Quick Thoughts: Truth, Overwork-from-Home, Benford's Law and more

(This is my weekly list of quick thoughts based on the most interesting topics I found in my feed.)

Quote of the week:

The purpose of free speech isn't to prevent suppression of the truth. Nobody is interested in the truth anyway. Free speech exists so that you can fight the establishment's lies with your own lies.

—paraphrased from Bryan Caplan

Fake News

Why is there so much fake news? Because that’s what people want.

Pankaj Jain @pj77in

Another Global pandemic - #FakeNews runs on this.

And how should you convince a believer that their belief is untrue? Providing facts and data will often backfire. The backfire effect happens because a direct contradiction of a person’s belief results in them going on the defensive, which triggers their fight-or-freeze response.

Krish Ashok @krishashok

Listening to the @notsmartblog gave me an alternative approach. Responding with a “why do you believe this? How do you know this is verified?” forces a different part of the brain to deal with it, and is more likely to result in a less defensive response

A Truth vs. The Truth

Sloww @SlowwCo

Truth & Perspectives:

Of course, this is saying the same thing as the poem about the six blind men who went to see the elephant, but it beautifully captures not only the fact that people have incomplete information, but also that there can be multiple levels of it. Someone should do a galaxy brain meme with this tweet.

This is related to what I said in the update 2 weeks back: if you think a top business or political leader is dumb, most likely you are mistaken, and you aren’t aware of all the information or all the constraints that they have. aka “If Kabil Sibal is such an idiot, how come you’re not the prime minister”.

AI mistakes vs non-AI mistakes

Vladimir Haltakov @haltakov

This is Karma. Karma is not a machine learning classifier 🐕‍🦺 Karma is a real dog trained to detect drugs. However, he would fail the simplest tests we apply in ML... Let me take you through this story from the eyes of an ML engineer. reason.com/2021/05/13/the… Thread 🧵

reason.comThe Police Dog Who Cried Drugs at Every Traffic StopCops laugh about “probable cause on four legs” but the damage to innocent lives is real.

Interesting thread by a machine-learning engineer pointing out how badly “drug sniffing dogs” would fail the same tests that are used to validate machine-learning algorithms.

Along similar lines, see this quote by Daniel Kahneman (the economist/psychologist who is one of the founders of the field of behavioural economics, and 2002 Nobel memorial prize winner) related to the difference in how people view AI mistakes vs similar mistakes in natural processes. This also explains why

Source

As more and more of the world is eaten by AI, we need to start thinking carefully about such aspects of AI.

Lessons

Here’s the first part of a tweet. What lessons do you think can be drawn from this event?

Think about that a bit before reading the full tweet below:

Heather Thompson Day @HeatherTDay

My sophomore year of college I got a C+ in Geology My grade was a 79.6 Professor refused to round up I sat in his office, told him a C would destroy my GPA He wouldn’t budge Now I’m a professor & I get to round up Sometimes what ppl teach us Is who we don’t want to be

But, that’s not the only lesson to be learnt from this. Here’s another takeaway:

Kevin J.S. Zollman @KevinZollman

I think there are many lessons to draw from this story. Let me point out a different one than the author does: she got a C+. It ruined her GPA. She became a professor anyway. If your professor doesn't round up, it's too bad. But you can succeed anyway.

Heather Thompson Day @HeatherTDay

My sophomore year of college I got a C+ in Geology My grade was a 79.6 Professor refused to round up I sat in his office, told him a C would destroy my GPA He wouldn’t budge Now I’m a professor & I get to round up Sometimes what ppl teach us Is who we don’t want to be

Overwork

Most people are working more time working now than they were spending on work + commute during the lockdown. More than one-third of them are working “much more”. This is not a good situation.

Why? Here’s a report which says that long working hours cause strokes and heart-attacks resulting in 745,000 deaths per year.

If you are a decision-maker at a company, do you think this is a problem? If not, why not? If yes, what are you doing about it?

Did You Know: Benford’s Law

Paul Graham @paulg

Which essays people read over the past 365 days.

Paul Graham tweeted this. And Paul’s essays are must-reads for everyone. But that’s not the point of this section. This follow-up is more interesting:

dileep kumar p @dileep31

I can't help but notice Benford's law in these numbers. It says that if you look at the frequency of leading digit among many naturally occuring numbers, smaller digits much frequent than bigger digits. Here, of the total 50 posts, 24 have no:of readers starting with digit 1.

Paul Graham @paulg

Which essays people read over the past 365 days. https://t.co/yGtYbSv5Cv

Benford’s Law claims that in many real-life data-sets, if you look at the first digit of the numbers involved, 1 appears far more often than 2, which appears more often than 3 and so on.

The distribution of first digits, according to Benford's law. Each bar represents a digit, and the height of the bar is the percentage of numbers that start with that digit.

This law has been used to detect instances of fraud in elections (there is a possibility of fraud if the voter counts don’t follow Benford’s law) or in data sets published in scientific papers, and even fraudulent data in Government numbers.

Read the Wikipedia article, it is quite interesting.

Subscribe and Share

Please share this post with others who might find it interesting

If you came to this page via a link someone sent you, you can subscribe to get the latest updates via email.