It's true cars have had long periods of slow progress. Things are going to get weird very quickly when fully self-driving cars arrive and we can just summon cars from a 24/7 active fleet via a phone app. - but that's a whole different thread.
I highly recommend reading a recent book entitled "You Look Like A Thing And I Love You" about AI. The author also writes a blog called AI Weirdness. She writes about the main types of AIs and their flaws. It was very interesting and puts into perspective some of the issues the world's top software engineers are struggling with, as they try to create this future. It's coming, but this book made it feel like we're a whole lot further out than we think.
Companies like ESRI(Geographical Information Systems company) have been implementing/developing AI systems for a while, Walmart started a thing called "Store No8" to use AI to help drive the next phase of consumer marketplace, Intel is also heavily invested in it. I'll have to read the book you mention, seems like there's an ever increasing attempt to implement in my(most) working spaces to see if the shoe fits and what can be done more efficiently or find unknown issues. A fun time in many market sectors.
I used to do a lot of AI around market pricing for a B2C store, on a top scale it was relatively dumb but when you deep dive and combine it with some current market tools driven by Google, credit card suppliers and social media it can give you some really scary shit
I highly recommend reading a recent book entitled "You Look Like A Thing And I Love You" about AI. The author also writes a blog called AI Weirdness. She writes about the main types of AIs and their flaws. It was very interesting and puts into perspective some of the issues the world's top software engineers are struggling with, as they try to create this future. It's coming, but this book made it feel like we're a whole lot further out than we think.
Companies like ESRI(Geographical Information Systems company) have been implementing/developing AI systems for a while, Walmart started a thing called "Store No8" to use AI to help drive the next phase of consumer marketplace, Intel is also heavily invested in it. I'll have to read the book you mention, seems like there's an ever increasing attempt to implement in my(most) working spaces to see if the shoe fits and what can be done more efficiently or find unknown issues. A fun time in many market sectors.
I used to do a lot of AI around market pricing for a B2C store, on a top scale it was relatively dumb but when you deep dive and combine it with some current market tools driven by Google, credit card suppliers and social media it can give you some really scary shit
Kind of amazing the amount of data companies have access to.
ESRI isn't that involved in machine learning. In can be used in predictive modelling of geospatial data and trends, but the data is created in underlying tables by outside sources, not by the GIS software itself.
They leverage it occasionally as a marketing tool for ESRI software, but the vast majority of companies who work with geospatial data are just doing on-the-fly analysis.
The unfortunate thing about big data is that its very rarely used to the benefit of society because often the policy changes that it might suggest are far more difficult than using it as a marketing tool.
Example, I could probably easily run analysis to determine if our local taxes on vacant land are high enough to drive liquidity in the market, and how vacant land prices might impact the minimum cost of housing on that land. However, proposing the increase in taxes on vacant land (policy) to spur development and reducing housing costs would be far more difficult than the few days of analysis necessary.
Meanwhile, I could go to a big marketing firm, and they'd be working with clients on target areas to sell Tequila and freakin killin it with sales, using social media and regional purchasing trends.
What's funny is that people complain when big data is used for public benefit, but they don't really seem to care when its used for corporate interests.
ESRI isn't that involved in machine learning. In can be used in predictive modelling of geospatial data and trends, but the data is created in underlying tables by outside sources, not by the GIS software itself.
They leverage it occasionally as a marketing tool for ESRI software, but the vast majority of companies who work with geospatial data are just doing on-the-fly analysis.
The unfortunate thing about big data is that its very rarely used to the benefit of society because often the policy changes that it might suggest are far more difficult than using it as a marketing tool.
Example, I could probably easily run analysis to determine if our local taxes on vacant land are high enough to drive liquidity in the market, and how vacant land prices might impact the minimum cost of housing on that land. However, proposing the increase in taxes on vacant land (policy) to spur development and reducing housing costs would be far more difficult than the few days of analysis necessary.
Meanwhile, I could go to a big marketing firm, and they'd be working with clients on target areas to sell Tequila and freakin killin it with sales, using social media and regional purchasing trends.
What's funny is that people complain when big data is used for public benefit, but they don't really seem to care when its used for corporate interests.
They are actually, a good friend helped start ESRI's AI team a few years ago.
ESRI is significantly more than the ArcGIS utilities you're referencing - that was simply their first-to-market platform.
What ESRI does in-house does not always apply to what is happening on a larger industry scale.
While yes, they do AI, I know hundreds of GIS professionals. None of them work with AI, and if they do, they are in academic setting.
Source - 10 years of working with GIS and going to lots of ESRI conferences and talking with lots of people who say "that AI stuff is cool, but who uses it?"
EDIT: ESRI does a lot of in-house development of technologies as a proof-of-concept. Maybe sometimes they'll work with some vendors or get some DOD contracts doing fun stuff, but again, what's possible, and what's commonly applied are two different things.
My point is, people think that everyone (especially the government) is using AI alongside Big Data to control us. It doesn't happen like that.
Your local government is WAY behind the curve of Big Data application. It's all happening at the corporate and academic levels.
It would not surprise me that Walmart is using it. They can afford, and rely on that level of market intelligence. They are a unique use case.
I think that's what was interesting about the AI book I recommended. It talked about specific applications, e.g. Amazon using AI to screen for candidates and the not-so hilarious biases that occurred. Same thing with prisons using AI to determine paroles.
AI's will always try to find shortcuts. The simplest solution and that often has unintended consequences. For example, Stanford used an AI to determine who should get vaccines. Somehow, it didn't select any front line workers! The people at highest risk.
Alot of it comes down to what we feed the AI. It needs huge data sets and prefers solving specific problems.
This is a far departure from bikes and I'm ok with it.
I remember that changeover where company folks were constantly correcting people in seminars and crap, but the guys who had been in the industry for awhile still called them "Ee-Sss-Are-I". Esri was acting like they had been deadnamed.