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

Don't Trust Google (2002)

idlewords.com

Post reply on HN
Print view View original

21–30 of 87 posts

Re: Don't Trust Google (2002)

#21

Post by ChuckMcM »

stock_toaster wrote: post #18

Something like 5 out of the first 16 comments on this post were all from accounts (1 each) created under 1 hour ago. Weird.

It isn't surprising. There are lots of reasons why people are unwilling to put their "primary" HN account at risk when they want to say something which can be used against them (or simply down voted into oblivion). When I worked at Google I was very careful about what I said outside of Google, there is very little the company doesn't know about its employees. Being a trouble maker always looked bad on your calibration scores and that translates into real money.

So when folks want to contribute to the conversation in some way (either to support, attack, or deflect it seems) but don't want to do so openly, the new accounts come out. One motivation for having a 24 hr waiting period on posting is to mitigate that, but sometimes folks do bring good data under an extra layer of anonymity.

Re: Don't Trust Google (2002)

#22

Post by mynameisvlad »

nick_riviera wrote: post #6
TeMPOraL wrote: post #2

And 14 years later there still "is no evidence or even reason to suspect that Google is not being an honest broker. The searches give good results, the rankings seem fair, the service remains free, and we haven’t heard of anyone being arrested for running a dodgy query." So while the author is right about potential dangers mentioned in the penultimate paragraph, it's been more than a decade and those fears haven't re…

Perhaps not yet but they have amassed all the data the author describes and much more. For example, I just drove half way across the UK. Google knows that because I was using Google maps. I stopped at Pizza Hut. Google knows that. I stopped at a BP garage. Google knows that. I searched for houses in an area in the UK via right move. Google knows that because the embedded map knows my Google cookie... And so forth...…

Google doesn't specifically know you stopped at Pizza Hut or the BP garage. It uses coarse location (GPS being too taxing on battery to get constant location updates) which has an accuracy of 100+ meters. So it would know you stopped somewhere around the Pizza Hut, but not specifically that you went inside. Could have just as easily gone to the Tesco next door.

Re: Don't Trust Google (2002)

#23

Post by mynameisvlad »

antocv wrote: post #8
tillywilly wrote: post #3

Seems like recently these anti-Google posts are being dredged up at a noticeable rate. Is there a reason behind it or is all just ad hoc? For instance I wonder how OP stumbled about the link.

Thanks to Google these links and many others are kept down under the rug.

This HN page is currently #3 for the search query "Don't Trust Google". So I'm going to call BS right there.

Re: Don't Trust Google (2002)

#24

Post by nick_riviera »

mynameisvlad wrote: post #22
nick_riviera wrote: post #6

Earlier quoted context omitted.

Perhaps not yet but they have amassed all the data the author describes and much more. For example, I just drove half way across the UK. Google knows that because I was using Google maps. I stopped at Pizza Hut. Google knows that. I stopped at a BP garage. Google knows that. I searched for houses in an area in the UK via right move. Google knows that because the embedded map knows my Google cookie... And so forth...…

Google doesn't specifically know you stopped at Pizza Hut or the BP garage. It uses coarse location (GPS being too taxing on battery to get constant location updates) which has an accuracy of 100+ meters. So it would know you stopped somewhere around the Pizza Hut, but not specifically that you went inside. Could have just as easily gone to the Tesco next door.

Actually yes it does know I stopped at both of those location.

1. The BP garage has a wireless AP with a known SSID to them. Location services uses this, the cell network and the GPS to build an accurate location. There are no other wireless networks at that location. I was at this location for around 10 minutes. Google location history shows this location on the mark.

2. They know that I was at the Pizza hut because I left a review in Google+ and was approximately near it at the time and stayed in roughly the same location for about 90 minutes. Google location history shows this location on the mark.

It's all about correlating events and data, not specifically accurate bits of data.

Granted these are all assumptions but they would stand as evidence in a court of law.

Now I'm not fussed about this myself as it's something I trade for the service, but the data is there.

Re: Don't Trust Google (2002)

#25

Post by kephra »

> I have one idea for how to do that, through a network of peer-to-peer search engines, and I hope to post it soon.

This was 2002 ... Is there any follow up posting on it?

I'm actively developing a censor free distributed search engine, with the abilities to parse the layout of a page into fields like author, title, date, abstract, article, pictures. The system is able to login into sites, hide itself as a normal browser to scrap the dark side of the web. Nodes and leafs of the network will communicate with tor (nyi), and act like a torrent of search engines.

Use the /chat/ or contact from my homepage, if you want to discuss my vision of a Google killer.

