Daniel Miessler's Blog, page 76

So You Want to Start a Blog…

I am often asked the best way to start a blog. My answer hasn’t varied much over the years, so I’m finally going to capture my answer as a series of concise admonitions.

I started this website in 1999, which, at the time of writing, makes 20 years blogging. In that time I’ve probably used every major blog platform that’s existed, which I’d guess has been no less than 20. I’ve done everything from pure HTML, Ruby, Node, Python, self-hosting, multiple third-party providers, all the way to my current setup.

Anyway, all that means is I’ve made lots of mistakes in these last two decades. And they’re mistakes that I can hopefully keep you from repeating.

First principles

No matter what you do, pay attention to these.

1. Your domain is your identity; pick it carefully

You need to decide if it’s a site about you, or a site about something specific and temporary.

I mostly recommend firstnamelastname.com, like this site because it gives you maximum freedom to explore stuff you care about. But if you have a specific project that you expect to have for at least 70 years, feel free to use something different. Err on the side of being too broad with your domain rather than too narrow.

Domain name changes are damaging to site reputation, so try to avoid them.

You don’t want to start a blog at 25 called blackberrytips.com, find out in 15 years that you’re actually into Buddhism, and end up writing about meditation on a domain with Blackberry in the name.

2. Use a solid domain registrar, like Google or Cloudflare

Be sure to lock these down with two-factor authentication.

I personally would stay away from groups that talk about how cheap they are, or who use other silly advertising to get you to sign up. Again, your domain is everything, so protect it.

Protect your domain’s DNS the same way, and consider using your domain provider as your DNS host as well.

3. Your URL Structure Matters a Lot

Spend a lot of time in the very beginning thinking about the structure of your site and the structure of your URLs. In general you want as few / slashes as possible in a URL, and I don’t recommend using the date as your leading structure.

Ideally you want maybe one level deep (like /blog, /articles, /research, etc), followed by 3-5 primary words that describe the post, e.g., /blog/so-you-want-start-blog. That’s what makes it easiest for Google to read your content.

What to Avoid (Don’ts)

These are the things to avoid.

Avoid Third-party Blogging Platforms

Every few years a new blogging platform comes along that is the slickest looking thing around. Blogger, Tumblr, Medium, Squarespace, etc—they all promise the world and end up going out of business eventually. Or worse, they sell out due to financial hardship or greed.

Either way you’re left doing a migration that takes tons of time and probably messes your site up in the process. Don’t use them.

What to do (Do’s)

Use Common and Popular Software

That doesn’t mean it has to be number one, but don’t pick number nine either. Going with one of the top 2 or 3 options means your platform will have maximum scrutiny and therefore faster patching if something goes wrong.

It also means it’s likely to last for a very long time.

Pick Software Without Lock-in

Make sure you pick a software package that has incentives that are truly aligned with independent content publishing. This is why you want to avoid the platforms mentioned above: they always (eventually) will try to do things in non-standard ways, make it difficult to migrate away from them, and/or will store your content in a proprietary format that doesn’t export well.

If You’re Not Technical, Consider Using a Service

It’s not a problem to use a hosted blogging service, you just want to make sure that it’s a pure one.

That means that it’s clean hosting of one of the software options above that’s popular and that doesn’t lock you in—as opposed to their own software on their own platform that doesn’ really export well.

Summary

Ok, so since I’ve just given advice so far, here’s my tangible recommendation.

Get your domain from Google or Cloudflare.
Get your DNS from them too.
Lock them down with 2FA.
Either build yourself a VPS or find a host for your blog.
If you’re self-hosting, I recommend AWS.
If you’re using a service I recommend FlyWheel or Pantheon.
For the blogging software itself, I recommend WordPress, Hugo, or Host—in that order.
Make sure you (or your provider) keeps your entire stack updated regularly.
The primary thing that will make your blog popular is consistent creation of content that’s at least 1,000 words in length. Everything else is highly secondary.
Don’t overthink it. Focus on the writing, not on the tech.

—

If you like my content, you can support it directly for less than a latte a month ($50/year) which also gets you the Unsupervised Learning podcast and newsletter every week instead of just twice a month.

So You Want to Start a Blog

I am often asked the best way to start a blog. My answer hasn’t varied much over the years, so I’m finally going to capture my answer as a series of concise admonitions.

I started this website in 1999, which, at the time of writing, makes 20 years blogging. In that time I’ve probably used every major blog platform that’s existed, which I’d guess has been no less than 20. I’ve done everything from pure HTML, Ruby, Node, Python, self-hosting, multiple third-party providers, all the way to my current setup.

