Quote: (05-14-2019 11:15 AM)kaotic Wrote:
So break down how AI is snake oil?
How is the radiology example debunked?
I want to know both sides of the story.
Here's why I'm inclined to understand his conversation.
I work in the software/technology sector, our bread and butter is manufacturing and defense contracting.
Automation isn't end times or global warming myth - it's already here.
We help customers help get prototypes done fast, automate internal processes, ECO's, life cycles, faster time to market, etc.
Some of our customer ARE using AI when developing products and they are replacing workers on the design/mfg teams.
While yeah we still need highly skilled workers in the design processes and in software tech, I can tell you that ALOT of our customers are looking at automation, whether it's internally or for their customers.
Automation is alot more reliable, makes us money, saves them money, time and material, etc.
Walter Black just pointed to Amazon as well, which I was going to mention, we've worked with simliar companies who the same type of logistical fulfillment jobs, same goes for agriculture here in the central valley.
"Lights out warehouse" is a very well known term in the industry we're in, it's not just some buzzword either.
While it's not Yang is screaming fire in a crowded theater, but he does see smoke on the horizon.
You mention his isssues are a couple decades away, but think about this.
Any time there is technological advancement it only accelerates every singe year forward, just like it did with computers and the internet.
I agreee there's bigger fish to fry, but Yang does bring up some valid points we need to be concerned about in the future.
Obviously there's quite a bit I disagree with him on.
I perceive you are earnest in wanting to know the sides of the conversation, and I figured you'd ask, so let me explain:
AI is a very generic term. So is futurism. As many examples as you can bring up of "look at what happened" I can show you 60s ideas and movies (2001) that are crazy off the mark as far as such sensationalism goes. And that's what it is, and for many reasons.
There has always been automation. Huge strides in automation took place within the last 150 years, just on a different level. They didn't cause more unemployment. Did they cause disturbances of sorts in life though? Sure. It's like a socratic understanding of the world, though. The more you know the more you don't know. The more you invent the more you need jobs to adapt to the new technology, even though much is displaced into the unknown category ... but look, we already have a track record showing this! It produced more jobs, greater standard of living, more prosperity. The right scenarios that caused this could be different now (bad gov't, evil globalists, etc), yes, but that's a different issue not purely automation or tech increases that in themselves are disturbing.
Humans still run AI and control it, if it is even allowed, and if (here's my point) you can even define it. Let me run a tangent to the radiology example. We have had AI in radiology for over 30 years. It is called computer aided detection in mammography. It would take too long for me to explain all the reasons, and there are many, why this has not improved outcomes for patients one iota, but that's the case. But you can talk about it being a computer that "detected" something, and you can sell it to people, or the government, or women's groups, about how much better it is. But it isn't. It does nothing, and if anything, it causes more problems.
Genghis Khan posts above and embarrassing lack of critical thinking, the stuff of NPCs. Yang posts a video of a marketing tool that just says "AI in radiology beats chinese radiologists" and you just take it on face value because it has people talking, a video of computers and a few percentages? I was flabbergasted when I saw that as a response. It means nothing. Yes, it is fake news. How could AI be "whooping the shit out of medical doctors" when you don't even know what you claim AI is doing, and you don't even know what (for example) a radiologist even does (or how many things he does?). You don't. At all. That's why you fall for this stuff. You have no idea what you don't know.
Let me set another thing straight: I'm not making the outrageous claims here. Futurists and AI sensationalists are. They bear the burden to substantiate their claims. I don't have to disprove them. They don't have anything, literally ZERO, to disprove (by definition it can't be disproved, since there is nothing to talk about). It's a backwards way of thinking on multiple levels.
Automation is one thing, computer learning is another. The human mind clearly lacks particular processing speed of computers, but it's not without advantage in other realms: it is far more flexible and adaptable than a computer will ever be, by definition. That's the problem that strict mathematician logic like the Architect guy from Matrix Reloaded, employ. They want graphs, algorithms, and have this idea that judgment and troubleshooting can be written into such paradigms. It can't. Insomuch as it can, it can only be added and adapted by humans. This is where most of the misunderstanding begins, if recognized. But there are technological triumphs of engineered machines and improved efficiencies in bridges and weaponry, etc. that fool us into thinking the higher level qualities of the human mind can be programmed by similar designs, but it is faith. They can't. The humans made the machines to begin with. That should help you understand the paradox, if you are humble enough to let it sink in.