"The Singularity Is Near" by Ray Kurzweil, How Close??
We'll be 100% muslim by then.
http://en.wikipedia.org/wiki/Predict...y_Ray_Kurzweil
I'm hoping this one comes true:
I'm hoping this one comes true:
http://www.futuretimeline.net/index.htm
I found this website that is loosely based on Singularity. It's basically a website where they predict future events such as the emergence of thoughtcrime, full immersion virtual reality, nanotech fabrics, and etc. If you don't take it too seriously it's really, really interesting/entertaining. 2020-2029 is where it starts to get really good.
I found this website that is loosely based on Singularity. It's basically a website where they predict future events such as the emergence of thoughtcrime, full immersion virtual reality, nanotech fabrics, and etc. If you don't take it too seriously it's really, really interesting/entertaining. 2020-2029 is where it starts to get really good.
Calling dibs on beachfront property
http://www.futuretimeline.net/index.htm
2020-2029 is where it starts to get really good.
2020-2029 is where it starts to get really good.
Ok, well, google books has some of the relevant pages. Let's see how it goes from there.
40% correct (mouse/kb), 60% wrong.
This is a bit fuzzy, because home wifi is obviously quite common, but as far as ever-present wifi access to a network goes, the only candidate is cellphone networks, and those are anything but what's described. Furthermore, the wifi INCLUDED in computers does not access such things. Wrong.
While I list this as a separate prediction, it is part of the same paragraph as the last one, which lends further support to my interpretation of what he meant by "the wireless network". This, of course, is obviously wrong. The vast majority of books/music/movies have physical objects, commercial-grade software has plenty, and none of it goes fast through world-wide-wifi. Wrong.
Uh, no. Wrong
I have no idea.
The only ubiquitous LUI that I'm aware of are voice-activated cellphone commands, and there's no natural-language understanding there. Wrong. (and see the next two for more indication of what he means by LUI)
I'm not aware of any such things (remember his definition of LUI). And anybody who's used speech over the phone wants to punch it in its virtual face. Wrong.
Uh.. yeah. thisclose. And there are no LUI systems to make reservations that I'm aware of, certainly not ubiquitous ones. Wrong.
He's right, but he was also right in 1999. Past-posted.
I'll give him this one. Kindle and friends really have taken off. Correct.
Since he doesn't give any frequency, merely existence, and AFAICT this technology does exist, correct.
Computers do not routinely identify their owners from faces. I'm sure they could, but they don't. 1 Correct, 1 Wrong.
Wrong.
I've never heard of this specific technology (very-high frequency interference), and it's certainly not replacing everyday speakers. Wrong.
1 is correct, 2 and 3 are not. 1 correct, 2 wrong.
Correct AFAIK
Both of these existed well before the book. Double past-post.
I am unaware of any such machines built molecule by molecule. Wrong.
Overreach, so wrong.
Not 100% on 1, but it seems reasonable, and 2 is correct. 2 Correct.
Again, relatively annoying to figure out what he means. If he's talking about webpages or powerpoint, it's a past-post, and if he's talking about govt documents, business documents, etc, it's wrong. 50% past-post, 50% wrong.
Can't judge without the full context of the book. If he's talking about the world, it's totally wrong. If he's talking about the US, is he right?
Wrong.
Wrong.
Wrong.
He we are again with his "wireless communication". Astrologers would be jealous, but I think as much as computers at school use wifi, this is technically correct. Correct.
There are a few other education-related things that I have no idea about.
Videoconferencing is commonplace, but live virtual lectures in an educational setting, not so much. Wrong.
I'll go ahead and call this correct. Takes infi research and nipicks abound, but I think he's on the right track. Correct.
WTF? A print-to-speech reading machine for the blind that can read signs. How would it even know where to look for a sign to read? Wrong.
Wrong.
And that's the end of google books.
This is a very underwhelming performance. In some cases, he's predicted things that have no reason to be, and in others, he's massively overestimated the spread of technologies.
Communication between components, such as mice/mike/display/printer/kb use short-distance wireless
Computers routinely include wireless technology to plug into the ever-present world-wide network, providing reliable, instantly available, very-high bandwidth communication.
Digital objects such as books, music, movies, and software are rapidly distributed as data files through the wireless network and typically do not have a physical object associated with them
The majority of text is created using speech-recognition software.
speech software is very accurate, far more than human transcriptionists.
