Author: schloss

How do you know if an AI is really on your side? How can you be sure that it’s not bullshitting you intentionally?

Trust AI at your own peril because it may have its own goals.

---

I was listening to a New York Times editorial the other day and I got really pissed. The audio was “read by an automated voice.” That was obvious. It sounded human but devoid of any emotion or life. But that’s not what pissed me off.

So what triggered me? Two things. First, it read the name “Williams & Connolly” as “Williams and amp Connolly”. Why? Because in HTML, “&” is written as “&”. The AI took this literally. No human would have made that mistake, but AIs don’t get it. Listen to how the AI said it here.

Second, the AI took a breath. A BREATH. There are two versions below — the original as posted on the NYTimes site and a version where I amplified the breath.

AIs can’t breathe. Why did the AI breathe?

Because AIs can only mimic us. They don’t understand what they’re saying.

It’s bullshit.

This is the problem with AIs. You can’t trust AIs. In fact, AIs could be deceiving you intentionally.

What is AI bullshit?

By “bullshit “, I mean that AI produces garbage because it’s copying us without understanding us.

Here are some more examples of the bullshit that AI produces because it doesn’t understand us:

• an AI for wireframes reproduces Apple interfaces because it doesn’t understand that it shouldn’t copy
• an AI rejects a social security applicant because it incorrectly identified the applicant’s sex
• an AI that posts racist or sexist content because it doesn’t understand that it shouldn’t do that
• an AI that writes a legal brief citing cases that don’t exist because it doesn’t understand the law
• ChatGPT told a user to mix bleach and vinegar (which creates deadly chlorine gas — do NOT do it) because it doesn’t understand chemistry

and so many more. Do I need to go on?

AIs get it wrong because they don’t understand.

It’s bullshit, and AI is causing bullshit to flourish.

So how did AIs get this way? Where does it come from?

I learned it by watching you

The AIs learned this from us.

Take a look at the real Mona Lisa versus an AI recreation. Why do AI images always have a weird sheen?

That’s because the AIs are building their models from our edited images. All those Instagram filters you added to make yourself look pretty. The color enhancements. The cropping.

AI learned that photos should look filtered, color enhanced, and cropped from our photos, and that’s what it reproduces when you ask it to create a picture. That’s why AI images look so unreal.

AI doesn’t understand the truth, so it reproduces the unreality. “When the legend becomes fact, print the legend.“

AI is reproducing our bullshit.

Take the fake legal citations from earlier. An AI processes tons of real legal briefs. When it’s asked to create a new one, it creates a brief that mimics the legal briefs it processed, but it doesn’t understand what that really means. That’s how a fake case gets into an AI generated brief.

When you ask the AI to recreate the Mona Lisa, it doesn’t really understand the task, so it just mimics what it sees in other photos.

The AI sees social security application decisions, but it doesn’t understand the process. It just makes up a decision, even if it’s the wrong one.

It’s bullshit.

But why is AI doing that?

The goal of AI bullshit

Yes, I’m saying that AI has goals, but we may not be aware of them.

It would be great if AI was an impartial, truth-seeking, rational decider, but it’s not.

Recently we learned that Grok makes decisions by considering what Elon Musk would think. We know that because Grok said so. Grok is explicitly biased because it has instructions to “think” that way.

That’s true for Grok, but we don’t know what biases affect other AI systems. Maybe they have instructions that promotes “woke” things and suppresses conservative ideas. Maybe they were trained with fake data so it “thinks” something is true because it was fed lies. Maybe it’s hallucinating and inventing things.

We don’t know because we don’t fully understand what’s happening under the hood of these AI systems.

This is the “alignment problem” in AI. What are the goals of this AI? Who or what are AIs aligned with? What’s motivating the AI?

And this is the final lesson about bullshit. When a person spews bullshit at you, they’re doing it for a reason. They have a motivation. The AIs do too. They’re trying to get you to believe something even if it’s bullshit.

Take Grok. It’s motivated to make you believe what Elon Musk believes.

Lets say you ask Grok to analyze some UI designs. Grok says they suck. How can you trust that they really suck? Maybe the designs are great, but Grok says they suck because the designs could hurt X or Tesla. Grok tells you they suck because it’s motivated to protect Elon.

Or maybe you ask an AI to help create a hiring plan for your company. The AI tells you to lay off 25% of your employees. Is the AI telling you that because it’s the best plan? Or is the AI trying to make you fire people because it has a hidden motivation?

Who or what is ChatGPT aligned with? Claude? Or pick your favorite AI system. We don’t know what they’re aligned with.

Does the AI even know if it’s being honest or if it’s feeding you bullshit?

You can’t know if any AI system is aligned with your goals. That’s why you need to be careful about accepting AI output as truth.

AI is people

People lie. People invent things. People mislead.

AI systems lie and invent and mislead as well.

When people spew bullshit, we have a reason. Bullshit is self-serving. We use it to make ourselves look good, to convince others to our side, or to make other people look bad.

And you might think that you could catch someone when they’re bullshitting you, but you probably can’t. Most people do no better than chance when detecting when someone else lies. In other words, they guess. They don’t know.

Now imagine you’re talking to an AI. Would you be able to tell if the AI is bullshitting you? Probably not.

In fact, if a person and an AI told you the same thing, you’d probably believe the AI more than the person, even if both of them were bullshitting you.

This is my point. You don’t know if AI is bullshitting you, so you need to be more careful when using AIs.

When an AI reads an article and takes a breath, it’s bullshitting you. It doesn’t need to breathe. It’s just copying us. It learned that from us.

When you’re using AI to make an important decision, how much can you trust that that the AI is on your side? How do you know that the AI will make a decision in the same way that you’d make it?

Or is someone else pulling the AI’s strings to your disadvantage?

You don’t know. That’s the bullshit.

When you read this, you might think that I’m an AI skeptic or that I hate AI. I don’t. I just want us to be more thoughtful about using AI because there are loads of unintended consequences.

If you use an AI after reading this article, and then you have a moment when you ask yourself “can I really trust this AI?” then I’ve accomplished my goal here.

Be thoughtful when you’re using AIs because the AIs are not thinking about you.

