08-05-2020, 12:52 PM
** 05-Aug-2020 World View: 7/30 testimony of the CEOs of Apple, Google, Amazon and Facebook
I didn't think that the CEOs said anything brilliant. I just said
that they played on the politicians' ignorance and stupidity, which is
quite a different thing, and trivially easy. Bezos told his
ridiculous weepy story about his dad Miguel. Pichai stumbled over
words and gave an impression of being overwhelmed. Both of them
avoided answering any real questions. In both cases, the response and
demeanor were well-practiced, and designed to disarm the posturing
politicians, who are too dumb to know what was going on anyway. And
it worked.
Let's take a couple of examples of questions that might have
embarrassed the CEOs, if they had been asked.
First, I assume you read something about the recent Twitter hack. A
malicious actor bribed or tricked or extorted a Twitter employee and
got control of a few dozen accounts. All the hacker did was send out
hundreds of tweets in order to collect a few hundred thousand dollars,
but as some media reports pointed out, the attack could be practicing
for a serious attack in November to affect the election.
Most people have no clue about the following: If you have an online
account with anyone -- a store, a bank, a social media service -- then
the company you're dealing with has an IT department of 5, 10, 20,
100, 500 employees who are responsible for updating the software or
providing customer support. All of those people have access to the
entire database, or they could not do their jobs, and so they all have
access to your data.
Companies can take various kinds of steps to mitigate the exposure.
Typically, your account password is encrypted, so no one in IT can get
it. Sometimes your social security number is encrypted, and that will
protect it from external hackers, but internal IT people will be able
to access the decryption functions, so those numbers are usually
unprotected internally. Sometimes your credit card numbers are stored
on another server, so that only select IT people can access them, but
they're still partially exposed. But basically, if you provide any
data to any online service or business, you should assume that 10-2000
IT people, depending on the size of the company, have free access to
your data, and to your entire account.
So the four CEOs must have been shaking in their boots, out of fear
that they might be grilled about this. Because if all of Twitter's
accounts can be stolen with the help of a crooked or rogue or idiot IT
employee (of which any company contains many, many), then Apple,
Amazon, etc., have exactly the same problem, and all of them can be
hacked in the same way, giving some malicious actor control of
multiple accounts, for at least a period of time.
Imagine the disaster that such a malicious actor could cause! And
Twitter has shown the way -- there's no defense against it. There are
plenty of questions that the four CEOs might have been asked about
this situation, but the politicians, who barely know how to turn on a
computer, just ignore the problem, and go on posturing about other
things.
The second embarrassing area would be those "magic AI" algorithms that
check for "hate speech." The obvious intent is that all hate speech
would be suppressed, and anything from Trump supporters would be
classified as "hate speech," and so therefore anything supporting
Trump should be suppressed.
The companies just say, "Oh, those decisions are made by algorithms --
AI algorithms -- which are non-partisan."
And the idiot politicians just accept that, as if there were some
magic involved. Actually, algorithms like that would be rules-driven,
and programmers would write the rules. As stupid as politicians are,
understanding a little bit about those rules is not beyond the
capabilities of those few who can still do fourth grade math.
The rules are confidential, of course, but we can speculate on how
some of them work. Let's suppose a tweet contains the text, "Make
America great." That alone wouldn't be enough to classify it as hate
speech, but it would add points in some sort of point system. If a
tweet contains another "racist" phrase like that, then there might be
enough points for the "non-partisan" algorithms to decide that the
tweet is hate speech.
So here are the questions for the four CEOs: We know from various blog
posts that everyone at Google, etc., is far-left pro-Democrat. So all
those hate speech rules are written by left-wingers, right? Do you
have any moderate or conservative people writing those algorithms? Of
course not. And isn't it true that even if there were a conservative
in the group, he would be sidelined by the others, and his work would
all be deleted during code reviews by all the left-wingers on the
team?
If you hired only whites to do programming, you would be accused of
biasing your algorithms to whites, and you would have to hire blacks.
If you hired only men to do programming, you would be accused of
biasing your algorithms to men, and you would have to hire women.
Isn't the same thing true that if you hire only left-wingers to do
programming, then you are biasing your algorthms to Democrats? And
isn't that against the law and your own supposed principles? Given
the importance of evaluation "hate speech," don't you think that you
have an obligation to hire a few conservatives, instead of just all
left-winger Democrats?
Higgenbotham Wrote:> Speaking of competence, I watched the testimony of the CEOs of
> Apple, Google, Amazon and Facebook yesterday.
> These guys are idiots, especially the one from Google. But what
> especially surprised me was that Bezos couldn't put together a
> better opening statement.
John Wrote:> ** 02-Aug-2020 World View: Idiots
> I actually disagree with this point of view. It's the politicians
> who were idiots because they barely know how to turn on a computer
> and have no idea what's going on. The CEOs were well prepared,
> and played on the politicians' ignorance and stupidity.