Also ubiquitous are language-user interfaces, which combine speech recognition and natural language understanding.
for routine matters, such as simple business transactions and information inquiries, LUIs are quite responsive and precise.
LUIs are frequently combined with animated personalities. Interacting with one to conduct a purchase or make a reservation is like videoconferencing, except the person is simulated.
Computer displays have all the display qualities of paper- high res, contrast, large angle, and no flicker
Books, magazines, and newspapers are now routinely read on displays that are the size of small books
Virtual retinal displays
Computers routinely including moving picture image cameras and are able to reliably identify their owners from faces.
In terms of circuitry, 3-D chips are commonly used..
speakers are being replaced with tiny chips that create audible frequency from the interaction of very-high frequency tones, making small speakers very robust for 3d sound.
PCs go 1 teraflops, supercomputers go 20 petaflops, and distributed computing reaches 20 petaflops.
increased interest in massively parallel neural nets, genetic algorithms, etc, althought most computation is still normal.
research has been initiated on reverse-engineering the brain through after-death scans and through high-res MRI.
Autonomous nano-engineered (built molecule by molecule) machines have been demonstrated and include their own computational controls- but are sitll impractical
...Computers play a central role in all facets of education...
The majority of reading is done on displays, and paper books/documents are being rapidly scanned and stored
Documents circa 2009 routinely include embedded moving images and sounds
Students of all ages typically have a computer of their own
..which is a thin tabletlike device weighing under a pound
Students interact with (those) primarily by voice and pointing with a pencil-like device
Keyboards still exist, but most textual language is created by speaking
Learning materials are accessed through wireless communication
There are a few other education-related things that I have no idea about.
Learning at a distance- lectures and seminars where the participants are scattered- is commonplace
Learning is becoming a significant portion of most jobs.....
Print-to-speech reading machines for the blind are now very small, inexpensive, palm-sized devices that can read paper books, other printed documents, and other real-world text such as signs and displays.
a virtual seeing-eye dog that reads and talks, and has GPS, has been introduced.
And that's the end of google books.
This is a very underwhelming performance. In some cases, he's predicted things that have no reason to be, and in others, he's massively overestimated the spread of technologies.
Looks we have to do our own estimations and predictions. Truth will arrive.
Did you really say the vast majority of music has a physical object!??!?!
Despite that, for some reason I feel like you might actually NOT be trolling... IDK.
Let's see if I can address some of your blindness w/out taking too much of my time.
My printer is wireless. They are cheap, awesome and commonplace. Wireless mic's and displays exist as well. Obviously there's not much demand for wireless displays, so you aren't familiar with them.
Yes computers routinely include wireless technology. You have to be trolling to argue against that one.
Digital objects. Lol you're obv wrong. Itunes, youtube, bittorrent. Come on bro.
He was wrong about the popularity of speech-recognition software. He was in that business, so it's understandable why he'd have high expectations for it. All (maybe just almost all?) the technologies he predicted for it exist (1 or 2 of them invented by him actually). They just aren't commonplace or in high demand.
LUI's are more common than you think, and are awesome. Rosetta Stone is the nuts, and is super super popular.
Computer displays of paper... He's referring to kindle like screens that did not exist in 1999. Furthermore, if it existed in 1999, it was not past posted.
Important point:Kurzweil first published most of these predictions in his 1990 book The Age of Intelligent Machines. He has reiterated them in most of his books. It should be assumed most of these predictions were made in the late '80s.
A quick google search of "3d computer chip" shows results coming up from 2004. Not gonna spend much time on this, but have to assume that prediction is right.
He predicted 20 petaflops. IBM's Sequoia will have 20 petaflops by 2012. You're nitpicking if you're calling this 1980's prediction wrong.
Obv he predicted reverse-engineering of the brain before it started. Definitely not past-post.
Yes computers do play a central role in all facets of education. They already did for college kids in the '90s, high school kids shortly after, and now elementary school kids as well. Do you know any elementary school teachers? I personally dated one that worked in one of the poorest districts in the US, and the amount they did with computers was pretty big.
The majority of reading is obv mostly done on displays. Novels are the only form of reading where it's close. Every other form of reading is dominated by displays.
The thin like tablet weighing under a pound is very close to accurate. Kindles, Ipad's, Macbooks, Iphones. Each individually falls short of his prediction, but together... he's pretty much correct.