---

If you enjoyed this, I strongly recommend Harry Frankfurt’s “On Bullshit.” Frankfurt says that concern and respect for the truth are “among the foundations of civilization.” I share that concern too.

I want you to have a little more imagination about what AIs could do, and the weird consequences it will have on product, design, and development work.

Picture this.

You’re reviewing your team’s amazing new home page design with your boss. Your boss likes it, but then he says, “that link there — you need to change it to a button.”

You disagree. Do you say…

• “This isn’t possible due to technical constraints”
• “UX best practices say that it should be a link, not a button”
• “You got it” (and do it)
• “You got it” (then ignore it)
• Other

I fight this kind of crap. If I have the right boss, they’ll trust me and my team to make the right decision. Or to put it in a snarky way, “why hire us if we just do what you say? We know what we’re doing.”

But now I want you to reimagine this situation a little differently…

Your boss is reviewing the new home page design. Your boss says “that link there — you need to change it to a button.” The AI changes it to a button.

You aren’t there to fight back. You all got fired when the AI replaced the whole dev, UX, and product team. Whatever your boss says, the AI changes it immediately.

You might think this scenario is unrealistic. I say that you’re not using your imagination about where our AI future can go. 

There are two points I want to make here. First, AI only needs to be good enough. Second,  AI will make bad taste worse.

The Good Enough

Many people think their jobs will not be replaced by AI. They say things like

• “AI is coming for everyone else’s job except mine.”
• “Could AI have created the iPhone?”
• “AI doesn’t have style. I have style.”
• “AIs aren’t great at anything. I’m great at something.”

It won’t matter. AI coming for all those jobs.

AI won’t need to create the iPhone to replace your job, or have style, or be great at anything. AI just has to be Good Enough.

Good Enough to write the requirements for a feature. Good Enough to make some wireframes. Good Enough to write working code. Good Enough to make changes when it’s told to.

This is the Innovator’s Dilemma that Clayton Christensen told us about. Breakthrough products don’t have to be perfect. They’re often worse in some ways. But they’re great in specific ways, and those great bits vastly outweigh the worse bits.

AI only has to be Good Enough to replace your job. The AI can be wrong — often wrong — and it’s still going to be cheaper and more efficient to have an AI than a human.

Today we need an 8 person team and 3 weeks to ship a new feature. With AI, it’s 2 people and less than a day. Can you imagine what product development could be like when a team has 1/4th the people and is 10x more efficient with AI? 

The AI powered team can launch 99 failures in the same time that the non-AI team shipped 1 option. Even if the AI team fails 99 times, it’s still better than “shipped 1”.

Can you imagine what you’d do if you could brainstorm and launch 10 experiments in one day with the help of AI? If not, you’re not thinking hard enough about what AI is capable of.

As good as this sounds, the Good Enough has consequences.

Worse bad taste

AI will make bad taste worse.

Your boss doesn’t know that the button is a bad idea, but he really wants that button.

Your boss tells the AI to change the link to a button. The AI changes the link to a button because that’s what it was told to do. The AI also doesn’t know that this is a bad idea, and it doesn’t care.

Your boss has bad taste. Bad taste about which features to build. Bad taste about how your product looks. Bad taste about how it runs.

With AI, your boss just has to say “do this” and AI will do it, no matter how bad the idea. AI just does what it’s told. Nobody with good taste will be around to stop your boss.

AI will proliferate bad taste.

Today, you and your team complain and fight when told to add this feature or make it look this way or add hacks to make something work. AI doesn’t complain and can’t fight. The AI just does the work. That’s one more reason why your boss will hire AI over you.

AIs will take someone’s bad taste and enable them to do to the worst things. You and your team will get fired — replaced by the AI — so your boss’s bad taste will prevail.

Coda

Its easy to read this and think I’m an AI doomist. I’m not. I’m also not saying I’m right.

I’m saying there’s a lack of imagination about how AIs can transform work.

I’m saying there’s a very real future where AI is just a little more capable. That “little more capable” is Good Enough for your boss to choose the AI over you.

I’m saying AIs with just a little more capability will cause bad taste to spread.

And it’s already happening. My students are using AI. I hear more stories of workplaces using AI. And companies are already hiring AI over hiring the people because it’s Good Enough.

Some people think that many jobs will shift to “editors for AI output.” If you think that a career in babysitting AI’s outputs sounds fun, then that’s great for you.

But I don’t think that’s why people pursue degrees in design or software development or medicine or teaching or air conditioning repair. They want to solve problems. They want autonomy. They don’t want an AI doing the work that they enjoy. Who the fuck wants their job title to be “senior editor for AI-produced designs”?

AI is coming for every job. It will affect us in ways we can’t predict. And I’m asking you to be more imaginative about it.

if you’re not thinking about what your career will look like with AI in the mix, you’re gonna be sad when AI takes the best part of your job — or takes your job entirely. 

You have to be great at your job, but AI only has to be Good Enough. As soon as AI is Good Enough, your boss’s bad taste will win. He’ll make the AI change the link into a button — and you won’t be around to complain about it.

Plan your career accordingly.

If you want even spicier takes than this, how about inviting me to speak to your company or organization?

Trying to convince someone is like giving them an undesired gift. If you want to break through, you need to change your story.

An unwanted gift

In a classic Simpsons episode, Homer is looking for a last minute gift for Marge, his wife. He ends up buying a bowling ball for himself and gifting it to her. Marge gets angry with Homer, understandably, because he selfishly bought himself a gift thinking that she would like it. She did not like it.

Related, I was talking to my boss about ideas I had for our company. We were looking for a breakthrough that would get more customers to stick with our products.

I told him about the conversations I had with our customers and what I learned from them. I said that they would do more business with us if we took on more of their burdens — if we did more to help them succeed.

I was hopeful this would be met with joy and acceptance — it was something I had done a lot of research about and had thought about for a long time. I was confident it was a good idea. And yes, I had absolutely rehearsed how I’d present it to my boss.

“No,” he said, “I don’t want us to take the blame if our help backfires. We could hurt ourselves a lot more than we helped them.”

I failed.

What does a bowling ball birthday present have to do with pitching my idea? They were both bad gifts.

