5 ms·

Michael Lewis (of the Moneyball and The Big Short fame) talks about Cambridge Analytica in his podcast https://atrpodcast.com/episodes/the-alex-kogan-experience

Michael Lewis (of the Moneyball and The Big Short fame) talks about Cambridge Analytica in his podcast https://atrpodcast.com/episodes/the-alex-kogan-experience-s1!d20f3 https://atrpodcast.com/episodes/the-alex-kogan-experience-s1.... Long story short: the CA scandal was way overblown. They could barely identify 1% of the population with all that data.

Pfhreak 7y ago
> They could barely identify 1% of the population with all that data. 1% of what population? 1% of Facebook users? 1% of the American population? Both of those are pretty staggeringly large numbers of people...
- dirtybird04 7y ago
  1% of the testing population. Yes, you are rightly identifying a staggering large number of people, but the flipside is you're misidentifying an even larger number of people. Which is why the scandal was way overblown.
- weego 7y ago
  By their own admission in pitch decks the CA system was no better resolving an individual accurately than existing ways. The main architect of the system's only previous experience was trying (and apparently failing) to model fashion forecasting. To say then CA situation was overblown in terms of scope and insight is an understatement imo. The irony is their sales pitch was so good it caused an international incident.
  - antod 7y ago
    > The irony is their sales pitch was so good it caused an international incident. As a non American, I find the "it's not so bad, it wasn't very effective" arguments I'm seeing in these comments a bit strange. I personally don't really care about the effectiveness of their analysis versus them (and others) having access to and collecting all that data in the first place. Once all the data is out there and the lack of privacy is normalised/accepted there will be all kinds of other attempts to use it in the future. Americans on the whole seem quite nonchalant about their privacy.
    - dirtybird04 7y ago
      >Americans on the whole seem quite nonchalant about their privacy I disagree, and let me elaborate. The CA scandal started two different conversations, which people get mixed up a lot. The primary one was about how shit Facebook's privacy settings are, and how all your data has been easily available for a few bucks. This stirred up a LOT of privacy concerns in the US, causing a whole bunch of people to quit Facebook. Zuckerberg was on the cover of magazines, for all the wrong reason. That conversation is still going on. The secondary conversation was ok now that we know the data was out there, how effective was it? Did it help swing the election? This documentary wants you to convince that it is true, that CA was the reason why Trump got elected. And as the article and Michael Lewis' podcast lays out, that's just not the case. tldr: We're talking about two different things. Talking about secondary concerns does not negate the primary ones.
      - antod 7y ago
        Yeah fair point. I suppose for non Americans like myself, we would naturally put far less weight on the secondary discussion. Also I hadn't watched the documentary in question, so wasn't referring to its arguments directly. I do have some vague hopes that California's upcoming privacy regulations (the ones that from the outside look a little GDPR like) will prod the tech giants enough to take this shit seriously. And I hope you're right about what this has stirred up in the US. I'm hoping that my view (ie the apparent nonchalance) from the outside is inaccurate. For the rest of the world, we rely on American sentiment to influence these companies.
TheHypnotist 7y ago
That sample size is better than most or all main stream polls.
lern_too_spel 7y ago
1% match is more than enough to build lookalike models in the Facebook ad platform.

Pfhreak 7y ago

> They could barely identify 1% of the population with all that data. 1% of what population? 1% of Facebook users? 1% of the American population? Both of those are pretty staggeringly large numbers of people...

dirtybird04 7y ago
1% of the testing population. Yes, you are rightly identifying a staggering large number of people, but the flipside is you're misidentifying an even larger number of people. Which is why the scandal was way overblown.
weego 7y ago
By their own admission in pitch decks the CA system was no better resolving an individual accurately than existing ways. The main architect of the system's only previous experience was trying (and apparently failing) to model fashion forecasting. To say then CA situation was overblown in terms of scope and insight is an understatement imo. The irony is their sales pitch was so good it caused an international incident.
- antod 7y ago
  > The irony is their sales pitch was so good it caused an international incident. As a non American, I find the "it's not so bad, it wasn't very effective" arguments I'm seeing in these comments a bit strange. I personally don't really care about the effectiveness of their analysis versus them (and others) having access to and collecting all that data in the first place. Once all the data is out there and the lack of privacy is normalised/accepted there will be all kinds of other attempts to use it in the future. Americans on the whole seem quite nonchalant about their privacy.
  - dirtybird04 7y ago
    >Americans on the whole seem quite nonchalant about their privacy I disagree, and let me elaborate. The CA scandal started two different conversations, which people get mixed up a lot. The primary one was about how shit Facebook's privacy settings are, and how all your data has been easily available for a few bucks. This stirred up a LOT of privacy concerns in the US, causing a whole bunch of people to quit Facebook. Zuckerberg was on the cover of magazines, for all the wrong reason. That conversation is still going on. The secondary conversation was ok now that we know the data was out there, how effective was it? Did it help swing the election? This documentary wants you to convince that it is true, that CA was the reason why Trump got elected. And as the article and Michael Lewis' podcast lays out, that's just not the case. tldr: We're talking about two different things. Talking about secondary concerns does not negate the primary ones.
    - antod 7y ago
      Yeah fair point. I suppose for non Americans like myself, we would naturally put far less weight on the secondary discussion. Also I hadn't watched the documentary in question, so wasn't referring to its arguments directly. I do have some vague hopes that California's upcoming privacy regulations (the ones that from the outside look a little GDPR like) will prod the tech giants enough to take this shit seriously. And I hope you're right about what this has stirred up in the US. I'm hoping that my view (ie the apparent nonchalance) from the outside is inaccurate. For the rest of the world, we rely on American sentiment to influence these companies.

TheHypnotist 7y ago

That sample size is better than most or all main stream polls.

lern_too_spel 7y ago

1% match is more than enough to build lookalike models in the Facebook ad platform.