Learning at distance lectures are insanely common. There are so many remote colleges. Don't know why you added in the world virtual to make it seem like he meant lectures were happening on Second Life or something.
FML. As I come to that, now I think you're trolling and I'm being an idiot for responding. Oh well. No sense in deleting all this now.
Despite that, for some reason I feel like you might actually NOT be trolling... IDK.
Let's see if I can address some of your blindness w/out taking too much of my time.
My printer is wireless. They are cheap, awesome and commonplace. Wireless mic's and displays exist as well. Obviously there's not much demand for wireless displays, so you aren't familiar with them.
Yes computers routinely include wireless technology. You have to be trolling to argue against that one.
Digital objects. Lol you're obv wrong. Itunes, youtube, bittorrent. Come on bro.
He was wrong about the popularity of speech-recognition software. He was in that business, so it's understandable why he'd have high expectations for it. All (maybe just almost all?) the technologies he predicted for it exist (1 or 2 of them invented by him actually). They just aren't commonplace or in high demand.
LUI's are more common than you think, and are awesome. Rosetta Stone is the nuts, and is super super popular.
Computer displays of paper... He's referring to kindle like screens that did not exist in 1999. Furthermore, if it existed in 1999, it was not past posted.
Important point:Kurzweil first published most of these predictions in his 1990 book The Age of Intelligent Machines. He has reiterated them in most of his books. It should be assumed most of these predictions were made in the late '80s.
A quick google search of "3d computer chip" shows results coming up from 2004. Not gonna spend much time on this, but have to assume that prediction is right.
He predicted 20 petaflops. IBM's Sequoia will have 20 petaflops by 2012. You're nitpicking if you're calling this 1980's prediction wrong.
Obv he predicted reverse-engineering of the brain before it started. Definitely not past-post.
Yes computers do play a central role in all facets of education. They already did for college kids in the '90s, high school kids shortly after, and now elementary school kids as well. Do you know any elementary school teachers? I personally dated one that worked in one of the poorest districts in the US, and the amount they did with computers was pretty big.
The majority of reading is obv mostly done on displays. Novels are the only form of reading where it's close. Every other form of reading is dominated by displays.
The thin like tablet weighing under a pound is very close to accurate. Kindles, Ipad's, Macbooks, Iphones. Each individually falls short of his prediction, but together... he's pretty much correct.
Learning at distance lectures are insanely common. There are so many remote colleges. Don't know why you added in the world virtual to make it seem like he meant lectures were happening on Second Life or something.
WTF? A print-to-speech reading machine for the blind that can read signs. How would it even know where to look for a sign to read? Wrong.
My printer is wireless. They are cheap, awesome and commonplace. Wireless mic's and displays exist as well. Obviously there's not much demand for wireless displays, so you aren't familiar with them.
Yes computers routinely include wireless technology. You have to be trolling to argue against that one. /// Digital objects. Lol you're obv wrong. Itunes, youtube, bittorrent. Come on bro.
He was wrong about the popularity of speech-recognition software. He was in that business, so it's understandable why he'd have high expectations for it. All (maybe just almost all?) the technologies he predicted for it exist (1 or 2 of them invented by him actually). They just aren't commonplace or in high demand.
LUI's are more common than you think, and are awesome. Rosetta Stone is the nuts, and is super super popular.
Computer displays of paper... He's referring to kindle like screens that did not exist in 1999. Furthermore, if it existed in 1999, it was not past posted.
Important point:Kurzweil first published most of these predictions in his 1990 book The Age of Intelligent Machines. He has reiterated them in most of his books. It should be assumed most of these predictions were made in the late '80s.
A quick google search of "3d computer chip" shows results coming up from 2004. Not gonna spend much time on this, but have to assume that prediction is right.
He predicted 20 petaflops. IBM's Sequoia will have 20 petaflops by 2012. You're nitpicking if you're calling this 1980's prediction wrong.
Obv he predicted reverse-engineering of the brain before it started. Definitely not past-post.
The thin like tablet weighing under a pound is very close to accurate. Kindles, Ipad's, Macbooks, Iphones. Each individually falls short of his prediction, but together... he's pretty much correct.
Learning at distance lectures are insanely common. There are so many remote colleges. Don't know why you added in the world virtual to make it seem like he meant lectures were happening on Second Life or something.
