DeepMind's Losses and the Future of Artificial Intelligence

guru · Aug 23, 2019

bone said:
I found the linked article to be long on vague sweeping pronouncements and hopelessly short on substance quite honestly.

Almost like a Psychology Professor decided to pronounce himself AI subject matter expert

Oh, just like most trading articles, where people who know the least have the most to say. They can easily alternate jobs writing on AI and trading. AI development is also like creating trading strategies - sometimes it works, but when it crashes everyone plays dumb and blames someone or something.

GRULSTMRNN · Aug 24, 2019

I use Google hangout calls all the time to make global calls, best quality among all services, worlds better than Skype or the like. Every company, Amazon and Microsoft and Google included venture into all sorts of project and it is the nature of the game that many such projects don't work out but the few successful ones shoot to the moon. Same with pharmaceuticals. This is the business model.

AI was not hot 20 or 30 years ago. It is hot right now and has been for the past 5-6 years. Don't confuse ai with ml.

guru said:
I never used their Google Hangouts, never used Google+, and really don't remember using Google for much except searching. All their "projects" don't matter if no one is using them and they keep dying. So Google is a failure outside of their main business.
But AI and reinforcement learning will be useful, just the question is to what extent and what are their limitations. AI was actually "hot" 20 or 30 years ago. Also, 38% of American cars were electric in 1900: https://cleantechnica.com/2018/02/25/38-percent-american-cars-electric-1900/

guru · Aug 24, 2019

GRULSTMRNN said:
I use Google hangout calls all the time to make global calls, best quality among all services, worlds better than Skype or the like. Every company, Amazon and Microsoft and Google included venture into all sorts of project and it is the nature of the game that many such projects don't work out but the few successful ones shoot to the moon. Same with pharmaceuticals. This is the business model.

AI was not hot 20 or 30 years ago. It is hot right now and has been for the past 5-6 years. Don't confuse ai with ml.

Well, I had to write basic artificial neural network code in school, in 1980s. We had "Artificial Intelligence" magazine in 1988. And we thought AI will advance much faster. But I guess you're not a tech guy, just like to pretend you know tech more than people who do tech.

GRULSTMRNN · Aug 24, 2019

What particular tools did you use in school in the 1980s, would love to hear that. Because there was no efficient programming language nor hardware available to properly train neural networks at that time. Not sure what you are talking about but you surely do not seem to understand how DNNs operate and what their computational requirements are. Did LSTMs exist back then? GRUs? RNs? Most were invented in the late 1990s.When were Billions of data points and labeled data available for training? Fast enough CPUs and GPUs? Highly programmable FPGAs? Sorry but I don't like to continue talking with you on this issue. I work with dnns every single day and don't think you understand the advances and applications in today's environment. Perhaps 30 or 40 years ago you were some hot shit but it does not sound like you fully grasp the potential and what problems AI already solves today

guru said:
Well, I had to write basic artificial neural network code in school, in 1980s. We had "Artificial Intelligence" magazine in 1988. And we thought AI will advance much faster. But I guess you're not a tech guy, just like to pretend you know tech more than people who do tech.

Simples · Aug 24, 2019

GRULSTMRNN said:
What particular tools did you use in school in the 1980s, would love to hear that. Because there was no efficient programming language nor hardware available to properly train neural networks at that time. Not sure what you are talking about but you surely do not seem to understand how DNNs operate and what their computational requirements are. Did LSTMs exist back then? GRUs? RNs? Most were invented in the late 1990s.When were Billions of data points and labeled data available for training? Fast enough CPUs and GPUs? Highly programmable FPGAs? Sorry but I don't like to continue talking with you on this issue. I work with dnns every single day and don't think you understand the advances and applications in today's environment. Perhaps 30 or 40 years ago you were some hot shit but it does not sound like you fully grasp the potential and what problems AI already solves today

Ai still haven't solved the human condition.

GRULSTMRNN · Aug 24, 2019

It never was the aim to accomplish such.

Simples said:
Ai still haven't solved the human condition.

KeLo · Aug 24, 2019

“It is very hard to predict, especially the future.”
- Niels Bohr (world renowned physicist)

tommcginnis · Aug 24, 2019

GRULSTMRNN said:
AI was not hot 20 or 30 years ago. It is hot right now and has been for the past 5-6 years. Don't confuse ai with ml.

The movie was quite topical, stemming from a scifi piece from 1969.
Don't know where you were.
https://m.imdb.com/title/tt0212720/

GRULSTMRNN said:
What particular tools did you use in school in the 1980s, would love to hear that. Because there was no efficient programming language nor hardware available to properly train neural networks at that time.

Sorry, dude. You're Deep-Wrong at this point. I entered grad school in the early 80s, and machine learning was A Thing then. First semester's Math Econ. Learn or die. The term itself is another couple of decades older than that. And yeah, the tie between AI and improvements to the human condition was picked up by science fiction writers from the AI founders/popularizers. (And, this is old news.)

guru · Aug 24, 2019

GRULSTMRNN said:
What particular tools did you use in school in the 1980s, would love to hear that. Because there was no efficient programming language nor hardware available to properly train neural networks at that time. Not sure what you are talking about but you surely do not seem to understand how DNNs operate and what their computational requirements are. Did LSTMs exist back then? GRUs? RNs? Most were invented in the late 1990s.When were Billions of data points and labeled data available for training? Fast enough CPUs and GPUs? Highly programmable FPGAs? Sorry but I don't like to continue talking with you on this issue. I work with dnns every single day and don't think you understand the advances and applications in today's environment. Perhaps 30 or 40 years ago you were some hot shit but it does not sound like you fully grasp the potential and what problems AI already solves today

We used assembly, C, and BASIC. You can write anything, including basic neural network in any programming language. I didn’t say they were advanced or deep or had billions of data pints, or anything else at that time. AI was advancing since then, and it encompasses machine learning, so when ML and NNs were hot then people talked about AI. 40 years from now people may not remember what was DNN or RL, and may use a different form of AI, arguing that 40 years ago AI couldn’t have been hot. Quantum computers are also hot for a while, every 5-10 years.
And I’m very aware of AI capabilities today, and I’m not saying that it’s useless, or that it will be useless in the future. It’s useful now but it has its limits, which right now sometimes can be solved without AI, or accepted as limitations. I do have plans to use AI/RL myself but can’t know yet whether it will do better than I can do without it. Unlike you, I’m not a psychic who can predict future, you hot shit.

GRULSTMRNN · Aug 24, 2019

ML != Deep Learning.

tommcginnis said:
The movie was quite topical, stemming from a scifi piece from 1969.
Don't know where you were.
https://m.imdb.com/title/tt0212720/

Sorry, dude. You're Deep-Wrong at this point. I entered grad school in the early 80s, and machine learning was A Thing then. First semester's Math Econ. Learn or die. The term itself is another couple of decades older than that. And yeah, the tie between AI and improvements to the human condition was picked up by science fiction writers from the AI founders/popularizers. (And, this is old news.)