Re: Don't Trust Google (2002)

#26

Post by kryptiskt »

kephra wrote: post #25

> I have one idea for how to do that, through a network of peer-to-peer search engines, and I hope to post it soon. This was 2002 ... Is there any follow up posting on it? I'm actively developing a censor free distributed search engine, with the abilities to parse the layout of a page into fields like author, title, date, abstract, article, pictures. The system is able to login into sites, hide itself as a normal bro…

There were some blog posts in 2003, referring to a paper. Hmmm... the link is broken, but this should be it: http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.58.4...

Re: Don't Trust Google (2002)

#27

Post by pmiller2 »

ChuckMcM wrote: post #21
stock_toaster wrote: post #18

Something like 5 out of the first 16 comments on this post were all from accounts (1 each) created under 1 hour ago. Weird.

It isn't surprising. There are lots of reasons why people are unwilling to put their "primary" HN account at risk when they want to say something which can be used against them (or simply down voted into oblivion). When I worked at Google I was very careful about what I said outside of Google, there is very little the company doesn't know about its employees. Being a trouble maker always looked bad on your calibratio…

> Being a trouble maker always looked bad on your calibration scores and that translates into real money.

Wow. Did you actually see first hand anyone who was punished by Google for things said outside the workplace? And did this extend to topics other than Google?

Re: Don't Trust Google (2002)

#28

Post by TeMPOraL »

pmiller2 wrote: post #27
ChuckMcM wrote: post #21

Earlier quoted context omitted.

It isn't surprising. There are lots of reasons why people are unwilling to put their "primary" HN account at risk when they want to say something which can be used against them (or simply down voted into oblivion). When I worked at Google I was very careful about what I said outside of Google, there is very little the company doesn't know about its employees. Being a trouble maker always looked bad on your calibratio…

> Being a trouble maker always looked bad on your calibration scores and that translates into real money. Wow. Did you actually see first hand anyone who was punished by Google for things said outside the workplace? And did this extend to topics other than Google?

Seconded. Could you elaborate on how "being a trouble maker" affected calibration scores?

Re: Don't Trust Google (2002)

#29

Post by scoot »

nick_riviera wrote: post #24
mynameisvlad wrote: post #22

Earlier quoted context omitted.

Google doesn't specifically know you stopped at Pizza Hut or the BP garage. It uses coarse location (GPS being too taxing on battery to get constant location updates) which has an accuracy of 100+ meters. So it would know you stopped somewhere around the Pizza Hut, but not specifically that you went inside. Could have just as easily gone to the Tesco next door.

Actually yes it does know I stopped at both of those location. 1. The BP garage has a wireless AP with a known SSID to them. Location services uses this, the cell network and the GPS to build an accurate location. There are no other wireless networks at that location. I was at this location for around 10 minutes. Google location history shows this location on the mark. 2. They know that I was at the Pizza hut because…

Besides which, consumer GPS has a resolution down to 3M, 95th percentile - not 100M [1].

[1] http://www.gps.gov/technical/ps/2008-SPS-performance-standar...

Re: Don't Trust Google (2002)

#30

Post by mynameisvlad »

nick_riviera wrote: post #24
mynameisvlad wrote: post #22

Earlier quoted context omitted.

Google doesn't specifically know you stopped at Pizza Hut or the BP garage. It uses coarse location (GPS being too taxing on battery to get constant location updates) which has an accuracy of 100+ meters. So it would know you stopped somewhere around the Pizza Hut, but not specifically that you went inside. Could have just as easily gone to the Tesco next door.

Actually yes it does know I stopped at both of those location. 1. The BP garage has a wireless AP with a known SSID to them. Location services uses this, the cell network and the GPS to build an accurate location. There are no other wireless networks at that location. I was at this location for around 10 minutes. Google location history shows this location on the mark. 2. They know that I was at the Pizza hut because…

You didn't explain how they knew about the PH/BP, which makes people assume it was due to location data (like your first example).

However, as for comment #1, Google doesn't use the SSIDs to determine locations on mobile, they use the [MAC Addresses](http://static.googleusercontent.com/media/www.google.com/en/...). They do have a database that also includes SSIDs for these MAC addresses, but there could be a dozen SSIDs around you at any given time, and you could be at any one of those businesses. I have already explained that Google uses coarse location to determine your location, not fine because of battery issues (and you can check this in Android, it's listed under the lower-power location use). Coarse doesn't use GPS, and its accuracy is generally 100+ meters because of this. It will know the general area you were at (and, no, it won't know you were there for 10 minutes, as it only picks up your location at fixed intervals which IIRC is above 10 minutes), but it won't know, specifically, where you were in that area.

I will concede on the Pizza Hut location, but it would have only known you were there because of the review, not because of the location (for the same reason as above).

Post reply on HN