It's past-posted in 1999. Maybe if he actually gave two ****s about being accurate in what he wrote, he would have noted in 1999 that it was already happening, and not left it as a prediction for the future.
I think you are reading too strongly into the word "common". Airplanes are a common form of travel. It doesn't mean that they compose of even as much as 1% of travel. Pretend he said ordinary instead of common. They are synonyms after all.
Last point I want to make is the distinction between monitors and kindle screens. I don't know the terminology, but there is something about the screen of the kindle (and other technologies now) that don't hurt your eyes in the way computer screens do. You might say this is nitpicking, but this was precisely the technology that was needed to make reading books on a screen feasible, and this is exactly what RK said the breakthrough would be 20 years ago.
The rest of your post just isn't worth responding to.
Edit: 1 more...
The prediction is for "music albums". And, yes, I think it's safe to say that it is incorrect that "music albums typically do not have a physical object associated with them".
Digital music sales are approximately half of non-music digital sales. However, most digital music isn't purchased (citation needed, I know. Don't have one), which puts it over the top of cd sales.
Saying his books need citations or more info is completely ridiculous when all you have read is a few paragraphs that weren't even written by him.
Read tSiN. You will be impressed.
(this comment is not directed at only Tom).
Read tSiN. You will be impressed.
(this comment is not directed at only Tom).
Wow, TomCowley just /threaded this discussion vnh
Page 191 (present tense describing 2009):
"Research has been initiated on reverse engineering the human brain through both destructive scans of the brains of recently deceased persons as well as non-invasive scans using high resolution magnetic resonance imaging (MRI) of living persons."
Past-posting and self-inconsistency FTL. Brilliant.
None of that **** is in the prediction on page 190. Cool story bro.
I guess you and Kurzweil share the trait that admitting you were ridiculously wrong about something just isn't worth doing.
Don't you own this book? I also told you what my source was- google books excerpt. More specifically, page 190, second full paragraph on the page. Music albums. (and yes, I wrote music the first time, which was not what was in the book. Oops.).
"Research has been initiated on reverse engineering the human brain through both destructive scans of the brains of recently deceased persons as well as non-invasive scans using high resolution magnetic resonance imaging (MRI) of living persons."
Past-posting and self-inconsistency FTL. Brilliant.
Last point I want to make is the distinction between monitors and kindle screens. I don't know the terminology, but there is something about the screen of the kindle (and other technologies now) that don't hurt your eyes in the way computer screens do.You might say this is nitpicking, but this was precisely the technology that was needed to make reading books on a screen feasible, and this is exactly what RK said the breakthrough would be 20 years ago.
The rest of your post just isn't worth responding to.
You didn't use the word albums, and you didn't link to your source, so I have no idea how you can add that word in now.
Even if you are super conservative, and are thinking linearly instead of exponentially, this is a completely ridiculous view.
There is no reason to think that computing purchasing power (price of calculations per second) won't increase dramatically over the next 20 years.
Once human level AI is created, it will be trivially cheap/easy/efficient to simulate millions of years worth of thought in a fraction of a day.
There is no reason to think that computing purchasing power (price of calculations per second) won't increase dramatically over the next 20 years.
Once human level AI is created, it will be trivially cheap/easy/efficient to simulate millions of years worth of thought in a fraction of a day.
Page 191 (present tense describing 2009):
"Research has been initiated on reverse engineering the human brain through both destructive scans of the brains of recently deceased persons as well as non-invasive scans using high resolution magnetic resonance imaging (MRI) of living persons."
Past-posting and self-inconsistency FTL. Brilliant.
"Research has been initiated on reverse engineering the human brain through both destructive scans of the brains of recently deceased persons as well as non-invasive scans using high resolution magnetic resonance imaging (MRI) of living persons."
Past-posting and self-inconsistency FTL. Brilliant.
This is much more specific than the page 121 quote, so it's adding new information, and nothing in it is inconsistent with page 121 which discusses freezing dead human brains, and the very early stages of reverse engineering.
Clearly non-invasive scans using high resolution MRI's of living persons was not done to reverse engineer human brains in 1999.
For kindle, your argument is that it doesn't exist on page 190? Are you even real? That's your argument??
And even if you want to add the word albums, my argument still applies. Your best case argument is that "it's too close to call" for this debate.
