8 ms·

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

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 "Fool me once, shame on you. Fool me twice, shame on me". Almost none of these companies can be trusted at this point. [1] Their refusal to ask OpenAI to be at the table really does not reflect well on them [2]. And the less said about the tenured professors who are now becoming company mouthpieces saying things like "we create products which cannot make profit but which is meant purely for data collection" the better [3]. And lastly, if these companies had such a sincere desire to "improve AI for the sake of humanity", how about they start by letting OpenAI (or a similar company) do a data audit of all the information they share so that we can actually be certain it is not just a data brokerage masquerading as a public service? I wanted to say that I wish the AI community will boycott this effort completely. I find it a bit worrying that this community now resides almost entirely within the walls of corporate America. [1] Interestingly, the only company which is even making noises about user privacy is Apple. Is it possible they saw something in this partnership that they didn't like? [2] https://twitter.com/OpenAI/status/781243032582578177 https://twitter.com/OpenAI/status/781243032582578177 [3] https://news.ycombinator.com/item?id=12428883 https://news.ycombinator.com/item?id=12428883

HNridiculous 10y ago
This has nothing to do with data collection or any other crazy conspiracy theories you are trying to peddle such as Open AI being intentionally excluded etc. Hell half of your comment does not even makes any sense. Allowing OpenAI to do "Data audit" LOL are you seriously high on drugs. Do you know what OpenAI is or does?? Nothing with Auditing! And yet this is the most upvoted comment in this thread along with second one where people have another set of crazy theories about Apple being "magically" superior.
- edem 10y ago
  Yet you didn't back your comment up with facts.
Taek 10y ago
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 that threshold is attainable only by a company with XXX million users, you inherently get a system where only those companies can innovate on the most cutting edge in machine learning. That's a bad thing for us, especially if we find those companies turning against us. Did Google start putting a liberal spin on search results to try to influence the election? We don't know! But we do know that they are using their position to manipulate ISIS recruitment, which means they aren't afraid to get their hands dirty with this kind of stuff. THERE IS NO OVERSIGHT! And you aren't going to get a search engine that can come close to competing with Google unless you can find a way to match their data volume before you have something competitive. It's a massive barrier. Facebook is enjoying the same sort of entrenchment. I've switched away from Facebook, and now I've lost touch with a lot of friends. They are my friends, why can't I export them? Why can't I export the data that belongs to me, and use it in some open-source format that allows me to reap the benefits of all the work that I spent tracking them down inside of Facebook's private garden? As a society, we need to start addressing these things, or we are going to end up with Internet giants that we don't have the means to get rid of, and that aren't at risk of being out-competed.
- crististm 10y ago
  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.
  - Taek 10y ago
    When John Green from Portland, Maine goes to Google and types in a search query for 'dirt on Mayor', I would imagine that Google is going to have an advantage over any algorithm, human or otherwise, that does not have access to things like the fact that it's John Green from Portland, does not have access to all the news coming out of Portland (at least not in the same, automated way that Google does), does not have access to Mr. Green's search history, and does not have data on the results millions of other searches related to 'dirt on mayor'. If you're going to compete with Google as the one-stop-shop for searches, you're going to need to be able to provide search results that are consistently competitive. And I think Google has gotten good enough that you simply can't do that without access to some portion of their moutains of data. Nobody else has search history for every single American, and without that you are going to be crippled when {Random American} tries to search for something.
patkai 10y ago
You just made me think about a few issues: 1. wouldn't these five be a cute quasi monopoly? (is pentapoly a word?) 2. isn't it brilliant how "they" - including many others, like Musk - make a big fuss about AI that is mostly already open and published, while what we should worry about is data? 3. is there any hope for an open society when most of us nerds - who are supposed to be all rebellious, undergroundy, security aware etc - with a lot of power to build the future can't wait to run into the arms of these big guys? We could use more @dhh type opinions.
- qznc 10y ago
  wrt 1.: "oligopoly"
- edem 10y ago
  The word you are referring to is oligarchy.
void-star 10y ago
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.
yAnonymous 10y ago
>Now that they have announced a partnership which potentially combines all this data together, I feel so much better! That's not going to happen. Their data is worth too much and it would lose its value if they shared it with competitors. >Interestingly, the only company which is even making noises about user privacy is Apple. Is it possible they saw something in this partnership that they didn't like? Apple aren't good at developing software. Having them in on this wouldn't benefit anyone but themselves so I doubt they were even asked. They're a company that makes shiny looking things with shiny looking interfaces.
mtgx 10y ago
So you're basically saying that this is a "Cartel on AI". What I'd like to see is all of the consortium members sign a pledge that none of them will sell or use their AI technology for defense/military purposes (mainly looking at you Microsoft/IBM/Amazon). I could see how 10 years down the line, the US government tries to co-opt the group to help it in its wars. This needs to be preempted now, in the same way they should've preempted calls for encryption backdoors by adopting end-to-end encryption years before and making it the status quo, before the government even gets a chance to think about it. They mostly failed in doing that, but I hope they won't fail with this.
- sharemywin 10y ago
  I wouldn't exclude facebook or google. I'm sure they've got big fat pipes that lead right to the NSA.
ptero 10y ago
They chose their initial partners somehow and did not include OpenAI. Maybe they thought this would make everything too transparent, maybe they see little financial benefits (those are private companies in general aiming to benefit shareholders), whatever. I see no problem with this. I wish OpenAI plays there eventually, but even now some of the partners go beyond what I would expect from a private company on fostering open AI research. I am however worried that this would make it even easier for a government agency to get bigger and bigger datasets of personal information
- wu-ikkyu 10y ago
  Why does it matter whether it is a government agency that has those data sets versus a private corporation?
  - ptero 10y ago
    Because: (1) government can ignore some protections that corporations are supposed to follow and (2) government can act on this in a way corporations cannot (use this information to argue for or issue search / seize warrants, etc.).
    - wu-ikkyu 10y ago
      >government can ignore some protections that corporations are supposed to follow Vice versa is also true. Google and Facebook can censor free speech more effectively than any other, legally. >government can act on this in a way corporations cannot (use this information to argue for or issue search / seize warrants, etc.) Google and Facebook don't need a search warrant to scour and social engineer billions of people's lives.