Anyway, all that means is I’ve made lots of mistakes in these last two decades. And they’re mistakes that I can hopefully keep you from repeating.

First principles

No matter what you do, pay attention to these.

1. Your domain is your identity; pick it carefully

You need to decide if it’s a site about you, or a site about something specific and temporary.

I mostly recommend firstnamelastname.com, like this site because it gives you maximum freedom to explore stuff you care about. But if you have a specific project that you expect to have for at least 70 years, feel free to use something different. Err on the side of being too broad with your domain rather than too narrow.

Domain name changes are damaging to site reputation, so try to avoid them.

You don’t want to start a blog at 25 called blackberrytips.com, find out in 15 years that you’re actually into Buddhism, and end up writing about meditation on a domain with Blackberry in the name.

2. Use a solid domain registrar, like Google or Cloudflare

Be sure to lock these down with two-factor authentication.

I personally would stay away from groups that talk about how cheap they are, or who use other silly advertising to get you to sign up. Again, your domain is everything, so protect it.

Protect your domain’s DNS the same way, and consider using your domain provider as your DNS host as well.

3. Your URL Structure Matters a Lot

Spend a lot of time in the very beginning thinking about the structure of your site and the structure of your URLs. In general you want as few / slashes as possible in a URL, and I don’t recommend using the date as your leading structure.

Ideally you want maybe one level deep (like /blog, /articles, /research, etc), followed by 3-5 primary words that describe the post, e.g., /blog/so-you-want-start-blog. That’s what makes it easiest for Google to read your content.

What to Avoid (Don’ts)

These are the things to avoid.

Avoid Third-party Blogging Platforms

Every few years a new blogging platform comes along that is the slickest looking thing around. Blogger, Tumblr, Medium, Squarespace, etc—they all promise the world and end up going out of business eventually. Or worse, they sell out due to financial hardship or greed.

Either way you’re left doing a migration that takes tons of time and probably messes your site up in the process. Don’t use them.

What to do (Do’s)

Use Common and Popular Software

That doesn’t mean it has to be number one, but don’t pick number nine either. Going with one of the top 2 or 3 options means your platform will have maximum scrutiny and therefore faster patching if something goes wrong.

It also means it’s likely to last for a very long time.

Pick Software Without Lock-in

Make sure you pick a software package that has incentives that are truly aligned with independent content publishing. This is why you want to avoid the platforms mentioned above: they always (eventually) will try to do things in non-standard ways, make it difficult to migrate away from them, and/or will store your content in a proprietary format that doesn’t export well.

If You’re Not Technical, Consider Using a Service

It’s not a problem to use a hosted blogging service, you just want to make sure that it’s a pure one.

That means that it’s clean hosting of one of the software options above that’s popular and that doesn’t lock you in—as opposed to their own software on their own platform that doesn’ really export well.

Summary

Ok, so since I’ve just given advice so far, here’s my tangible recommendation.

Get your domain from Google or Cloudflare.
Get your DNS from them too.
Lock them down with 2FA.
Either build yourself a VPS or find a host for your blog.
If you’re self-hosting, I recommend AWS.
If you’re using a service I recommend FlyWheel or Pantheon.
For the blogging software itself, I recommend WordPress, Hugo, or Host—in that order.
Make sure you (or your provider) keeps your entire stack updated regularly.
The primary thing that will make your blog popular is consistent creation of content that’s at least 1,000 words in length. Everything else is highly secondary.
Don’t overthink it. Focus on the writing, not on the tech.

—

If you like my content, you can support it directly for less than a latte a month ($50/year) which also gets you the Unsupervised Learning podcast and newsletter every week instead of just twice a month.

Unsupervised Learning: No. 201

.errordiv { padding:10px; margin:10px; border: 1px solid #555555;color: #000000;background-color: #f8f8f8; width:500px; }#advanced_iframe {visibility:visible;opacity:1;}#ai-layer-div-advanced_iframe p {height:100%;margin:0;padding:0}

—

If you like my content, you can support it directly for less than a latte a month ($50/year) which also gets you the Unsupervised Learning podcast and newsletter every week instead of just twice a month.

How Many Major Companies Have Lost Their Leadership Position Due to a Breach?

I was just thinking about the biggest breaches we’ve had in history, from companies like Adobe, LinkedIn, Equifax, Marriott, Target, etc., and wondering how badly they’ve been affected long-term.