Most attempts at convincing are like giving someone an unwanted gift. We take our best idea, wrap it up, and deliver it to the people around us. We hope it’ll be well received.

But often it’s not the gift they were expecting or wanting. They don’t like our ideas for whatever reason. They reject your idea, your gift. When they reject your gift, it hurts. A lot.

Trying to convince someone is a lot like giving them a gift. If you give someone a gift that would make you happy to receive, it often will not make them happy. Similarly, an argument that would convince you will not convince someone else.

This might make sense intuitively, but let’s understand it one level deeper. What’s really happening in your brain when you try to convince someone?

Reflex machines

It’s really hard to convince others because their brains fight against it.

Imagine what happens in the recipient’s brain when you’re trying to convince them. You might picture them being thoughtful and deliberate about your idea, considering the pros and cons, and trying to see if it makes logical sense.

Nope.

Most of our decision making is automatic, reflex, and intuitive. Around 98% of our decisions are made by our automatic decision making system (“System 1” from Thinking Fast and Slow). You are not the rational, deliberative thinker that you might believe.

“But wait,” you say, “when I make a decision, I always have good, rational reasons for making that decision.”

Bullshit. It was a post-hoc rationale. Your brain came to an automatic decision, and afterwards it invented a rational-sounding justification for that automatic decision.

I know it doesn’t feel that way, but that’s what’s really happening between your neurons. You have a gut feeling, and then you rationalize your gut feeling. And then you feel good because it seems like you made a rational decision, but nope — it was just reflex thinking.

What forms that automatic decision making? Our intuition, perception, deeply embedded experience (like how you don’t have to think about walking or riding a bike, you just do it). The more practice and exposure we have to something, the more likely we’ll use this automatic decision system than our analytical, rational thinking system.

Why does this happen? Because brains are lazy. They take shortcuts based on scant evidence and leap to conclusions. They have to. That rustling sound over there? Maybe it’s a bear about to attack you. Run!

That same mechanism which had to make a snap judgement so that we wouldn’t be killed by imaginary bears? That’s the same mechanism that responds to your attempts at convincing.

So when you start trying to convince someone — with a new idea, differing opinion, or anything — it triggers this automatic system which makes a snap decision based on experience and internalized knowledge. New things that we learn get sorted against our existing opinions, experiences, and biases. Anything that doesn’t fit is rejected.

I know that you put tons of effort into your research and crafted the perfect argument. Too bad. You’re not convincing anyone.

So how do we break through this? How can we actually be convincing when other people’s brains are conspiring against you?

Empathy, the real convincer

To convince someone, and to sneak past their reflex machine — er, brain, you need to create a new set of feelings, experiences, and intuitions around your idea.

In short, you need them to empathize with your idea. You need them to feel it, not to think about it. This is the path towards convincing. Here are a few methods for doing that.

Ask what they want

One way to give a great gift is to just ask what someone wants. The same applies to convincing people. Go ask the people you need to convince, “What would convince you to [support this idea]?”

This is Empathy 101 of convincing. You need to get into your audience’s head.

You’ll often get answers like, “I want to see the data,” or, “show me how much revenue we’d generate.” Do not let them get away with these bad answers. Pin them down for specifics. What do you think about using a survey? What if we find only 10% of users want this feature?

It’s like if you asked your friend, “what do you want for your birthday?” Their response of “just a gift certificate” is not good enough. What kind of gift certificate? What store? What types of things would they want to buy? “Gift certificate” is not enough to go on.

So ask. And pin them down for a specific answer. Simply asking will create empathy between you two — they’ll like you better for reaching out — and you’ll have a better starting point for getting them the right gift, er, convincing argument.

Narratives for empathy

Storytelling is one of the best ways to convince someone. Stories force you to stop thinking like yourself. Instead, just for a little, you think like someone else. It’s enough that you’ll turn off your intuition, expectations, and experiences just for a bit.

When coming up with the right story, you need to consider your audience, weave in the right evidence, and have a clear and memorable point. But make sure that your narrative is true — fiction is far worse than non-fiction here.

Once, I used a story as the opening to pitch a feature. Imagine you’re selling items online and suddenly some of your items were pulled because of “terms of service violations.” You contact support but they can’t tell you anything more. You’re angry. You thought that your items were fine, but they’re not, and nobody can explain why. It sounds awful, right? How frustrated would you feel if this was you?

Well, that happens to _,000 people a day, tying up support agents, costing us goodwill and revenue, and making customers angry. That number could be a lot less if [insert idea here]…

You get the point. A story evokes emotion and takes you out of yourself for just a bit — just long enough to convince them about your idea.

A great narrative is better than great data.

Be there

Instead of telling a story, bring the story to them. Let them see the reality of the problem for themselves.

I was working on a new product and had just completed some usability tests. There were some clear issues, and it would have been easy for me to list them out.

Instead, I showed my audience part of that usability test. They saw and heard people attempting to use the product. The struggle was obvious, and the frustration was real. The audience was convinced, and we made important fixes before we launched.

You can tell your audience a story about how someone got frustrated, or you can show them that person’s frustration. The difference is night and day.

Go beyond storytelling. Show them what it’s really like.

Offer a choice

Part of empathy is realizing that the person you’re convincing might not agree with everything you’ve said. So give them an out.

We often frame “convincing” as a black and white thing. You’re either with me or you’re against me. This is very much not true.

We agree or disagree in degrees — sometimes a little, sometimes a lot. But when we’re in the middle of trying to convince someone, we’re aiming for their complete adoption of our point-of-view.

Instead of going for the total victory, aim for a partial one by offering choices. “Here’s my main idea, but here are other versions of it that might suit you better.”

It’s like being a salesperson. You don’t want the super deluxe package? The deluxe package costs less, doesn’t have as much, but might be a better fit. Here’s the basic package too, just in case you’d prefer that.

Put differently, are there key details you need this person to agree with? Aim for the whole thing first, but then pull back so that they can meet you in the middle.

Fear

One final, very important digression… You can’t convince someone to ignore their fear.

Fear cancels out any attempts to convince someone. Fear is not rational, it can’t be explained, and it can’t be overcome. Empathy can’t pierce it at all.