HNridiculous 10y ago

This has nothing to do with data collection or any other crazy conspiracy theories you are trying to peddle such as Open AI being intentionally excluded etc. Hell half of your comment does not even makes any sense. Allowing OpenAI to do "Data audit" LOL are you seriously high on drugs. Do you know what OpenAI is or does?? Nothing with Auditing! And yet this is the most upvoted comment in this thread along with second one where people have another set of crazy theories about Apple being "magically" superior.

edem 10y ago
Yet you didn't back your comment up with facts.

Taek 10y ago

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 that threshold is attainable only by a company with XXX million users, you inherently get a system where only those companies can innovate on the most cutting edge in machine learning. That's a bad thing for us, especially if we find those companies turning against us. Did Google start putting a liberal spin on search results to try to influence the election? We don't know! But we do know that they are using their position to manipulate ISIS recruitment, which means they aren't afraid to get their hands dirty with this kind of stuff. THERE IS NO OVERSIGHT! And you aren't going to get a search engine that can come close to competing with Google unless you can find a way to match their data volume before you have something competitive. It's a massive barrier. Facebook is enjoying the same sort of entrenchment. I've switched away from Facebook, and now I've lost touch with a lot of friends. They are my friends, why can't I export them? Why can't I export the data that belongs to me, and use it in some open-source format that allows me to reap the benefits of all the work that I spent tracking them down inside of Facebook's private garden? As a society, we need to start addressing these things, or we are going to end up with Internet giants that we don't have the means to get rid of, and that aren't at risk of being out-competed.

crististm 10y ago
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.
- Taek 10y ago
  When John Green from Portland, Maine goes to Google and types in a search query for 'dirt on Mayor', I would imagine that Google is going to have an advantage over any algorithm, human or otherwise, that does not have access to things like the fact that it's John Green from Portland, does not have access to all the news coming out of Portland (at least not in the same, automated way that Google does), does not have access to Mr. Green's search history, and does not have data on the results millions of other searches related to 'dirt on mayor'. If you're going to compete with Google as the one-stop-shop for searches, you're going to need to be able to provide search results that are consistently competitive. And I think Google has gotten good enough that you simply can't do that without access to some portion of their moutains of data. Nobody else has search history for every single American, and without that you are going to be crippled when {Random American} tries to search for something.

patkai 10y ago

You just made me think about a few issues: 1. wouldn't these five be a cute quasi monopoly? (is pentapoly a word?) 2. isn't it brilliant how "they" - including many others, like Musk - make a big fuss about AI that is mostly already open and published, while what we should worry about is data? 3. is there any hope for an open society when most of us nerds - who are supposed to be all rebellious, undergroundy, security aware etc - with a lot of power to build the future can't wait to run into the arms of these big guys? We could use more @dhh type opinions.

qznc 10y ago
wrt 1.: "oligopoly"
edem 10y ago
The word you are referring to is oligarchy.

void-star 10y ago

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.

yAnonymous 10y ago

>Now that they have announced a partnership which potentially combines all this data together, I feel so much better! That's not going to happen. Their data is worth too much and it would lose its value if they shared it with competitors. >Interestingly, the only company which is even making noises about user privacy is Apple. Is it possible they saw something in this partnership that they didn't like? Apple aren't good at developing software. Having them in on this wouldn't benefit anyone but themselves so I doubt they were even asked. They're a company that makes shiny looking things with shiny looking interfaces.

mtgx 10y ago

So you're basically saying that this is a "Cartel on AI". What I'd like to see is all of the consortium members sign a pledge that none of them will sell or use their AI technology for defense/military purposes (mainly looking at you Microsoft/IBM/Amazon). I could see how 10 years down the line, the US government tries to co-opt the group to help it in its wars. This needs to be preempted now, in the same way they should've preempted calls for encryption backdoors by adopting end-to-end encryption years before and making it the status quo, before the government even gets a chance to think about it. They mostly failed in doing that, but I hope they won't fail with this.

sharemywin 10y ago
I wouldn't exclude facebook or google. I'm sure they've got big fat pipes that lead right to the NSA.

ptero 10y ago

They chose their initial partners somehow and did not include OpenAI. Maybe they thought this would make everything too transparent, maybe they see little financial benefits (those are private companies in general aiming to benefit shareholders), whatever. I see no problem with this. I wish OpenAI plays there eventually, but even now some of the partners go beyond what I would expect from a private company on fostering open AI research. I am however worried that this would make it even easier for a government agency to get bigger and bigger datasets of personal information

wu-ikkyu 10y ago
Why does it matter whether it is a government agency that has those data sets versus a private corporation?
- ptero 10y ago
  Because: (1) government can ignore some protections that corporations are supposed to follow and (2) government can act on this in a way corporations cannot (use this information to argue for or issue search / seize warrants, etc.).
  - wu-ikkyu 10y ago
    >government can ignore some protections that corporations are supposed to follow Vice versa is also true. Google and Facebook can censor free speech more effectively than any other, legally. >government can act on this in a way corporations cannot (use this information to argue for or issue search / seize warrants, etc.) Google and Facebook don't need a search warrant to scour and social engineer billions of people's lives.