Stock price doesn’t tell the full story of whether something impacted a company.

I’m wondering one specific thing about these top companies with the biggest breaches: What percentage of companies that were top-three in their industry, were dropped out of their top-three spot as a result of their incident?

Here’s the list I’m looking at, which may not be perfect.

I removed Yahoo and eBay because they were dying due to market forces unrelated to security breaches, and OPM because it’s part of the government.

Adobe
LinkedIn
Equifax
Home Depot
Target
Marriott

By my count, every single one of these companies has maintained its industry-leader position years after the incident. So the answer to the question of “what percentage lost their leadership position?” seems to be a big fat:

0%

None. Out of these six—and I’m sure I’m forgetting some—they’ve all maintained their dominant position as if nothing ever happened.

There are of course many smaller companies—and especially startups—that had a bad incident early on and went out of business because of it. But that seems to be a case of chasing away investors more than the stock market or customers.

It’s just interesting to me that Adobe is still the market leader. Marriott is winning at the hotel Game of Thrones more than ever. Equifax is annoyed, I’m sure, but their position as an industry leader hasn’t been shaken as far as I can tell.

Target? Still #2 to Walmart in that space. Home Depot? Yep, still doing their thing and either #1 or #2. LinkedIn is just fine as well.

So what?

So why did I bother to notice this, or point it out?

It just seems really interesting to me that for top-N industry leaders, both stock price and competitive position seem immune in the long-term to even the largest breaches that we’ve seen.

That doesn’t mean it doesn’t cost them money. And effort. And the opportunity cost. So it’s not pleasant or desirable or cheap for this to happen.

But it also doesn’t seem to be an existential risk for top companies, which is a belief that many people still hold.

I think lots of CEOs and CISOs and security teams proceed every day under the assumption that a big breach could be the end of their entire company. And maybe that’s best. Maybe people become good at their jobs—regardless of what it is—by convincing themselves that it’s more important than the reality.

But I can’t help but be intrigued by disconnects like this, where the general opinion among practitioners is divorced from the actual case.

Curious what others think—both about my assessment of infosec’s collective opinion on the existential threat, and about the overall analysis of how much mega-breaches affect mega-companies.

—

If you like my content, you can support it directly for less than a latte a month ($50/year) which also gets you the Unsupervised Learning podcast and newsletter every week instead of just twice a month.

We’re Seeking Approval From Too Many People

The United States is less happy and more depressed than it’s been in a long time, and people are trying to figure out why.

The World Happiness Report shows us down for 3 years in a row, and we’re currently holding only the #19 spot in the world.

Many point immediately to personal mobile devices and social media, but I had an idea earlier tonight that I thought was worth capturing.

What if we’re less happy because the people we’re trying to please have changed?

Being a student of evolutionary psychology, I firmly believe that happiness is all about pleasing evolution. Or, more specifically, convincing evolution that we’re winning, which tells our body to send us happiness signals.

So the question is, what do you have to do today—vs. decades or centuries ago—to convince evolution that you’re winning?

I the US in like 1950 or so, one basically had to please your God, your family, and maybe your boss (who made up a distant third, if that). With religion you know you’re imperfect already, so you’re just striving to be a little better every day. And with your family, you’re trying to do be a good parent, spouse, etc.

This doesn’t take into account the many who were struggling with oppression and survival still.

For many it was a relatively stable situation. And if you go back even further the number of people who you had to impress shrinks even further.

Feedback times

In addition to the number of people one has to worry about getting signals of winning from, there’s also the factor of update cycles, i.e., how often are you getting new or widely changing signals from those that determine if you’re winning?

How often do you get changing signals from God about how you’re doing? From your family. From your close friends. Or from the community—which back in the day was just the local people in the town.

People have always had wins and losses on all those fronts on a regular basis, and sometimes you’re more popular or less popular with your kids or spouse or church. But in general, I don’t think one’s overall reputation among those groups changed too drastically, or too often.

I think it has to do with how often you put yourself out there to be judged, like a performer. If you’re a church pastor and a new pastor takes your flock, that’s a strong signal of denial and loss. But if you’re one of the flock you’re not broadcasting much, and thus don’t have as many opportunities to be rejected.

Social media makes people performers to the world

So it could be that the reason depression and suicide are rising so quickly is that the number of people we’re trying to please is rising, combined with the exponential growth in the opportunities to either be cheered or ostracised.

Social media is like vanilla extract, but for life.