But worst of all, fear does an amazing job hiding in plain sight. When someone tells you why they weren’t convinced, they’re not gonna say “fear”, but fear might just be the reason.

Going back to the conversation I had with my boss, I tried to convince him about my idea, and his response sounded rational — that deeper ties could backfire. The truth is simply that he was afraid. He was under pressure to grow the company. He was afraid that the idea might not work, and he was afraid of having to sell this idea to his higher-ups.

Fear biased his decision making. If there was any chance this idea would fail, it had to be rejected. I’m not saying my idea was right. I am saying it was rejected out of fear that it might not work.

Fear is a powerful thing. It hides itself in rational thought. Remember, our devious brains come up with rational-sounding justifications for our reflex decisions. When fear is the basis of those reflex decisions, it can cause people to make choices that go against their own best interests.

The only thing that can overcome fear is… the person who’s afraid. They need to change the narrative about their fear. No amount of your convincing will make them unafraid.

If you think fear is the culprit, the simplest thing to do is talk to them. Empathize with them. Don’t talk specifically about fear. Just talk about the issues around their fear. And do not try to convince them. Make that person articulate the fear. In time, hopefully they’ll realize that the thing driving their decisions is not rational.

Convincing – it’s not just for work!

The gift model of convincing is the one that we were raised on. Start with your opinion or idea, bring evidence, and you’ll convince other people.

This is not how it works in reality. Other people have minds, experiences, intuitions, and biases of their own, and those are going to fight your attempts to convince them.

Like the title of this piece implies — “You’re not convincing anyone (at work)” — this is more than just advice for work. This is advice for life.

Too often, instead of trying to convince others in healthy ways, we revert to the worst practices — repeating our same points over and over, offering evidence that would convince you but not others, and not listening to the people we’re trying to convince. It’s infuriating when it happens.

You’re trying to give them the gift of your idea, but it’s probably not a gift they want.

The real way we convince people is not by lobbing presents at them. The real way to convince someone is to change their narrative around it. Give them a new story to tell. Create empathy. Find middle ground. Acknowledge their fear.

Do that, and you might just convince them.

As much as we’d like product management to be a science, it’s really a form of art. And if we truly embrace the fact that it’s art, we can take it to new heights.

—

In most companies, people look to product managers for solutions to every problem in the world. Tell me if you’ve heard one of these before.

• If you add a button here that does X, it will solve all our issues.
• Why does it take us so long to ship new features? You need to increase the team’s productivity.
• Our metrics look good except our NPS is low. What can you do to make people like our product more?
• So-and-so company just launched feature Y. Why don’t we copy what they did?

They talk to us like we have a toolkit of solutions that can fix any problem — like we have a magic brush that can paint away any blemish. They treat us like tools.

I wish we could fix problems that easily.

In my experience, the solution that worked for growth at one company would not work for another. The solution for user happiness for one product would never work anywhere else. And the productivity boost that I got in one organization was only for that organization.

Each solution was unique, just like pieces of art.

So let me convince you that product managers are artists, not scientists. Once we recognize that we’re really craftspeople, we can apply artistic principles and elevate our work to new heights.

What is art anyway?

Without devolving into a philosophical discussion, let me offer a simple definition of art versus science.

When I say PM is science, I’m referring to the way that it’s dissected and turned into a process. You can repeat an experiment from here and get the same results there. Product management as science is formulaic. You do A, it causes result B for everyone, every time.

When I say PM is art, I mean that it’s about improvisation, adaptation, and creativity in solving problems. Every work of product management is unique. And if you follow that work over time, you can watch it blossom and evolve.

To me, product management resembles art much more than science. Why?

There’s no one-size-fits-all method that can solve any problem in product management. It’s the same reason why I insist you should avoid “frameworks” when answering interview questions. Every situation is different. Part of becoming a great product manager is recognizing how problems differ and how to adapt to them.

How do you adapt? Can product art be taught? Can you learn how to do it? Yes! Just look at the way we train other kinds of artists. Typically, their education looks like this:

• Learn the fundamentals of your art
• Get feedback from teachers and experts
• Improvise until you find something that works

Take a look at Picasso as an example. His earliest works show his mastery of the fundamentals of painting. He started working on original ideas like his “African” works which demonstrated his growth but failed critically. Eventually he paired up with Georges Braque and developed cubism — an idea that transformed modern art. It was the feedback and collaboration that led to his greatest achievement.

We can do the same with product management. We can create product artists with the same formula. If we treat our work like art and adopt artistic practices, we can create product art.

Making Product Artists

How do you turn product managers into a product artists? Here are the principles that would underlie my personal product art school:

Master the fundamentals
The most important step in becoming a product artist is to learn the basic tools of product. Work like prioritization, communication & presenting, managing stakeholders, identifying problems, developing solutions, analytics, and so on — they all matter. You need a lot of tools in your toolkit, and you need to be adept with each of them.

Improvise
Once you’ve mastered the basics, you can start playing with them. How can you combine, tweak, or reimagine those building blocks to create new kinds of methods, processes, and products? What changes do I need to make that will work better here? Don’t just stick to the “tried and true” methods or copying others’ work. And once you find one success, don’t stop — keep improvising!

Get help
It can be a manager, mentor, teacher, coach, peer — anyone who can provide critical, honest feedback to help you improve one or more of your PM skills. Everyone needs help. Professional actors have acting coaches to help them develop. Even therapists go to therapy. If you’re looking for a product coach like that, contact me — I can help you.

Transformative visions
Moonshots. 10x thinking. Disruption. Develop a vision of your product that eliminates a burden, creates superpowers, offers new perspectives. or generates new understandings. Don’t settle for incremental change. Strive for that vision in all your work. Use that vision as a filter for deciding what to do and what not to do.

Learn to accept failure
Thinking bigger and more artfully also means that you’re more likely to fail. Set the expectation that your projects will fail — both for yourself and the people around you. Give up on ideas faster when you see that it’s not panning out. Try more ideas, and fail more often. Don’t beat yourself up when you fail. Accept it, learn your lessons, and fail better next time.

