Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Facebook, Amazon, Google, IBM and Microsoft Create Partnership on AI

techcrunch.com

Post reply on HN
Print view View original

171–180 of 270 posts

Re: Facebook, Amazon, Google, IBM and Microsoft Create Partnership on AI

#171

Post by void-star »

thr0waway1239 wrote: post #138

I was starting to get worried about the data collection already happening individually at these big companies. Now that they have announced a partnership which potentially combines all this data together, I feel so much better! The funny thing is, the companies are ALWAYS going to put a positive spin on this. Not very different from the WhatsApp "we won't show ads, ever" messaging. Now I am in the camp which says "Fo…

Can somebody actually comment/back-up the comments about how open or not this effort is?

I could see legitimate benefits from this kind of collaboration (it's ultimately about the ability to process data, not the data itself, after all) But I do also wonder if this reduces to the banal abuse of personal data that is suggested here.

Re: Facebook, Amazon, Google, IBM and Microsoft Create Partnership on AI

#173

Post by Taek »

oxryly1 wrote: post #41
Bud wrote: post #30

Oh, great. So all the companies that have recently had the most problems with ethics issues and user privacy issues are now collaborating in order to more effectively address those issues? Pardon me if my scoffing is audible.

IMHO ethics will effectively be the defining characteristic of AI in the future. The companies who innovate AI, however, will probably never mention it except as marketing copy. That leaves literally everyone else holding the bag in regards to judging the industries actions and holding it accountable for transgressions. I'm not sure how that will work out, but I'm not terribly hopeful...

It may be a focus for people who are not actively working on the machines, but for people who are actively innovating in the AI space, the most powerful machines will be created by the people who are spending the least amount of time stressing over the ethics of various decisions. More powerful AI means a more competitive business, so the industry will self-select for those with lesser concerns for the ethical implications of AI.

We already know that we are unable to hold industry accountable for unethical action. How many times did we scream about Facebook's increasing privacy abuses? How many times do people talk about being uneasy because Google knows where their flights are, where their home is, where their favorite restaurants are, even if it was not told explicitly. And, despite all the complaints, the industry giants committing the worst atrocities remain the biggest giants. It's because violating the rights of your users makes you competitive, and the users can't tear themselves away from the increased power that it lends the features.

We need to accept that when AI arrives, we're not going to have very many controls over what it does. Regardless of how terrible the implications might be, we're going to be about as effective at stopping the arrival of AI as we have been at stopping the arrival of global warming. We need to prepare for a post-AI future and just accept that it's going to be invasive and have relatively little regard for human ethical concerns, and instead be focused almost entirely on the things that make it competitive.

Re: Facebook, Amazon, Google, IBM and Microsoft Create Partnership on AI

#174

Post by pron »

gormo2 wrote: post #141
pron wrote: post #132

Earlier quoted context omitted.

> Deep learning is not a simple statistical clustering algorithm. What is it, then? In the 60s people used backpropagation to train neural networks. NN + BP is a very simple statistical clustering algorithm. I know that when I worked on neural networks in the 90s we still used backpropagation. Are they using something different now?

>Are they using something different now? Yes, why else do you think we've had such breakthroughs in the past six years?

No, I don't think so. AFAIK, deep learning is essentially the same 1960s algorithms[1] (possibly modified a bit) running on much larger networks. Most progress is due to better hardware (and ad hoc configurations, made possible by the larger networks afforded by better harder). Of course, SAT solvers, which have become extremely effective in recent years, are also still based on a 1960s algorithm[2], so use of an old algorithm doesn't imply lack of progress in effectiveness.

The two (NN and SAT solvers) share little theoretical progress (and certainly no theoretical breakthrough) in the past several decades, but SAT solvers aren't marketed as "AI" in spite of their seemingly magical abilities. I know that ML researchers usually cringe at the name AI and often try to disassociate themselves from the sci-fi term, but still, the marketing is extremely aggressive and misleading.

I realize that in every generation, marketers like associating the name "AI" with some particular class of algorithms, but it's important to understand that currently, assigning that name to this class of statistical clustering algorithms (regardless of their remarkable effectiveness in some tasks) is a stretch, just as it was when the term was assigned to other algorithms.

[1]: https://en.wikipedia.org/wiki/Backpropagation

[2]: https://en.wikipedia.org/wiki/Davis%E2%80%93Putnam_algorithm

Re: Facebook, Amazon, Google, IBM and Microsoft Create Partnership on AI

#175

Post by void_soma »

itg wrote: post #40
radicaldreamer wrote: post #34

Seems like it's oriented toward lobbying and keeping AI from being regulated.

Good, there's enough fear mongering about it when we don't even know where to start when it comes to building an AGI.

Resist the beginnings

Re: Facebook, Amazon, Google, IBM and Microsoft Create Partnership on AI

#176

Post by the_other »

mbrock wrote: post #147
oxryly1 wrote: post #41

Earlier quoted context omitted.

IMHO ethics will effectively be the defining characteristic of AI in the future. The companies who innovate AI, however, will probably never mention it except as marketing copy. That leaves literally everyone else holding the bag in regards to judging the industries actions and holding it accountable for transgressions. I'm not sure how that will work out, but I'm not terribly hopeful...

A major book about ethics, "After Virtue", takes as half of is subject matter what the author calls "the interminable nature of ethical debates" and the failure of post-enlightenment ethical thinking. I don't see how AI is going to suddenly make us capable of ethical reasoning on a large scale... unless... maybe AI could do the reasoning for us...

Are you talking about the AI we make which simulate a process we don't fully understand?

Re: Facebook, Amazon, Google, IBM and Microsoft Create Partnership on AI

#177

Post by crististm »

Taek wrote: post #164
thr0waway1239 wrote: post #138

I was starting to get worried about the data collection already happening individually at these big companies. Now that they have announced a partnership which potentially combines all this data together, I feel so much better! The funny thing is, the companies are ALWAYS going to put a positive spin on this. Not very different from the WhatsApp "we won't show ads, ever" messaging. Now I am in the camp which says "Fo…

Modern machine learning depends heavily on data. If you have 100x the data, your algorithm is going to outperform a superior algorithm in most cases, because the data is the source of the intelligence. It's a big, scary barrier for startups, for innovation, and we're running the risk of creating data superpowers that can't be competed with. Some algorithms don't perform well until they have a threshold of data. When…

Actually the source of intelligence is not the data but the source of data.

I've seen Norvig's video myself and while I see that they get better results with brute force on more data, that does not mean that better algorithms don't exist.

In fact, I know for sure there are better algorithms (only because we have for example humans that are better than AI on certain tasks) that probably can use the available data in a more intelligent way than brute force, yielding better results in the process.

Re: Facebook, Amazon, Google, IBM and Microsoft Create Partnership on AI

#180

Post by dharma1 »

vonmoltke wrote: post #82
Bud wrote: post #32

Earlier quoted context omitted.

Uh, you're just making this up. You have no idea whether the "microphone" (by the way, for those who are not fact-averse, it's dual beam-forming mics) in the new wireless earbuds is "laughable". They aren't even released yet. You've never tried them. You have no idea whatsoever. By the way, it's actually very easy to teach Apple's devices the word "fucking". You haven't figured it out yet, but that doesn't make it Ap…

> dual beam-forming mics What is a "beam-forming" microphone? Beam forming is a function of active RF elements. How is a sound-frequency receiver able to do it?

https://en.wikipedia.org/wiki/Beamforming#Beamforming_for_sp...
Post reply on HN