These two things combine to massively speed up (and exaggerate) the messages to evolution that we’re either winning or losing at life.

Imagine that you’re a basic housewife or merchant living in the United States in the 1950s, and that you live in a small town in Missouri. You have a spouse, two kids, and a house that looks a lot like the others around it. Everyone goes to the same Church.

Occasionally you get a promotion or a demotion, or you have a new kid, or you start playing an instrument at the church. And these things do change how your family, and the church, and the people at work see you—but only very slightly.

Over time, you’re just you, and the amount of rank and popularity shift you experience in life is relatively small and gradual, when it does happen.

Compare that to someone today, living in a big city or a small town. You’re a stay-at-home parent, or you work at a local business, but you don’t really go to church.

You have a few local friends but you mostly talk to them online, along with the other hundreds of people you have collected from high school to your mid-thirties.

Your main activity is using social media. You’re constantly posting pictures of your food, or your kids, or the vacation you’re going to go on. And you’re constantly watching which of your hundreds of friends are liking which pictures. And how much they like them compared to how much they like other peoples’ content.

When you post lots of stuff, and like lots of stuff, and then take a break to do something at work, or with the kids, your brain can’t wait to get back on Facebook to see how you did. Turns out, you only got a couple of likes for the picture of your kid’s birthday, but the girls you hung out with the week before gave over a hundred likes to another mother who posted pics of her kid at the same time. Why do they like her more?

You suddenly feel a sense of loss and dread and rejection like nothing else.

Why?

Because evolution is constantly looking for signals that you’re winning or losing, and it sits ready to punish you or reward you accordingly.

Evolution is tirelessly keeping score.

If you put out all those story shares, and cute little posts, and likes of other people’s content—those just got entered into the Evolutionary Ledger as pings of success. And evolution listens closely to whether those pings come back or not.

When they do, you get a rush. You’re winning! And you just programmed your hopeless built-thousands-of-years-ago brain to immediately repeat those pings out to the world so you can get another rush.

Every time you pick up your phone and check social media, that’s what you’re doing. You’re checking to see if those pings came back—hoping for replies.

When they don’t come back—and you hear silence instead—you can feel Evolution staring at you disapprovingly. And the failure burns through you.

A simple solution

The great news is that this model—if it’s at least somewhat accurate—admits immediately of a solution: reduce the number and frequency of pings for approval. And when you do ask for them, make sure you’re asking groups that are stable, trusted, and that you actually care about.

Family, close friends (that you actually visit in person or talk to on the phone), and cherished colleagues.

People like that are 1) more likely to give you generally positive feedback, and 2) the feedback is likely to be quite stable and predictable. When you talk to your spouse or your best three friends, they’re not likely to tell you that they hate you one day, and tell you they love you the next. Whereas the internet will absolutely do that.

The easiest solution is to just massively reduce your number of success beacons you’re sending out in hopes of replies. That means largely getting off of social media, and my first recommendation there is to abandon Facebook.

Give evolution fewer inputs to track.

Second, take a very serious look at who you care about in life, i.e., who you should care about, and limit your pings just to them as much as possible. That means spending more time working on your own projects instead of worrying about what everyone on the internet is doing, and spending quality time with the people you care about.

In short, massively reduce your circle by cutting out the internet from your friend group. The internet is not a friend—it’s an amorphous blob that can either sprinkle pixie dust or depression on your life. Don’t treat it like a spouse or friend that you need the approval of.

And finally, if you are someone who creates things and shares them, do so in a healthy way. What does that mean? It means putting things out there without attaching a success ping to them, so that Evolution isn’t listening over your shoulder for a response.

Establish your goals as self-betterment, the love of your family, the love of your friends, and the improvement of the world. And then go about your craft—whatever that is.

When you make something, like a blog post about Evolution Success Pings for example, just put it out there and move on to the next thing. Consider all the work to be done—regardless of whether nobody notices or you get some great vibes.

Evolution’s hardware and firmware

What studying Evolutionary Biology has taught me is how much our bodies—and therefore brains and minds—are fish out of water. We weren’t built for modern society. We’re GEICO commercial neanderthals walking around in the world of Star Trek, and we’re doing a bad job of it.

Using a computer/hacking metaphor, Evolution is the very old firmware running on our brains, and we’re trying to run modern society software on top.

It’s not working.

That’s why I study this stuff. So I can see how bad our hardware, firmware, and software is. So I can hack it. And that’s precisely what this is—this recommendation to reduce your group size and the number of success pings. It’s a hack for the bad code running on all our brains right now.