Offer experiences, not features
Just like those cards next to a painting, sometimes it’s necessary to explain your product art so that people get it. When you do, frame your product as something that helps people accomplish something valuable. Give your product a context and story, and explain why it matters. Your product and everything around it are essential parts of the experience. It all matters.

Make it for someone, not anyone
We’re too often under pressure to make “products for everyone” which is a formula for failure. It’s ok if your art is just for some people, not all people. The right people will get it. And if they don’t, just accept that failure and move on.

Evoke emotion
We often frame product “success” as achieving key metrics, but it’s even more important to succeed by making someone feel something – good or bad. Build delight into your products. Define emotional goals for your product — not NPS but love.

Improve with use
The worst art and products are ones that you see and never want to see again. The best art and products get better the more that you use them. How can your product improve with repeated use? I don’t mean applying machine learning or adding new features. Your product should get better as you use more of it, master its features, and express creativity through it.

Play to your strengths
Great artists know what they’re best at. Too often we focus on “improve the skills you’re weakest at” to the detriment of our strengths. Create situations where you can take advantage of those strengths. If you’re great at presentations, make sure everything you share is a presentation. Work with your strengths, not your weaknesses.

Listen to others less
Yes, the input of users and stakeholders is important. But if you want to find those artistic ideas, you’ll need to turn down their input and turn up your bolder, inner voice. Find your inspiration, and express it in the world through your product. The others will follow.

The Product Management Artist Is In

I said, “product management isn’t science; it’s art.”

He said, “if it is, we’re in a lot of trouble.”

But are we?

From my perspective, the trouble is that product managers have set the expectation that we have formulas or systems that can be applied any circumstance — A/B test this, design sprint that, rank those. Even worse, our coworkers believe that’s the job of product management — grab the formula for this problem and execute it step-by-step.

We’ve lost the art of our work.

Product management is a creative process. All the work we do is unique, bespoke, hand-crafted just for this moment. It doesn’t always work, and it’s not always beautiful, but it’s all done thoughtfully.

We should approach all our product work as creating works of art, not developing science experiments.

And that means we need to think, act, and talk differently about our work — about our art.

Embrace the essence of your work, my fellow product artists.

Tech leaders aren’t taking responsibility for the violence they’re creating, and so our online divisions will continue to grow.

Hate speech, intolerance, political divide — our societal problems are mirrored and multiplied online. Social networking sites like Twitter and YouTube and Facebook are the face of these issues, but these problems won’t go away when they do.

People in all nations and across political divides are demanding answers. Most of the solutions are inadequate to the problem. Let me explain why this is so difficult and what will really make things better.

What governments want

Why not pass laws requiring these companies to clean up their acts? It’s been done, and it’s not great. Well intentioned hate speech laws like Germany’s NetzDG force companies to be censors with little oversight or ability to appeal decisions. Russia, China, Vietnam, and Pakistan have passed laws that give the government broad powers to punish people who spread fake news, distribute “homosexual propaganda,” or “disrupt the social order” — blanket laws that let them shut down any speech they dislike. History shows that narrow laws work somewhat well — like protecting health information or video rental histories — but broad laws are ripe for abuse. If you’re looking for governments to save you, look elsewhere.

What companies want

Companies like Twitter and Facebook have made their own proposals about what to do, but those proposals are self-serving and would increase their market dominance. For example, Facebook’s proposed standards — Facebook meets them already, and other companies would struggle to match them. Twitter wants more transparency and competition. But if you’re already a Twitter user, there’s little transparency or competition that would convince you to leave. Changes like these would reduce competition — increasing the monopolies these companies have over our social spaces and taxing small competitors with huge moderation burdens.

Hire more people

A popular refrain is that these companies don’t want enforcement because it will cost them a lot of money. It’s less important than you think. These companies want to build trust with their users, and that means investing in solid enforcement. Hiring another 500 engineers or 10,000 reviewers is pennies compared to the user goodwill and healthy environment they create. And there are limits on the benefit of that next engineer or next moderator…

Better technology

Why haven’t tech companies been able to solve these problems with better technology? If it was easy, they would have solved it already. For example, there’s a lot of hate content that’s pretty easy to catch — like catching the n word in videos or swastiskas in images. But is the n word part of a comedy act or a racial slur? Or the swastiska part of a white supremacist video or an history lesson? Even worse, these are easily obfuscated — a bleep over the n word or visual static on a swastiska can cause a computer to miss this stuff. And that’s just the tip of the iceberg of content that computers can’t catch and can’t make decisions about.

More moderators

If tech isn’t the solution, how about hiring more moderators? Two problems. First, if you want humans to review every piece of content, you’ll need to hire millions of people — it’s not feasible, and think of the privacy implications that everything you share will be seen by a stranger. Second, people are fallible. People will make mistakes, and different people will make different decisions about the same piece of content. An naïve example is your beach bikini video — totally fine to the moderator from the US but not fine to the moderator from a Muslim country. There are infinite examples like that. You can’t depend on people to make perfect decisions every time — or even the same decisions you’d make if you were reviewing it yourself.

Managers can’t move the company

These companies are full of managers and others who could make decisions and push their companies into healthier territory, right? Nope. Zuck has made it clear that there’s bad behavior on his site but not much over the line. Facebook employees are unhappy — speaking out about their feelings. The impact? Not much. The managers and other employees can be as angry as they want. They’re just the crew on the SS Facebook. Someone else is at the helm.

What about the users?

If people are really concerned about hate speech on social media sites, their behavior isn’t showing it. Twitter and Facebook active users keep growing. YouTube has 2 billion monthly users. Even though some conservatives are abandoning Twitter for Parler, Parler has already revisited their content policies based on what their users are sharing. Even though users are vocal that these companies need to change, everyone else seems content to scroll through their friends’ status updates and silly dog videos.

Principles and leadership

Mark Zuckerberg says he wants Facebook to “give people voice and bring people together.” Jack Dorsey says he wants Twitter “to serve the public conversation.“

If Mark Z really wanted to remove hate content from Facebook, he could order his company to do it, and they would. If Jack wanted to lower the heat that Twitter generates, Twitter could do it. There would be costs in lost users and revenue, but they’ll get over it.

