Your certainty isn't required. This is a strong empirical result from studies done on computer aided diagnosis.
The problem stems from the fact that AI systems are very strongly general (e.g. machine learning creates a generalization or abtraction from the training data), this specifically waters down and essentially excludes input which might still be pathological but deviates from the abstraction.
One of the most significant differences between people and the majority of machine learning AI systems is that people are much better at recognizing exemplars they have previously seen and "extra list features" (input which was not present at all in the training data), but not so good at recognizing *every* instance that matches the abstraction (i.e. we tend to miss boring, repetitive stimuli and focus on novel ones)
AI responds MUCH better to input that matches the generalization, but not as well to specific exemplars from the training set. (e.g. the AI's recognition "score" is always higher for a new input that is close to the learned generalization, than it is to an item from the training data that doesn't match the generalization as much). Humans are the opposite.
The human's diagnosis and management plan should be double checked by the AI, not the other way around
AI diagnostics is very good at catching all the "routine" findings, not so much at outliers. Human experts are better at catching outliers and novel cases, but will sometimes miss the "routine" ones for various reasons (e.g. fatigue etc...)
When you have the AI go first, it's results *further* constrain the human's diagnosis, leading to the outliers and novel cases being missed. Essentially the human just confirms the AI results and doesn't do any more work. This is the worst of both worlds - you're getting the weaknesses of both the AI and the human compounded together.
When you have the human go first, they're more likely to catch the outliers and novel cases, and then any of the "routine" results will be caught by the AI if they were missed by the human. This way you get the *best* of both worlds - is double plus good, yes?
If we weren't so focused on the failing strategy of "look for MOAR antibiotics", we could have had the superbug problem licked already with alternatives like phage therapy (which you can only get right now if you fly to Georgia).
What does it tell you when the famed "merkan innovation" can't even out innovate a tiny former soviet republic using 90yr old tech?
Software changing under me for idiotic reasons JUST after I have finished tweaking my workflow exactly the way I want it; is the NUMBER ONE reason I no longer use open-source software for production work at all. Also why I disable very goddamn "auto-update" feature out there...
Just use AdBlock. I've been using it for years and when I see a Slashdot page loaded by someone NOT using adblock it looks absolutely horrifying!!!
Install AdBlock, and enjoy a nice clean UI on just about every website anywhere. It really is refreshing not to see all that crap.
The Uncanny Valley has mostly to do with nearly perfect visual representations pushing our expectations for other aspects of a given simulation (e.g. body language, facial expressions etc...) beyond what we are able to produce. The discrepancy between the two then exacerbates our perception that "something is off"
The Uncanny Valley is mostly mental/perceptual and has virtually nothing to do with our currently crappy implementation of 3D film/TV. The issues there are mostly physiological (e.g. my stereoscopic convergence says this thing is right in front of my face, but my eyes are focusing on the TV 8ft away, WTF!)
When your stereoscopic vision converges at one point, and your monocular vision focuses at a different point entirely, your body tends to make the reasonable assumption that you've been poisoned, because NOPE.
Slashdot Mirror
... is not in changing the epsilon value of P
The real answer is in *requiring* 2 things:
Nobody could have possibly known in advance that hooking *everything* up to the internet was a security risk, right?
... this is like Alberta claiming they "don't have any rats" LOL
I'm not so sure about that.
Your certainty isn't required. This is a strong empirical result from studies done on computer aided diagnosis.
The problem stems from the fact that AI systems are very strongly general (e.g. machine learning creates a generalization or abtraction from the training data), this specifically waters down and essentially excludes input which might still be pathological but deviates from the abstraction.
One of the most significant differences between people and the majority of machine learning AI systems is that people are much better at recognizing exemplars they have previously seen and "extra list features" (input which was not present at all in the training data), but not so good at recognizing *every* instance that matches the abstraction (i.e. we tend to miss boring, repetitive stimuli and focus on novel ones)
AI responds MUCH better to input that matches the generalization, but not as well to specific exemplars from the training set. (e.g. the AI's recognition "score" is always higher for a new input that is close to the learned generalization, than it is to an item from the training data that doesn't match the generalization as much). Humans are the opposite.