Let me be even more direct.

Evolution is our bad firmware, and it’s tuned for signals of success and failure coming from our environment, which it then rewards with bliss or depression. In older societies, this worked fine because our groups were small and those inputs didn’t change often or vary that much.

Now that’s different.

Social Media is malware that’s been developed to massively increase the number and frequency of your success pings (seeking of approval), and it’s designed to interact directly with your firmware (Evolution), which in turn affects your hardware (mood, energy, hormones).

Everything you see, hear, and read is code running on your brain, and social media is malware.

It’s code designed to compel you to do things. To click things. To like things. To share things. And it works because it has a direct line to your firmware/hardware.

This is why everyone’s comparing social media to a drug, along with the requisite addiction metaphors. It’s because both social media and drugs are hacks that affect our underlying hardware.

That’s all fine. That’s the beauty of hacking: once you know how a system works, you can change the inputs to get what you want.

And in our case, that means controlling the code running on our firmware/hardware. Now that we know our tech stack is extremely vulnerable to this type of malware, the answer is simply to avoid it.

Summary


Drop Facebook as much as possible given family requirements.
If you post to Twitter/Instagram, do so without caring about the response.
Focus on your own projects, and use social media to syndicate, not to seek approval.
Use social media to find and interact with other like-minded creators/thinkers, not trolls.
Reconnect with your inner circle of family and friends, and spend time with them and talk to them via voice/video.

In short, ask yourself what projects and efforts you really care about, and focus on those things knowing that the only approval you really need is from your close friends, your family, and the small group of wonderful people you meet when sharing your work with the world.

The internet is not your friend, so stop treating it like one.

—

If you like my content, you can support it directly for less than a latte a month ($50/year) which also gets you the Unsupervised Learning podcast and newsletter every week instead of just twice a month.

Unsupervised Learning: No. 200 (Member Edition)

This is a Member-only episode. Members get the newsletter every week, and have access to the Member Portal with all existing Member content.

Non-members get every other episode.

Sign in

or…

—

If you like my content, you can support it directly for less than a latte a month ($50/year) which also gets you the Unsupervised Learning podcast and newsletter every week instead of just twice a month.

Unsupervised Learning: No. 199

.errordiv { padding:10px; margin:10px; border: 1px solid #555555;color: #000000;background-color: #f8f8f8; width:500px; }#advanced_iframe {visibility:visible;opacity:1;}#ai-layer-div-advanced_iframe p {height:100%;margin:0;padding:0}

—

If you like my content, you can support it directly for less than a latte a month ($50/year) which also gets you the Unsupervised Learning podcast and newsletter every week instead of just twice a month.

Unsupervised Learning: No. 198 (Member Edition)

This is a Member-only episode. Members get the newsletter every week, and have access to the Member Portal with all existing Member content.

Non-members get every other episode.

Sign in

or…

—

If you like my content, you can support it directly for less than a latte a month ($50/year) which also gets you the Unsupervised Learning podcast and newsletter every week instead of just twice a month.

Corporate Incentives Are Diverging From Employee and National Incentives

The 1950s were not as perfect as people with red hats like to think they were, but there was one dynamic that was fairly healthy: it used to be that three primary incentives in the U.S. were aligned:

Individual Families
Corporations
The Country

Corporations would hire workers, which they would pay enough to: live on, save for retirement, save for college, go on vacations, buy presents for Christmas, and maybe have enough left over for a movie or a family dinner at the restaurant. And somehow you’d also not be constantly stressed about healthcare.

And in turn, those thriving families would make up a thriving country, which was working generally together towards shared goals.

Corporations saw themselves as contributing to America. And people saw themselves as doing the same by contributing to companies.

We just grew apart

I’m sure there are hundreds of books on what caused the following changes, but here’s where we seem to have ended up.

Most corporations are functioning either like sovereign countries within the United States, or like supernovas that are designed to make as much money as possible for early employees before exploding.
The country is run by a combination of lobbyists for those companies, and self-serving politicians who function like their own corporations.
The average family is left to battle for its own resources. Most corporate jobs aren’t nearly as stable, don’t pay nearly as much relative to the full costs of raising a family, and they’re always relatively close to being fired due to an “innovation” at work that makes their work redundant.

In short, we’ve lost the mutually-supporting structure of the family helping the corporation, the corporation helping the country, and the country doing its best to strengthen families and businesses. I think that structure is a major part of what made the country strong. But things are just different now.