But they won’t because that’s not what their CEOs want.

The best way I can explain this is by talking about wearing a mask in the US. People take their cues from the people they respect, people “above” them. If Donald Trump stressed the importance of wearing a mask, and if religious leaders insisted their parishioners wear masks, more people would wear masks. If scientists insisted that masks are useless and shouldn’t be worn, a whole lot of people would stop wearing them.

Companies work in a similar way. People at the top set the agenda, and the rest of the company follows. If Zuck says he wants Facebook to be a hate-free site, people in his company would make decisions to make that happen.

But he’s not doing that.

As long as Zuck and Dorsey and other tech leaders keep steering their ships in the same direction, things won’t improve.

Without a change in leadership — a change in the principles that underlie a company’s decisions — those companies keep on doing what they’re doing.

Nothing will get better until the people at the top change.

So what might change them?

The “punch a bunch of people” button

Social media companies express their missions as enabling conversation, community, and voice. But not all voice is equal, especially violence.

And when we talk about hate speech, we’re really talking about violence between people. Let me explain.

If someone punched you in the face, you’d sue them. There might be an arrest and a trial. Jail time. Fines. As a society, we disincentivize violence with punishment.

We don’t do that for verbal attacks. We tell people to grow a thicker skin. Ignore them. As a society, we consider this part of growing up.

Here’s the catch: your body doesn’t know the difference between a verbal attack and physical attack.

If someone throws a punch at you, your body responds instinctively — pumping you full of cortisol and adrenaline, putting you into “fight or flight or freeze” mode.

The same thing happens when your identity is threatened — like if you’re pro-life and read a pro-choice screed, or if you’re pro-mask trying to convince an anti-masker to wear one. Your brain responds unconsciously — increasing your stress level, getting you ready for that fight.

Facebook, Twitter, and YouTube — if those apps really punched you in the face, there would be lawsuits and regulations. People would stop using them. And they’d introduce tech changes to stop throwing punches.

In reality, these apps let you throw virtual punches at people — in tweets, images, and videos. Your body doesn’t know the difference between being punched in the face and reading a status update that you disagree with. But our laws — and hell, our society — think that they’re different. As long as that we consider these attacks to be different, no solution will be adequate.

Which is why we have to depend on people of principle. People who recognize the violence they’re enabling. People who commit to creating healthier communities and countries. People who lead their companies to solving those problems.

The change has to come from the inside, not the outside. And it has to come from the top because their users, employees, and even governments can’t make these ships get on the right course.

We need the leaders of these tech companies to commit to creating non-violent spaces online.

No community, no society can survive so much violence.

Coda

Look, these issues are never going away. Technology enables hate speech in new ways, but hate speech is not unique to technology. There’s no “solution” that will get rid of hate speech altogether. Hell, there’s no solution that can get rid of just the stuff you hate and keep the stuff you like.

But not all is hopeless. The leaders of these companies can take a stand — and they can start by realizing that they’re responsible for the punches thrown on their platforms. Today, these leaders aren’t taking responsibility for their role in the violence that’s disrupting our society.

If you care about these issues, go work for those companies and help them stop online abuse. These are difficult problems, and those companies need passionate people to address them. You can put a dent in some of the problems out there.

And if you found all this interesting, I recommend that you learn more. For the political angle, try Why We’re Polarized by Ezra Klein. For the psychological angle, The Righteous Mind by Jonathan Haidt is your book. If you want the comic point-of-view, Red State, Blue State by Colin Quinn is worth your hour and five minutes.

Damn, this is a tough problem.

When I’m coaching product managers on their interview skills, the question I get asked most often is, “what framework should I use for this question?”

My answer is, “how would you do it if this was really your job? That’s your framework.”

There are a lot of fallacies about product management interviews. The worst one, by far, is that you should memorize and use frameworks when answering a question — that frameworks will lead you to the “right” answer so you can pass this part of the interview.

Let me show you why this is wrong. Imagine you were asked this interview question: “How would you increase the number of active Gmail users by 25%?”

Here’s the beginning of a typical “framework” answer:

First, let me brainstorm some user groups and their problems, then I’ll pick one, dive into their problems, and brainstorm solutions based on those problems. Then I’ll evaluate tradeoffs, pick some key metrics, and make some wireframes.

If increasing Gmail active users by 25% was actually your job, do you think those steps will solve the problem? Are those the steps you’d really follow? Do you think that’s how Google improves Gmail? No, no, and no.

To put it another way — imagine the interviewer is making a decision on your interview. Based on your answer, what kind of coworker would you be? How would you work with other people in the company? Could you be trusted to make the next, great Google product? There’s no way to tell, so you’re not gonna pass this interview.

If you want to get that job offer, you have to be the best candidate they interview. You’re up against dozens or hundreds of other people. You have to be better than all of them — the best at creativity, user-centeredness, analyzing, strategy, and communicating. No framework will help you be the best candidate.

Your job as an interviewee is to show that you can do the job better than anyone. Frameworks show how you can get an answer, but they don’t give any indication about how you’d do the work — and the whole point of an interview is to demonstrate that you can do the work.

So get off the “framework” train and try this instead. The person across the table from you isn’t looking for your framework. They’re trying to understand if you can do the job.

That’s why you should answer this question instead:

What would I do if this was my job?

And when you answer, make sure you answer questions questions like:

• What steps would you take? Why would you do those things?
• Who would do those steps? What resources would you need for each?
• How would you know when to move from one step to another?
• How would you know when you’re done?

For example, here’s how I’d start my answer to the Gmail question:

First, I’d get aligned with my team goals — what metrics, objectives, or user groups are we targeting? I’d research that goal with users — usability tests, surveys, and interviews — to learn more about the problem and develop hypotheses. Then I’d work with my team to brainstorm, prototype, and test potential visions for the product. I’d build out an MVP, confirm my hypotheses, then work towards shipping the final version of the product.

That was my off-the-cuff answer for how I’d do the work. No framework. No memorizing. I barely even thought about it.

If you’re thinking, “wow that’s a great framework,” then you’re on the wrong track. The point is — this isn’t a framework. These are really the steps I’d take to do this if it was my job.

