[lbo-talk] A feast of Google math

Eleutherios eleutherios.rizooto at gmail.com
Sun Sep 30 15:14:45 PDT 2012


Cool link, Joanna.

I long ago became deeply distrustful of Google and avoid it as much as possible. This means using Do Not Track setting on browsers, controlling cookies, running Google/Fb/etc. in separate clean browser environments, sometimes using Tor https://www.torproject.org/ (provides anonymity, no IP address or location tracking), and https https://www.eff.org/https-everywhere (provides encryption for transactions on websites that support it, not directly an averting Google/Fb thing but provides privacy from third parties).

It's not fully up to par yet, but it has improved enough and in some ways now outperforms Google that I mostly use https://duckduckgo.com/ . It's well worth trying. Occasionally, I revert to Google, but that's mainly for Image, Scholar, and such specialty searches. Much of the Advanced Search capability is now at DDG, it's results have become good, are directed at finding info and are nicely organized, and it doesn't track. For certain types of searches or calculations, I also use http://www.wolframalpha.com/ --which is quite nifty for the uninitiated (especially those liking to play with numbers, data, graphs, sciences/maths).

On 29-Sep-12 16:52, Chuck Grimes wrote:
> Everything you ever wanted to know about the how of Google
>>
>> http://static.googleusercontent.com/external_content/untrusted_dlcp/research.google.com/en/us/pubs/archive/38331.pdf
>>
>>
>> Joanna
> --------------
>
> Thanks.
>
> Comments. I use Google as a spell checker, since my text processor
> (Emacs) doesn't have such an animal.
>
> I have become deeply suspicious of google over the years because the
> first few offerings in a search are nearly always heavy commercial
> links, so I go down a few and then click. Another technique is to
> start with some obvious word or two, and if that doesn't work, I click
> on something to some other words that might get closer like a specific
> name. Getting the right video is particularly difficult because I am
> often looking for posted panels, lectures, and other material from the
> left. And these often have a thousand or less views.
>
> I figured they used sophisticated stats, but I had no idea they depend
> so heavily on linear algebra. Subspaces, nodes, wow.
>
> The face recognition engine is scary. In fact thinking back on the
> whole presentation all of it has applications for the national
> surveillance state. Surveillance is one of those words I can almost
> never spell without looking.
>
> CG



More information about the lbo-talk mailing list