The problem isn’t that corporations became evil where they used to be good. No. They were “good” in the past only in that their incentives just happened to align with those of families and the country. They needed people, and people needed jobs. So it was a great fit.

But most companies’ mission has always been to make a profit for shareholders.

Well, now there are simply better ways to do that, i.e., by automating human jobs away and lobbying to get laws changed in their favor.

Same goal—different ways of doing it. And these new ways just happen to be really bad for human work, human meaning, and for democracy.

Breaking it down

And this leaves us with the following:

Families don’t have loyalty to companies or the country, because they don’t feel they’re advocating for them
Corporations don’t have loyalty to workers or to the country, because their primary goal is profit
The country doesn’t have any direction or morals because it’s run by people with short-term and selfish goals, similar to those in the corporate world

So the family is alone, the corporation has sold out, and the country has lost any identity or direction.

That’s a problem. It’s not a foundation on which to build a future.

It’s a foundation for revolution, basically. For unrest and then rebirth.

And that’s why I’m not sold on Steven Pinker’s optimism. It doesn’t matter how bad it used to be compared to what we have now.

What matters is how much worse it’s getting relative to how good we remember it—even if that positive memory is flawed.

If we want to return to a healthy state we need to decide who we are. As a country. As companies. As families. And as individuals.

—

If you like my content, you can support it directly for less than a latte a month ($50/year) which also gets you the Unsupervised Learning podcast and newsletter every week instead of just twice a month.

Job Losses from Automation Are Being Actively Engineered and Funded

Many people seem to believe that there’s some sort of gentle pressure towards automation, and a counterbalancing gentle pressure towards job creation—and these two things live in happy equilibrium.

That’s a bad model for how the world works, especially today where most businesses are focused more than ever on maximizing margins.

Our move towards automation is active, deliberate, and extremely well-funded.

We’re not gracefully moving towards automation. We’re not just seeing technology slowly seep into our work lives. And this isn’t being counteracted by an opposing force of increased creativity and innovation in our fellow humans.

That’s a fairytale, and one that you should be cautious of.

Wells Fargo just released a report talking about how great it will be when banking needs 200,000 fewer workers.

What’s happening instead is you have massive industries, like transportation, banking, customer service, food service, etc.—all spending billions upon billions of dollars doing active research on how to automate their businesses.

McDonald’s is installing kiosks all over the country

That doesn’t mean just robots, or just machine learning, or whatever kind of tech—it means inventing anything that can reduce the number of unpredictable, flawed, and prone-to-get-sick-or-sue employees that consume a massive percentage of profits every year.

It’s not that these companies hate people—it’s that they hate inefficient and undependable parts of their business that hurt their margins. And people just happen to be those undependable inefficiencies.

Around 3 million people work in customer service, just in the United States.

Think about how bad customer service is, and how fast AI is improving. And then ask yourself where the crossover point is, after which AI becomes a better option going forward.

That point is different based on the industry and use case, but we’ve already crossed it for some, and many more are likely coming soon. When exactly? Nobody can say, but it’s no longer smart or responsible to assume things are resistant by default. We should assume the opposite.

The higher-end the product, the higher the chance you’ll still get a human on the other end when you call for help.

Automated reps don’t get tired. They don’t forget the training they just had on Monday. And you can upgrade them across the entire fleet in a relative instant. And to train more, you largely just copy and paste.

High-end products and services will still have human representatives for the foreseeable future, but most services aren’t high-end. Most customer service calls are extremely monotonous, routine, and simple to resolve. And those calls will increasingly be going to bots and AI-based agents.

Because humans are so static, and the types of things we need on a customer service call are so similar, the AI Customer Service Agent problem is similar to attacks against cryptography. Once you get a foothold, the attacks predictably get better, and they never get worse.

These attacks just happen to be against the experience of meaning humans get from working.

Summary


Human job loss to automation is not some gentle force that gets naturally counterbalanced by new types of work.
Automation is a profit-center, which means humans losing their jobs is a profit-center.
Companies and industries are spending hundreds of billions of dollars to replace humans in their workforces, which will in turn lead to billions more in profits for them.
Most people don’t realize this extremely active and funded component to the equation, and it’s time they start to.


Notes


I couldn’t find the artist who drew the horses, but I did the labeling and captions.

—

If you like my content, you can support it directly for less than a latte a month ($50/year) which also gets you the Unsupervised Learning podcast and newsletter every week instead of just twice a month.