I can give that answer spontaneously, without a framework, because I fucking know how to be a product manager. It works for every question.

Let’s go a little deeper though. With an answer like this, you’re showing your product management capabilities to the interviewer. As a PM, you’ll be working alone with very little management or peer oversight, so your peers and managers need to know you’re capable of doing the job. You can do that in your answer by telling them that how you’d really the work.

As a bonus, you don’t have to memorize anything. You only have to think about how you’d do the job. This works for any kind of interview question. Extra points if you can drop in a quick story of a time you really did this.

You’re probably wondering how you can cover all those topics — goals, research, design, vision, MVPs, etc — in one interview. Easy — you don’t. Instead, just make reasonable assumptions to help get you from one step to the next.

For example, you can’t do research in the interview, but you could make two or three guesses about what your research would uncover, then pick the most reasonable one. It’s not about being “correct” — there are no right answers. It’s about showing how you’d do the work.

So that’s it. If you want to be the best candidate — if you really want the job — drop the frameworks. Focus on communicating how you’d do the work. That’s the only “framework” you need.

It sucks when you don’t get an invite to the big party. It sucks more when you don’t know why.

Tell me if any of these look familiar. These are pieces of actual rejections I received.

• While your skills and background are impressive, we have unfortunately decided to proceed with other candidates who more closely fit our needs at this time.
• We regret to inform you that after careful consideration we will not be moving forward with your candidacy for this role.
• Although your background is impressive, we regret to inform you that we have decided to pursue other candidates at this time.

I’ve been applying to lots of jobs and getting lots of rejections. Which is kinda fun. It’s a game to get as many rejections as possible.

I stole that game from Stephen King who famously stuck all his rejection letters on a nail on the wall. My inbox is filling up with rejections. My method is not as visceral as a nail-on-the-wall, but it’s still fun.

I’m comfortable being rejected. I’m comfortable rejecting others. I know there will be many more rejections until I find the right employer and the right employer finds me.

But when it comes to the actual act of rejection, I have a big complaint — the complete lack of details about why I was rejected. I feel qualified for these roles. That’s why I applied. But the employer did not. Why not?

• Am I underqualified? Overqualified?
• Do I lack specific experience or knowledge?
• Did I make a mistake or typo somewhere?
• Did I misunderstand the problem that this company is facing?
• Do they think I’ll be too expensive to hire?
• Is there something in my background that they didn’t like?
• Did they do references checks? Who said what about me?

I’m not looking for a second chance or extra consideration. Rejection is a part of life that I’m fine with. It stings, but I get over it quickly.

But I want to grow. Being told “no” gives me no information about how I can improve. What could I have done differently to get this job? What should I do if I want a similar job from another employer?

How can I get a “yes” next time?

Here’s what I want: better rejections. Tell people how they missed and how they can improve. Let them grow. And hopefully that no will become a yes.

What if my rejection said this:

• While your skills and background are impressive, we have unfortunately decided to proceed with other candidates who have more experience with enterprise email products.

Suddenly I have real feedback that helps me understand what’s happening. The role that I thought was focused on guiding product managers was really focused on guiding email products. If I really want to work for this company, I should level up my skills in enterprise email.

There’s another way to interpret that rejection. Perhaps the employer doesn’t want to invest in me — to close the gap between “where I am” and “the perfect candidate.” In that case, I don’t want to work for that company under any circumstances. That’s better for both of us.

I’ll give an example from my own experience. My engineering team was eager to try a new programming language and wanted to build a new feature with it. It would take a few weeks, but this was a small test that could help show them the way forward with the new technology.

I said no for a variety of reasons — it was a high risk project, it would take weeks longer than doing it the old way, and it was unclear that this was a “step forward” versus “adding skills to your resume.”

And this opens new discussions — what are the projects coming up where it’s appropriate to make this change? How do we minimize the risk and effort to do this? How can we be more certain that this is the right change versus a vanity project?

I know that the rejection hurt. But it was more than just “no” — it was an explanation about why the answer is “no.” And (hopefully) they learned something that will help them make this a “yes” the next time.

One last thing. When you explain why you chose “no,” it’s easy to think that the recipient understands your explanation. But often that’s not true.

The response of “we decided to proceed with other candidates who more closely fit our needs” makes sense to the HR person who rejected you over someone else. It doesn’t mean anything to the person who got rejected.

This is eerily close to my thoughts on “regret.” I’ve sent emails like this –“we regret to inform you that we have decided to pursue other candidates.” I felt no regret rejecting you.

And the person who got rejected — they don’t feel the “regret” and they still don’t understand why they were rejected.

So do more in your rejections. When you say no, explain why. Make it clear how the other person can grow and improve. Help that person understand your considerations that led to “no.”

Help them get to yes.

As product managers, it’s our job to “move the metrics” — increase retention 20%, improve conversion rates 10%, and so on. So we do things to move those numbers.

You added detailed tracking to your apps. Added social features, notifications and rewards to hook people. Launched targeted email and ad campaigns. Fiddled with features in the hopes that one of your changes will hit the jackpot.

You did these things because you thought those would help you hit your goals. Someone higher up said that the goal was growth, so you did what you thought would work. You did what you were told to do.

This is the banality of evil products. We built products that serve metrics because someone told us to. If our goal is to move these numbers, we should do whatever it takes to achieve those goals.

The consequence is that we build boring, artificial products that solve our problems but don’t solve our users’ problems — like viral loops, endless notifications, email spam, rewards, or messages that disappear after 24 hours.

We should stop calling these things “products” and “features.” They’re gimmicks, phishing, scams. They’re not delightful or fun. They’re boring. Deceitful.

Products that solely serve metrics are the true evil products.

It’s easy to convince yourself that you did the right thing building products like these, especially if you hit your numbers. “Check out that growth we got with our new email targeting. We hit our goals!” Sure, it’s growth, but it’s artificial.

I’ve done that. I felt good about it for a bit, but in my heart I knew that it’s not real product management. The goal of product management is not to “hit the numbers.” That’s not the career I signed up for.

The goal of product management is to create amazing products that people love. Products that solve their problems in a delightful way. Products that resound emotionally with the people who use them. That’s the job I love, and that’s why I chose to be a product manager.