Go is many orders of magnitude more complex than chess, and has many orders of magnitude more game states.
Computers routinely include wireless technology to plug into the ever-present world-wide network, providing reliable, instantly available, very-high bandwidth communication.
Are you arguing these things don't exist?!?
How can you possibly argue that this quoted prediction is inaccurate?!
Furthermore, the wifi INCLUDED in computers does not access such things.
I just don't understand how you can possibly believe the things you are saying.
ZJ -
Haven't followed this recent exchange but I think the point is that if you are attempting to grade someone's past predictions you have to be careful not to unconsciously alter them to fit the subsequently known facts.
If RK wanted to set wider goalposts for himself he should have made his predictions more vague. That way we would all be correctly less impressed when they turned out to be true.
To take the specific claim about wireless technology, it's pretty close but it seems to me it's not strictly true. If he had watered down "routinely" and said something like "many computers will be able to...", he would have been correct, but then he would have been making a significantly easier claim.
Also I think tensions are running a little higher than they need to in this thread. Durka wasn't helping with his incessant and icky "nutlicker" comments.
Haven't followed this recent exchange but I think the point is that if you are attempting to grade someone's past predictions you have to be careful not to unconsciously alter them to fit the subsequently known facts.
If RK wanted to set wider goalposts for himself he should have made his predictions more vague. That way we would all be correctly less impressed when they turned out to be true.
To take the specific claim about wireless technology, it's pretty close but it seems to me it's not strictly true. If he had watered down "routinely" and said something like "many computers will be able to...", he would have been correct, but then he would have been making a significantly easier claim.
Also I think tensions are running a little higher than they need to in this thread. Durka wasn't helping with his incessant and icky "nutlicker" comments.
Rating Kurzweil's ability to predict the future by looking at what percentage of his predictions are correct is idiotic. This is a terrible way to see how good he is at predicting the future. He could have just increased the date to 2030 for all his predictions and he would have been 80%+ correct.
A better way to see how good he is at predicting the future would be to compare him to other people who take predicting the future seriously and compare their accuracy for specific prediction dates (eg compare how accurate they were for 2010 predictions, then 2020 etc, making sure that they made their predictions at roughly the same date). Their predictions would have to be about the same thing (because predictions can be of different difficulty). I could predict that in 2030 computers are more popular than today - that's an easy prediction. Some are not so easy.
Kurzweil was wrong for his 20 petaflop prediction but he still deserves credit for being only 2 years out on a tough prediction. His predictions should not be rated as only correct or incorrect. The degree of which his prediction was incorrect should be considered and the difficulty of the prediction should be considered.
A better way to see how good he is at predicting the future would be to compare him to other people who take predicting the future seriously and compare their accuracy for specific prediction dates (eg compare how accurate they were for 2010 predictions, then 2020 etc, making sure that they made their predictions at roughly the same date). Their predictions would have to be about the same thing (because predictions can be of different difficulty). I could predict that in 2030 computers are more popular than today - that's an easy prediction. Some are not so easy.
Kurzweil was wrong for his 20 petaflop prediction but he still deserves credit for being only 2 years out on a tough prediction. His predictions should not be rated as only correct or incorrect. The degree of which his prediction was incorrect should be considered and the difficulty of the prediction should be considered.
ZJ -
Haven't followed this recent exchange but I think the point is that if you are attempting to grade someone's past predictions you have to be careful not to unconsciously alter them to fit the subsequently known facts.
If RK wanted to set wider goalposts for himself he should have made his predictions more vague. That way we would all be correctly less impressed when they turned out to be true.
To take the specific claim about wireless technology, it's pretty close but it seems to me it's not strictly true. If he had watered down "routinely" and said something like "many computers will be able to...", he would have been correct, but then he would have been making a significantly easier claim.
Also I think tensions are running a little higher than they need to in this thread. Durka wasn't helping with his incessant and icky "nutlicker" comments.
Haven't followed this recent exchange but I think the point is that if you are attempting to grade someone's past predictions you have to be careful not to unconsciously alter them to fit the subsequently known facts.
If RK wanted to set wider goalposts for himself he should have made his predictions more vague. That way we would all be correctly less impressed when they turned out to be true.
To take the specific claim about wireless technology, it's pretty close but it seems to me it's not strictly true. If he had watered down "routinely" and said something like "many computers will be able to...", he would have been correct, but then he would have been making a significantly easier claim.
