However, is there anything you can do to minimize the pain of having to click the “Web” option buried in a menu every single time? The answer to that question is yes. Google does not make it easy, because its URLs seem extra-loaded with cruft these days, but by adding a URL parameter to your search—in this case, “udm=14”—you can get directly to the Web results in a search. That sounds like extra work until you realize that many browsers allow you to add custom search engines by adding the %s entry as a stand-in for the search term you put in.
I believe we have the “My dog is stupid, I beat her at three out of five games of chess” problem.
Self driving cars are a reality. Their most human characteristic is that they sometimes crash. GPT is great at recall, but it is prone to rabbit holes. Sound familiar? Human acts in machines are viewed as failures.
Artificial Intelligence is a property of machines. AI is not human. It is unlikely to ever pass a scientifically administered Turing Test. The more likely path is that concepts and skills will spin off creating smaller, specialized systems that economically meet industry needs. The result - new career paths and increased opportunities for employment.
AI is an amazing human accomplishment that greatly enhances some applications, while replacing others. Sadly, it has been overhyped and misrepresented by the media.
History didn’t have AI. And if it’s actually possible to make AI that can replace humans in general then under our system I’d expect everyone except the people in charge to be replaced and rendered unemployed. Why employ people whom you have to pay and can disagree with you when you can have machines that don’t? The ideal business is a CEO with no employees.
I think the importance of the Turing Test has also been overvalued. You could argue AI surpasses this, some of the time. So what?
If a weather forecast is wrong, people will point out the human equivalent has the same problems. If AI can be much less wrong about the weather for much longer periods this is a clear win for many people.
The benefit of AI in niche things like protein folding or drug design is fairly obvious. Driving is a tough task, and doing slightly better than human operators will not spare the companies from concerns or liability. I don’t doubt they drive well in perfect conditions, but I drive in Canada where we had two feet of snow in the last couple days and are expecting the same again. How much training data reflects that specific condition?
AI has also been used in early cancer detection. They showed it lots of data. It took a long time to realize the reported chance of cancer went way up if there was a ruler in the picture measuring something.
Since the smartest practitioners cannot say or guess how AI arrived at a conclusion, it is clear AI is prone to these errors when used by people unaware it is following “unintended rules”, which are legion^ and can be far more subtle than rulers. There is much potential and some very impressive beginnings. By companies which have sometimes shown disregard for consumers or wellbeing - use what the elite let their own kids use. Because of needed investment, overhype is the name of the game.
The greatest gains will be in niche areas and the biggest problems will come from playing Ouija in areas where the outcome is important and needs to be error free to the greatest possible extent. Still, folding clothes is said to be so variable that despite research robots still can’t do it well.
^Compare this to being able to correlate two factors versus proving cause and effect. If flu season correlates with wintertime, flu might well correlate with Christmas decoration frequency or themes emphasized on the Hallmark channel.
A CEO with AI is not going to stay in business because other people will have AI too. That’s the way it is with tech. The point is not that tech can replace a particular people skill. It does for you me and everyone. Now we all have the tech, people are still out there, can I leverage the resource of people to differentiate? That is still going to matter.
I don’t think things are guaranteed one way or the other. Maybe people still matter, but it’s more important who you know than what you know. Or you only learn the what by knowing the who. So that’s a game changer in itself.
People are also going to pit AI against AI. They will try to get their rival to pick up some bad data and come to wrong conclusions, while keeping the bad data out of their own AI.
Big ‘IF’! That hasn’t happened and we are not even close.
The Turing Test does not evaluate performance, rather it evaluates the degree, and nature, of human intelligence displayed by a system. In the Turing Test a question is posed to a machine and a human. The responses are evaluated to determine which was from the machine and which was from the human. Not which answer was correct.
I play a TT game with GPT 4. Either of us poses a question then GPT evaluates the answers. I always win because my answers are based on experience and emotion. I do not know of a system that can pass a valid Turing Test.
I know; I don’t consider the so-called “AI” in the news to be real AI at all. But human-equivalent AI can replace humans by definition, and the elimination of as much of the workforce as possible is a central goal of business.
You could have said a hundred years ago that history didn’t have the vast number of robots we now use in our homes, so the servants displaced by them would never find work again. But that’s not what happened; people were slowly pushed out of those jobs but it didn’t create permanent unemployment.
Farm equipment replaced most jobs that EXISTED, but unemployment isn’t above 50 percent.
But those robots weren’t anywhere near as capable as a human, so they couldn’t replace everyone. If they had been then yes, those people would have been permanently unemployed (or just rounded up and killed as surplus to requirements).
But they replaced IMMENSE numbers of people, and in fact they WERE more capable as humans in many jobs, and that’s why they replaced humans. There was once a day when something like 40% of all working age women were full time domestic servants; that number is now under two percent in the Western world, and many of them are performing different sorts of duties under that generic term. Once upon a time, most people were farmers; almost all those jobs were displaced by technology.
No AI you come up with will replace everyone - even if you had a generic AI that could replace a lot of desktop jobs, many jobs are not done at a desktop (a fact that I think surprises many.) The vast majority of people I work with can’t be replaced by software. They could in theory be replaced by robots but the robots don’t exist yet and from what I have seen firsthand, won’t anytime soon.
The majority of jobs that have ever existed - and I don’t mean 51 percent, I mean 99% of them - have been eliminated by machines. People have never been “rounded up and killed as surplus to requirements” because of an invention, that’s just silly. The unemployment rate isn’t 99%. Why do you think that is? Honest question; do you know why 99% of all jobs have been eliminated but we aren’t at 99% unemployment?
Which just means that, relevant to the thread subject AI is being overhyped. But if it could replace everyone including robots the result in our present system would be the bottom 99% being discarded as worthless scum and the remaining “elite” living in little bubbles of robot-provided luxury in a nation otherwise composed of corpses.
No, of course it wouldn’t. You don’t know what happens when technology dispalce jobs, which is why I asked.
Indeed, that clearly makes no sense at all just based on the obvious fact that all your scenario does is remove one of of every hundred people from the existing economy; it would be as if those people simply vanished. The other 99 still have an economy amongst themselves.
99 of every hundred, not 1 of every hundred. The rest would be exterminated, since the military would be robots by that point like everything else and the “elite” looks at them as subhuman vermin who deserve suffering and death.
No, that’s ridiculous. Stop this nonsense, man, let’s get back to discussing AI in reality. It’s not going to put 99 percent of people out of work - you still can’t explain why it is most people are still employed despite most jobs being eliminated - not result in the extermination of 99 percent of all people (the reason for which you also cannot explain.) We’re trying to discuss real life, not fantasy.
And I recall how the Inclosure Acts sent hundreds of thousands of UK Citizens to starve to death in parks.
BTW–how many Homeless people have died, needlessly, on US streets?
And, do you know what the term Social Murder {LINK} means?