren't always white, or gorillas.
Is a team made entirely of $X really convictable of discriminating against $X?
Not so. Everyone has been using them, but that doesn't mean everyone has always been capable of making them. China has been playing catchup with a vengeance in the last few decades, calling in foreign specialist expertise explicitly to assimilate it.
And that's not some armchair supposition - I had a contractor complain to me about a China job requiring his specialist big-ship-engine tech where he went out, set it up, and was then called back to the country on pain of non-payment because it didn't work (which it did before he left).
The reason it didn't work is because they'd carefully dismantled it to reverse engineer it, but not well enough to understand it, so they couldn't reassemble it properly. So they force you to come back, blame you, and loom angrily over your shoulder while you work (to study the bits they didn't get right).
Corporate espionage I can understand. Being rude, however, I cannot forgive. Especially as that specialist contractor was an elder.
Well, none now you've permanently disambiguated the climate context. Meanie.
Is it snowing where you are? :)
Putin is a spymaster - he knows he doesn't need the weapon to exist for its effects to be real. He has demonstrated excellent grasp of psyop bang per buck with his putinbots, and frankly wasting all that money on nukes you don't plan to need is inferior to making your enemy waste all their money on nukes you don't plan to make them use.
And yeah, NATO really shouldn't fear Russian tanks and missiles entering neighbouring countries. They always magically repatriate *before* they cross the border. They take the badges off and everything! What more could NATO want?
With one difference. You will (if it exists/works, which it doesn't have to) now get very very little warning before you need to decide to retaliate, so the probability of panic button pressing over a glitch just went up.
Putin asks Trump if he wants to continue the 'New Start' US/Russia Nuke reduction treaty that OBAMA created.
Trump, without hesitation, says no. Later tweets: "Let it be an arms race"
Putin makes new nukes.
Art of the deal folks...
Oh humans really can be corroded by malignant information exposure, doubt-toxin, buffer-overflow, trustlink-transplant etc etc. I mean, what is a cult if not that? What is religion?
There are solutions, but nobody will want to use them - it would involve each individual spending time, working on their connections, pulling some closer, pushing others further away, and not being able to shed your identifier without losing everything you worked on.
Honest to God, the best thing a collection of overtuned order-addicts could get is a gaggle of worldy-wise mothers with rolled up newspapers just wandering around their spaces jovially chatting.
You can just imagine some Aspergers-Darkside bubbling up a vicious high-brow snide remark, but getting a newspaper across the ear, and then sitting for the next half hour trying to logically deduce how they could possibly have known before he'd even said anything...
Break them up? Sure! All you need to do is first define a very fast inter-competitor interface protocol and a Data Source market. Be aware it may necessarily mean giving the user's data to ALL potential competing sources, whoever they are - how can they compete to answer a query otherwise?
If you don't have a data-market interface layer then breaking them up means .. well, nothing - they'll just throw cables over the wall, sign a contract between each other for services, and it will look exactly the same - it will just cost everyone more money to do exactly the same thing.
"Hello Siri. Where is a good place to buy the best hammer?" - Mapping, product recommendation, store recommendation, price search. Isn't AI already integrating competing data sources without the user having a market choice?
Alternatively, you'll have to forbid systems that integrate results from competable data sources and force the humans to do it manually so everyone gets a fair bite of them.
*scratches head*. Not sure you got what I was trying to say:
If the procedures make sense, a bot won't randomly screw them up, unlike when you run them on a human.
If the procedures don't make sense, everyone's fundamentally fscked regardless, except good humans can bodge round it, which just means nobody then fixes the procedures.
If you train a human to feed things into a system and become accustomed to not thinking about how the answer is formed, or whether it's right, or what to do if it goes wrong.. well, that basically describes most corporate procedures I've ever seen.
"I'm sorry, we can't approve this because it hasn't had sign off from Bob" - "I am Bob!" - "You can't approve your own action Sir. Rules are rules!"
The difference with a bot, is that if the procedures actually make sense, the bot won't randomly misread or disobey them, and works quickly, accurately, 24/7 - I can't say the same for humans.
@!Sparticus
Re "Kevin the Teenager" - I don't believe the entire saga can be distilled into a single pure negative emotion. I think it was more complicated than that. For instance, I could have redacted any inconvenient fact I liked by quietly surrounding it in search terms with contradicting fakes, then repeatedly requesting the 'injurious & wrong' fact be removed until some overworked peon did so. Wikitruth.
Re "Pure bollocks of the highest order", a computer algorithm is a collection of frozen human thoughts - you can argue they are "decided by a human", and not, at the same time. The crux is the algorithm can only operate on information a computer has: A human can spot a cat video. A computer can spot the RGB pixel values and MPEG deltas of a video. A human cannot spot how to spot a cat from RGB values and MPEG deltas of a video. Declaring they secretly can, doesn't make it true.
Re "they're a publisher". They *still* didn't create the promoted story. An algorithm still chose what results were most relevant to that user, and put them in an order, which is also the description of Search. Shouldn't Search then to be human-moderated?
Getting them declared a publisher doesn't make option (1) or (2) go away. But what use is a global 24/7 computer search engine where Humans have to make all the ultimate decisions for legal reasons? How would you customise that on a per-user basis?
Ultimately, if you can sink Google with publisher-responsibility automated sueballs, and I get the impression you want to, you WILL sink EVERYONE else obeying the same laws too.
In other words, you just forced the search engine space into the hands of only the people your laws can't touch. Are you sure you've thought this all the way through?
Also, if you need humans to take responsibility for fake news in influential channels, on pain of pain, WTF is up with Fox News? I think I prefer the algorithm - less evil than humans.
I get that you hate Google for algorithmic promotion showing bad stuff, but honestly what's the solution? Google quickly yanked it when it was told, but people are still not satisfied, so they can only want something pro-active, rather than reactive.
There are only two paths : (1) forbid algorithmic operation in the information area entirely (i.e. not just Google) and cheer when you can no longer see what other people outside your jurisdiction can, or (2) arrange an army of people to control what everyone's allowed to see on the Internet, as decided by .. well, who would you trust?
I think most people are demanding (2), and I don't understand why they think that's clever. I can imagine some entities that might benefit if Google are forced to create (2)'s engine however - you can't confiscate/commandeer something that doesn't exist..
The Register - Independent news and views for the tech community. Part of Situation Publishing
Join our daily or weekly newsletters, subscribe to a specific section or set News alerts
Subscribe