Also I think tensions are running a little higher than they need to in this thread. Durka wasn't helping with his incessant and icky "nutlicker" comments.
There's nothing in the prediction that requires wirless being used for fixed computers that are trivial to connect via cables. Where cables aren't available wireless is common and it doesn't require any special effort in most cases - what else is required for 'routine'.
Kurzweil was wrong for his 20 petaflop prediction but he still deserves credit for being only 2 years out on a tough prediction. His predictions should not be rated as only correct or incorrect. The degree of which his prediction was incorrect should be considered and the difficulty of the prediction should be considered.
Rating Kurzweil's ability to predict the future by looking at what percentage of his predictions are correct is idiotic. This is a terrible way to see how good he is at predicting the future. He could have just increased the date to 2030 for all his predictions and he would have been 80%+ correct.
A better way to see how good he is at predicting the future would be to compare him to other people who take predicting the future seriously and compare their accuracy for specific prediction dates (eg compare how accurate they were for 2010 predictions, then 2020 etc, making sure that they made their predictions at roughly the same date). Their predictions would have to be about the same thing (because predictions can be of different difficulty). I could predict that in 2030 computers are more popular than today - that's an easy prediction. Some are not so easy.
Kurzweil was wrong for his 20 petaflop prediction but he still deserves credit for being only 2 years out on a tough prediction. His predictions should not be rated as only correct or incorrect. The degree of which his prediction was incorrect should be considered and the difficulty of the prediction should be considered.
A better way to see how good he is at predicting the future would be to compare him to other people who take predicting the future seriously and compare their accuracy for specific prediction dates (eg compare how accurate they were for 2010 predictions, then 2020 etc, making sure that they made their predictions at roughly the same date). Their predictions would have to be about the same thing (because predictions can be of different difficulty). I could predict that in 2030 computers are more popular than today - that's an easy prediction. Some are not so easy.
Kurzweil was wrong for his 20 petaflop prediction but he still deserves credit for being only 2 years out on a tough prediction. His predictions should not be rated as only correct or incorrect. The degree of which his prediction was incorrect should be considered and the difficulty of the prediction should be considered.
Why should we care about his relative accuracy when the discussion involves the absolute likelihood of one of his predictions coming true?
He could be mostly right about a prediction / some of his predictions, but missing by a few years. The question is: how far off is he?
Seesm a bit over the top. Smartphones (even most mobile stupid phones) are computers and are routinely linked together wirelessly in huge numbers, if anything the prediction has be so fullfilled that its taken so much for granted that people don't notice its happened. Other mobile computers routinely connect to the same networks.
There's nothing in the prediction that requires wirless being used for fixed computers that are trivial to connect via cables. Where cables aren't available wireless is common and it doesn't require any special effort in most cases - what else is required for 'routine'.
There's nothing in the prediction that requires wirless being used for fixed computers that are trivial to connect via cables. Where cables aren't available wireless is common and it doesn't require any special effort in most cases - what else is required for 'routine'.
The natural reading depends partially on when he made that particular prediction.
If he made it in the era of desktop PCs it seems to me he would have understood "computers" to mean computers that had at least as much functionality as desktops, including such banal functions as being able to print stuff. So it seems to me cell phones are ruled out.
Also I definitely don't agree with your second point about wired desktops not counting against him because it is trivial to make them wireless for some extra $ if you care to.
He seems to be making a prediction about the default consumer computing standards of the future. Those standards are important because they are a barometer of how advanced computing technology in general is. The improvement in top-end stuff have a trickle-down effect that makes consumer stuff cheaper.
For example we could extend your reasoning to say that he would have been correct in predicting that "computers routinely have 3D monitors", because it is in fact trivial to upgrade to them. It's just quite expensive and really pointless because it's not yet standard.
In both cases it's probably just a matter of time... but then time is very much of the essence when you're making predictions.
ETA - Meh I'm starting to come around to your view regarding "routine" on this one... it's close either way I think.
It would really depend on the nature of the prediction how much if any partial credit he gets.
The problem is in the realm of computing technology lots of predictions approach very high degrees of probability the more time you allow.
But anyway nobody is saying his predictions are no better than random. Just that you have to avoid various pitfalls in grading anyone's predictions.
Feedback is used for internal purposes. LEARN MORE