Emptying out the mailbag & clearing the history #21

Posted on September 11, 2017

(Last Updated On: September 11, 2017)

Other mailbags can be found hereKeep in mind a lot of this is email conversations, comment replies, or some random interesting things I’ve found. By their nature they are not as thorough or complete as a post on one topic.

Here’s what’s covered in this installment:

Our monthly reminder of what Facebook thinks of us

First, Facebook bought a data security app…so they could spy on customers who use that app. That is, if you have the security app, because you’re likely to allow the security app more access to your phone, you know, for security, Facebook takes advantage of this trust and monitors everything you do on your phone. They knew consumers wouldn’t allow Facebook that access, so they bought another app instead.

For instance, they want to see how much you use SnapChat, a competitor? They monitor the millions of people who have that app and also use SnapChat. This is how Facebook knows to copy certain features. Or why Instagram -owned by Facebook- looks so much like SnapChat now.

Facebook’s Onavo Gives Social-Media Firm Inside Peek at Rivals’ Users

Next, in my favorite Facebook news of all-time,

“At the age of 15-24, Facebook claims that you have a potential reach of 1.5 million people. This applies to the number of Facebook users living in Sweden, according to Facebook Ads Manager.

The problem is that, according to Statistics Sweden, SCB, there are only barely 1.17 million people in the age group registered in Sweden”

New digitbluff from Facebook? Has spelled out 330,000 Swedish youths

Reuters apparently followed up on that,

“Facebook’s ad-buying website tells advertisers that the world’s largest social network has a potential reach of 41 million 18 to 24 year olds in the United States, whereas U.S. census data shows that last year there were 31 million people living in the country between these ages”

Facebook digital ads figures differ from census data

Other countries have been examined as well. In every case, Facebook overestimates.

As if that wasn’t bad enough!

Russian firm tied to pro-Kremlin propaganda advertised on Facebook during election

THAT’S ALL FROM THE LAST MONTH! At this point, who knows if Facebook even makes any money. Maybe they’re the modern Enron.

Just remember, this how Mark Zuckerberg started Facebook:

(Those are real. He admitted to it. They were used in court against him.) A lot of people say “you have to be on Facebook” business wise these days. Personally, that’s equivalent of telling me “you need to do business with crappy people.” Doesn’t sound like a solid strategy.

Oh, and this is a guy quietly gearing up for a potential presidential run. He’s currently touring the country. If he actually runs, remember what he thinks of you.

And remember his platform is part of what’s destroyed American politics. (And that polarization was relatively stable until 2004, when his platform came online.)

-> None of the original founders are still at Facebook but Fuckerberg. My experience has been whenever most or all of the co-founders leave, the remaining one is a real dick.


Jawbone is out of business

Jawbone was a wearable company. Back in April I wrote,

“Wearables are already going through their downturn. Pebble sold to FitBit, who then laid-off 6% of its employees. There is about to be a graveyard of wearable companies.”

“Another one bites the dust.”

Pessimism regarding upcoming artificially intelligent personal trainers (part 9)

Speaking of AI issues

“There were two causes for the algorithm being correct…First, at the hospitals hosting this study, patients with a history of asthma who presented with pneumonia were usually admitted directly to intensive care units to prevent complications; this led to patients with pneumonia and asthma having better outcomes than patients diagnosed with pneumonia and without a history of asthma, with an approximately 50% mortality risk reduction.”

Because the algorithm couldn’t contextualize…

“thus the algorithm “correctly misinterpreted” the presence of asthma as a protective variable.”

Unintended Consequences of Machine Learning in Medicine

To be clear, this algorithm decided having asthma was better than not, if you had pneumonia. In reality, getting intensive care was better than not.

We’re entering a very dark cave of unintended consequences with this. Silicon Valley has a history of thinking healthcare is way easier than it is. But it’s infinitely harder than computer science. The amount of random associations we’re going to find with e.g. genetics is scary. The health world is pretty damn aware correlation does not equal causation. Yet the basis of these algorithms is correlation.