Higgenbotham Wrote:> Could you be more specific? What did the CEOs say that was
> brilliant? Because I saw not a glimmer.
> As far as being well prepared, all I remember any of them saying
> is I don't know. Although I would admit that much of what they
> said I don't know to I would not have necessarily expected them to
> know. But something beyond I don't know would be expected, such
> as how they go about those things in general. I do remember
> Zuckerberg essentially saying that, yes, it is Facebook's general
> practice to copy competitors.
I didn't think that the CEOs said anything brilliant. I just said
that they played on the politicians' ignorance and stupidity, which is
quite a different thing, and trivially easy. Bezos told his
ridiculous weepy story about his dad Miguel. Pichai stumbled over
words and gave an impression of being overwhelmed. Both of them
avoided answering any real questions. In both cases, the response and
demeanor were well-practiced, and designed to disarm the posturing
politicians, who are too dumb to know what was going on anyway. And
it worked.
Let's take a couple of examples of questions that might have
embarrassed the CEOs, if they had been asked.
First, I assume you read something about the recent Twitter hack. A
malicious actor bribed or tricked or extorted a Twitter employee and
got control of a few dozen accounts. All the hacker did was send out
hundreds of tweets in order to collect a few hundred thousand dollars,
but as some media reports pointed out, the attack could be practicing
for a serious attack in November to affect the election.
Most people have no clue about the following: If you have an online
account with anyone -- a store, a bank, a social media service -- then
the company you're dealing with has an IT department of 5, 10, 20,
100, 500 employees who are responsible for updating the software or
providing customer support. All of those people have access to the
entire database, or they could not do their jobs, and so they all have
access to your data.
Companies can take various kinds of steps to mitigate the exposure.
Typically, your account password is encrypted, so no one in IT can get
it. Sometimes your social security number is encrypted, and that will
protect it from external hackers, but internal IT people will be able
to access the decryption functions, so those numbers are usually
unprotected internally. Sometimes your credit card numbers are stored
on another server, so that only select IT people can access them, but
they're still partially exposed. But basically, if you provide any
data to any online service or business, you should assume that 10-2000
IT people, depending on the size of the company, have free access to
your data, and to your entire account.
So the four CEOs must have been shaking in their boots, out of fear
that they might be grilled about this. Because if all of Twitter's
accounts can be stolen with the help of a crooked or rogue or idiot IT
employee (of which any company contains many, many), then Apple,
Amazon, etc., have exactly the same problem, and all of them can be
hacked in the same way, giving some malicious actor control of
multiple accounts, for at least a period of time.
Imagine the disaster that such a malicious actor could cause! And
Twitter has shown the way -- there's no defense against it. There are
plenty of questions that the four CEOs might have been asked about
this situation, but the politicians, who barely know how to turn on a
computer, just ignore the problem, and go on posturing about other
things.
The second embarrassing area would be those "magic AI" algorithms that
check for "hate speech." The obvious intent is that all hate speech
would be suppressed, and anything from Trump supporters would be
classified as "hate speech," and so therefore anything supporting
Trump should be suppressed.
The companies just say, "Oh, those decisions are made by algorithms --
AI algorithms -- which are non-partisan."
And the idiot politicians just accept that, as if there were some
magic involved. Actually, algorithms like that would be rules-driven,
and programmers would write the rules. As stupid as politicians are,
understanding a little bit about those rules is not beyond the
capabilities of those few who can still do fourth grade math.
The rules are confidential, of course, but we can speculate on how
some of them work. Let's suppose a tweet contains the text, "Make
America great." That alone wouldn't be enough to classify it as hate
speech, but it would add points in some sort of point system. If a
tweet contains another "racist" phrase like that, then there might be
enough points for the "non-partisan" algorithms to decide that the
tweet is hate speech.
So here are the questions for the four CEOs: We know from various blog
posts that everyone at Google, etc., is far-left pro-Democrat. So all
those hate speech rules are written by left-wingers, right? Do you
have any moderate or conservative people writing those algorithms? Of
course not. And isn't it true that even if there were a conservative
in the group, he would be sidelined by the others, and his work would
all be deleted during code reviews by all the left-wingers on the
team?
If you hired only whites to do programming, you would be accused of
biasing your algorithms to whites, and you would have to hire blacks.
If you hired only men to do programming, you would be accused of
biasing your algorithms to men, and you would have to hire women.
Isn't the same thing true that if you hire only left-wingers to do
programming, then you are biasing your algorthms to Democrats? And
isn't that against the law and your own supposed principles? Given
the importance of evaluation "hate speech," don't you think that you
have an obligation to hire a few conservatives, instead of just all
left-winger Democrats?