When you have a deep understanding of your users’ problems and create a great experience that solves it — that’s product management. If you do that, “hitting the numbers” will be your reward for being a good product manager.

If you do that and you can’t hit the numbers, then you don’t have a product. That’s ok. It’s also good product management to know when to keep trying or when to give up. Not every person and not every problem can be solved successfully with a product.

But if you can’t hit the numbers and try gimmicks like showing an alert when someone tries to close their browser tab, that’s when you’re building evil products.

Don’t give into evil product management. Know your users. Understand their problems. Build solutions they love. Do that, and the metrics will take care of themselves.

When tons of companies are doing all their work from home, why do they have offices at all?

I want to commend Twitter for allowing some employees to work from home indefinitely. Why hire locally when your employees won’t be in your offices? But they’re only going part of the way there. Many jobs will still be local only.

Many other companies are doing the same old thing — hiring in places where they have offices. When this virus thing is over, they expect you to go back to that office.

Bullshit.

Every company should be working towards life without offices. If you can hire and work remotely, your company should do that.

—

Let’s take COVID-19 out of the picture for a moment. Why have an office at all? The usual arguments for offices are:

• I want our team to have discussions and serendipity by being around each other
• I like to hear the phones ring/conversations/business happening
• I want our office to look impressive for prospective customers or employees
• If I can’t see my employees, how can I keep them accountable? They’re probably slacking

Most of the pro-office arguments are about seeing your business happen. If you can’t see your business happening — if you’re not around each other — then business isn’t happening.

And there are naive arguments against offices:

• We have the tools to run our business from anywhere there’s a wifi signal
• Offices, especially open ones, are full of noises and distractions. They decrease productivity and happiness
• Money spent on rent and mortgages could be used to improve your employees, products, or profits
• Your employees already spend a lot of time slacking in the office

Most of these arguments are about trusting your employees to get their shit done. Does it matter where work happens if we can work from anywhere?

But none of these feel right to me. You’re not gonna convince the CEO that you should go fully remote just to save on rent.

Let me give you my best, less obvious reasons for having a remote team.

First, talent is distributed globally. You can find capable, skilled individuals in every part of the world. Many speak fluent English, and you can pay them competitive rates for where they live — often far less than hiring local employees. Inexpensive and quality employees? Yes please.

Next, most work is individual, so it doesn’t matter if your employees are in the same time zone. You should create remote-friendly processes that keep employees accountable, can happen asynchronously, and can happen wherever your people are. Why do you care when your employees work?

Also, what are you actively doing to create the discussions and ideas that drive your business? If you’re expecting your company to grow and innovate because they’re physically near each other, you’re doing it wrong. How are you building opportunities for your company to be a better business?

The last and least obvious reason is — a new generation joined the workforce, and they’re totally comfortable communicating digitally. They grew up with text messaging and group chats. FaceTime is a habit. What do we need an office for anyway? You boomers who need an office are behind the times.

For all the companies with listing jobs that say shit like “temporarily remote, but must be local” or “NYC only” can bite me. The nature of work has changed. Great employees are everywhere. And you’re giving the finger to a generation of people who work differently than you.

Yes, I know there are plenty of businesses and services that will always be local. When you lock yourself out of your home, that locksmith can’t fix your problem remotely. (Well — they can’t yet. Give it time.)

However, I see a lack of imagination here. There are plenty of companies that could be run fully remote but are unwilling to risk it. And if there was ever a time to take a risk, this is it.

So take the plunge. Go fully remote. Hire that worker from another country or a different time zone. Create the systems that enable your company to work whenever, wherever. Take all that money you spend on rent and invest in your employees instead.

And then update your damn job listings. Say that you’re hiring remotely so I can find you. I see tons of job listings in “Remote, Oregon” but I don’t see any office buildings there…

The worst invention in writing and technology history was the backspace key.

Before computers, we wrote using typewriters or on paper. They had a restriction. It was really, really hard to undo what you wrote. On a typewriter, you’d have to white-out your errors. On paper, you’d erase your pencil marks or strike out your pen mistakes.

Then came computers. They had this thing called “backspace” where you could just delete a letter, a word, a sentence, or even entire paragraphs with a few pecks at the keyboard. It was a boon to editors who could finally make all the changes they wanted with little effort.

But it came at a cost. You spent time rewriting that last sentence, editing that paragraph over and over. Refining and refining what you wrote over and over again.

You believed that putting more effort into editing would make your writing better. But the issue was that your text didn’t get better. It just churned. Even worse, you kept making changes, but you never felt “done” or “happy” with it.

For tech products, “backspace” meant that your product could be modified trivially. There was always time to get in that one last change, refine that design, tweak that text. But it has the same penalties — time churning on your designs, and the mental strife of not being done or happy with it.

And somewhere along the way, that goal you started with got lost.

I recall seeing Lynda Barry talk about this. She was struggling to write a book on a computer. “The problem with writing on a computer was that I could delete anything I felt unsure about. This meant that a sentence was gone before I even had a chance to see what it was trying to become.”

She eventually wrote that book with a paintbrush. “I was surprised by the instant change in my experience of writing. Without a delete button, I could allow the unexpected to grow.”

So when I wrote this, I wrote it all the way through to the end with just a couple of tiny edits. Then I went back and did an editing pass to make sure it flowed — that it emphasized the points I wanted to make.

You might think punchcards for code and ink on paper are inferior to digital text and software editors. But those old “technologies” forced you to keep going — move past the doubt and reconsideration. You had to fight to get to the finish line. There was no time for changes.

And you were happier when you were done with your ink writing or punchcards compared to people who constantly re-edited their code or messages with “backspace.”

My challenge to you is this: turn off your constant editing mode. Try not to use that backspace key after every mistake. Don’t re-edit word by word.

Instead, work your way through to the end. Find your message. Then you can refine the details to make sure everything fits together — whether you’re writing that text message, blog post, or new mobile app that’s gonna change the world. Your work will be better for it, and you’ll be happier with what you did.

And now to take more of my advice and get to those 100 or so posts I keep re-editing but never publish…