4 Steps to Changing People's Perceptions

Welcome to my newsletter! I'm Dave Anderson, an ex-Amazon Tech Director and GM. I write this newsletter I've called Scarlet Ink, which is a weekly newsletter on tech industry careers and tactical leadership advice.

Free members can read some amount of each article, while paid members can read the full article. For some, part of the article is plenty! But if you'd like to read more, I'd love you to consider becoming a paid member. 99% of my articles are intended to be evergreen (readable and valid forever). Some weeks I have fresh content; other weeks I'll update/rewrite something from 4+ years ago because I want to keep the quality of all articles high.

During a backpacking trip a couple of days ago, my techy group of friends were discussing why AI has made incredible improvements over the last years regarding writing code, but writing English has lagged.

I've been wondering that myself. I'm a happy Claude user, and I swear every model update has made it easier for me to build my stupid little pieces of hobby software. It's pretty fun to have clear and obvious improvements appear so rapidly. At this point I'm limited by my imagination, rather than my technical abilities.

As part of my usual tests when a new model appears, I ask Claude to create some social media in my writing style. I dislike writing social media, and it's a necessary evil for content creators. Well, to be honest, I've skipped it for a few months now because I'm super lazy, but that's beside the point. I certainly wouldn't want AI to write my newsletter, considering writing is my hobby. However, it's an interesting idea to see if Claude is capable of generating short-form content (based on my long-form articles) so I could repurpose it on various platforms.

And yet again, Claude fails miserably, according to my personal taste. I'll provide it with examples. I'll tell it that it shouldn't sound like an AI. I'll have it create a set of writing instructions based on 100k of my words and try again. And over and over again, it is obvious to me that the writing is AI and not my voice.

We came up with many interesting reasons why AI writes like AI, but here's the interesting one relative to this article.

Code is provably correct or not. When they train AI on code, this is usually compiled, tested, and correct code. Code has a right way of writing it and a wrong (literally not compiling) way. Sure, there might be seven ways to write a loop, but there aren't really an infinite number of creative ways to loop. Most people arrived at a coding standard, and so most production code will look vaguely similar if you're writing in the same language, in that product space. When you ask AI to write code, there are proper patterns to follow. Writing code "like an AI" means writing like the average coder who writes proper production code. It's not usually a bad thing.

Writing English? It's consistently nonsense. Look at how I write. I'd never write this way in an Amazon OP1 document, and I know that very few other humans would naturally write in this way. Style is a combination of word choice, sentence structure, convention following, and more. For example, I am thrilled with the usage of parenthetical statements, but I can't stand long paragraphs.

Digression over - let's talk about perception.

To wrap up my digression, computers are great when there are right answers. But humans perceive the world with a wide variety of unprovable nuance. We're biased, and inconsistent. One person might read an AI-generated article and think it was well written. I would likely read the same article and feel like it was the mental equivalent of nails on a chalkboard.

Clearly your view of the world is different from other people. We all know this as a fact, but it's difficult to wrap our brains around the consequences of this fact.

Obviously, we know that means that we each have different political opinions, or different opinions on the best sports player.

But it also means that we view other people's actions through our biased lens. And what we rarely accept is that our view of other people's motivations and actions is just as biased as whom we think the best sports player is.

If you were being honest, you'd comfortably admit that your opinion of the best sports player is just as right or wrong as another educated opinion.

But what about your opinion that a co-worker made a mistake with their project because they were being lazy? This is your opinion as well. It's your perception of their work and motivation, not a fact.

How everyone interprets the world is the truth and a fact in their world. Can someone tell you that you're wrong to think that your co-worker is lazy?

Perhaps if they gave you new facts, you might change your perception. But if someone said, "No, I'm not lazy!" it wouldn't change a thing. Everything you encounter in the world is filtered through your perceptions. It's based on everything you've gone through in your life to this point. How can someone convince you to see that person differently? They would need to provide new inputs into your perceptions.

It's not your interpretation of events that they can change. Instead, they need to change the events so that you will interpret things differently.

I put that in italics because I want to emphasize that statement. It's a key to changing your view on human interactions.

Here's a classic situation.

A husband says, "I feel that you care more about work than about me."

His wife responds, "You're wrong; I care more about you."

Was the husband wrong? No, they have spoken honestly about their feelings. That can't be wrong, since it's their perception of the situation. Was the wife wrong? No, again, that's also how they feel. And that's not wrong, since it's also their perception of the situation.

How can this be true? What's the reality? Who's right? Of course the answer is that both of them are.

That's the tragedy of most relationship fights (at work and at home). We're fighting over differing perceptions when there's zero chance we'll agree to change our mind without changing the inputs to our perceptions.

As I mentioned before, the key is changing the inputs. You need to change what the husband sees and hears so that the husband changes their interpretation of the situation.

At work, your manager says, "I feel like your work output is low."

You immediately think, "But you're so wrong! I get a ton done!"

Do you think that response would change your manager's mind? I think it's obvious that it has zero chance of changing their mind. However, I constantly heard people take that approach.

You need to address what created that perception for the manager and address those inputs.

This doesn't necessarily mean working harder or creating more work output. This is about directly addressing how the perception was reached and how to influence that perception.

My experience.

When I look at my successes at Amazon, I'm fairly sure that this was a superpower of mine. I can't understate how frequently I turned what might have been interpreted as a failure into a success. Or made a small success seem like a huge one. Or made one of my relatively average team members look like a rockstar.

Doing real work is valuable. But being able to look past reality and manipulate perceptions is critical in any moderately large workplace. Follow along, and I'll explain my slightly evil methods.

Operations and team perception.

Sherri managed a high operational load team. The team had complex systems, high customer usage, and regularly had operational tickets related to various systems failing. Her team had a fantastic alarm setup that identified outages quickly, and Sherri would hop onto calls, spin up the various ops teams to help handle the next emergency. She often handled tickets on the weekends, at 2am, and on holidays. No one else would want the job as manager of that team, as the personal load on that leader was high. Their team was only functional in large part due to the heroics of their manager. They still managed to hit most of their feature launch dates amid the high operations storm.

When we brought up Sherri in the "how much of a raise should we give people" meeting, it came down to a debate on perception.

• Some leaders viewed her as an indispensable hero. She was a leader in an incredibly challenging space. Directly in the line of fire for most operational events, she'd kept her team from having high turnover. She was deeply technical and personally handled many events herself. She kept projects on track by ensuring her team also completed the necessary feature work. These leaders believed she deserved a huge raise. Perhaps even a promotion.
• Other leaders viewed her as a hacker. She had to do those heroics because she hadn't managed to address the operational issues in her space. She had been in place for years, but the situation was no better than when she joined. If anything, it was worse. She was patching things when she should be looking at root causes. The fact they kept doing feature work instead of resolving their operational issues was part of the cause of their continued problems. These leaders felt that Sherri was the main problem for the team, and they should find a new manager to fix this disaster.

Which leaders were right? The answer (if you've been following along) is that there is no right answer.

Sherri did solve operational problems quickly and professionally. She personally went out of her way to ensure that customers were rarely impacted for long. This is a good thing.

She also should have made progress on improving their operational situation and should have pushed back on the roadmap for her team until she got things under control.

Depending on every leader's bias, they viewed one of those things as more important than the other, and this influenced their perception of Sherri.