Right now, humans can reason “Uh, giving someone asthma is not going to help their pneumonia.” But when it comes to some random part of the genome, we’ll have no idea.

This is not new to the health world. A few years ago I wrote,

“For a long time SIDS had no understandable cause. Eventually, in the early 1900s, we delved into this disease and examined a bunch of SIDS cadavers. This was a great opportunity as cadavers represent one of our best diagnostic tools, especially a hundred years ago. We can look inside the body and go, “This group, who died of this disease, seems to have this commonality.”

Upon opening the SIDS cadavers we found an obvious congruity. All the SIDS babies had enlarged thymus glands. Understanding their death became easy: Too enlarged of a thymus gland could asphyxiate the infant, causing quick death.

We immediately formed a solution: Irradiate the thymus gland, causing it to atrophy, and boom, no risk of suffocation.

The result? Ten thousand people were killed, and many more harmed.”

One of the dirty secrets of medicine is we have no clue how many people we’ve killed / hurt relative to how many we’ve saved / helped. In aggregate, if you simply took away basic hygiene and vaccines, I wouldn’t be surprised if we’re net killers. It’s that hard, and medicine was brutal until only recently.

Even in modernity, give people hospice care rather than aggressive treatment for lung cancer? They live 25% longer. (Detailed in Atul Gawande’s Being Mortal.)

The SIDS cadavers all had large thymus glands because they were from poor people. Stress causes hypertrophy of the thymus, and poor people are chronically stressed.

-> One insight you can’t avoid when looking deeply enough into human health is where you’re born, who your parents are, irreversibly matters.

Tattoos inhibit sweat

Nice idea for an experiment-

“follow-up studies could be of significant interest to athletes with tattoos who train in the sun and are at risk of overheating.

Others who may also be impacted include firefighters and soldiers.

“Military personnel have a high incidence of tattoos and are often exercising vigorously with heavy equipment in a hot environment,” Luetkemeier said.”

Do Tattoos Affect Sweat, Inhibit Exercise?

Supinating or pronating during dorsiflexion?

Someone asked about this in an email. My response,

“If the aim is purely dorsiflexion, then I wouldn’t be supinating or pronating. I’d keep the foot flat on the ground.

That said, more people overpronate than excessively supinate. So adding some supination can be a nice way to offset potential muscular imbalances.”

Talked about more in The best damn IT band stretch ever.

Exercise changes how *firm* your brain is

This is such a cool study.

Firmer, fitter frame linked to firmer, fitter brain

Think of a tempurpedic / memory foam pillow. Initially, it’s very responsive to pressure. You push on it, release, it comes back quickly. After it gets worn out, it doesn’t come back as quickly. It’s mushier. This study found this happens to the brain!

Solid Jeff Bezos business advice

My Advice To Anyone Starting A Business Is To Remember That Someday I Will Crush You

“In the early days of your startup, you may have to do a little bit of everything: design, accounting, marketing, legal, even cleaning up the office. That takes hard work, perseverance, and—I can’t emphasize this enough—the understanding that once you’ve succeeded, I will begin systematically choking off your revenue streams. It’s all part of the process, as integral as drafting your business plan, scaling up your company, and coming to terms with the fact that I will ultimately force you to take out a second mortgage just to make payroll.

Beyond that, you need to be willing to pivot. Too many times I’ve seen startups get mired in ideas that were clearly unfeasible from the get-go. And even if your core business is sound, you may need to reboot your product several times before you create something so truly innovative that I move into the market myself and proceed to undercut you at every turn, putting such enormous strain on you and your partners that your organization descends into bitter infighting.

Now, suppose your startup really catches on and you build up an impressive customer base. That’s great—but there are still many trials ahead. Be prepared, among other things, for my company to duplicate your product or service and sell it at prices you can never compete with, all the while turning your board members against you one by one and, eventually, buying your company for less than it’s worth just so we can shut it down. Nonetheless, congratulations! You’ve made it further than most.”

Subscribe to Blog via Email

Enter your email address to subscribe to this blog and receive notifications of new posts by email.