It should actually be the opposite right now
The human's diagnosis and management plan should be double checked by the AI, not the other way around
AI diagnostics is very good at catching all the "routine" findings, not so much at outliers. Human experts are better at catching outliers and novel cases, but will sometimes miss the "routine" ones for various reasons (e.g. fatigue etc...)
When you have the AI go first, it's results *further* constrain the human's diagnosis, leading to the outliers and novel cases being missed. Essentially the human just confirms the AI results and doesn't do any more work. This is the worst of both worlds - you're getting the weaknesses of both the AI and the human compounded together.
When you have the human go first, they're more likely to catch the outliers and novel cases, and then any of the "routine" results will be caught by the AI if they were missed by the human. This way you get the *best* of both worlds - is double plus good, yes?
...or wherever, but this article IS a masterpiece of trolling aimed *directly* at Slashdots heart!
Don't feed the beast peeps
Yup^ The author is apparently either an idiot, or has never set foot in a software dev shop in their life!
THIS^
LOL
You have no clue what AI is, do you?
Also, your entire comment is a logical fallacy, as has already been pointed out.
What ridiculously hyperbolic garbage.
If we weren't so focused on the failing strategy of "look for MOAR antibiotics", we could have had the superbug problem licked already with alternatives like phage therapy (which you can only get right now if you fly to Georgia).
What does it tell you when the famed "merkan innovation" can't even out innovate a tiny former soviet republic using 90yr old tech?
Yeah, thought so.
...that if you had all software devs. from North America take the same test, you'd get comparable results.
Failing to include ANY measure of how this compares to performance HERE - marks this as just so much xenophobic propaganda
Same
I've been a software dev. for 20+ years, and this is the most accurate description of the planning process I have ever seen.
Kudos, coward.
... what's the YouTube of video games?
http://www.abandonia.com/
IoT ready.
The T in IoT is ... your brain
I for one, welcome our new Internet of Brains (IoB) overlords!
Same
Exactly.
If I had mod points, they would be yours, Mr. Coward sir. This is spot on.
Great! So you're totally fine if I swap all your QWERTY keyboards for Dvorak ones while you sleep?
Dvorak is clearly a better keyboard layout, so your user experience and productivity are going to be SO MUCH BETTER!
No, no, you don't need to thank me! Just trying to be helpful is all.
Software changing under me for idiotic reasons JUST after I have finished tweaking my workflow exactly the way I want it; is the NUMBER ONE reason I no longer use open-source software for production work at all. Also why I disable very goddamn "auto-update" feature out there...
Just use AdBlock. I've been using it for years and when I see a Slashdot page loaded by someone NOT using adblock it looks absolutely horrifying!!! Install AdBlock, and enjoy a nice clean UI on just about every website anywhere. It really is refreshing not to see all that crap.
it doesn't scale - you run out of stack pretty quickly
LOL - Clearly you skipped out on the "Algorithms 101" lecture on "Tail Recursion"!!!
LOL!
So I'm imagining the fact that 5.1 sound is basically the standard for home audio now?
??? Not at all.
The Uncanny Valley has mostly to do with nearly perfect visual representations pushing our expectations for other aspects of a given simulation (e.g. body language, facial expressions etc...) beyond what we are able to produce. The discrepancy between the two then exacerbates our perception that "something is off"
The Uncanny Valley is mostly mental/perceptual and has virtually nothing to do with our currently crappy implementation of 3D film/TV. The issues there are mostly physiological (e.g. my stereoscopic convergence says this thing is right in front of my face, but my eyes are focusing on the TV 8ft away, WTF!)
^^This.
When your stereoscopic vision converges at one point, and your monocular vision focuses at a different point entirely, your body tends to make the reasonable assumption that you've been poisoned, because NOPE.