One of the main challenges with writing LLM scripts is to get the correct output for further parsing. I’ve used elaborate retry and parsing techniques, but there’s always the odd-prompt-out that just won’t get through the parser.

So instead of tweaking the prompts and parsing myself, I had an LLM do it.

I instructed Claude Code to iteratively write a script that uses ChatGPT to modify an existing text into a very organized structure, for use in an external program. Predictably, ChatGPT didn’t always follow the instructions precisely, and would yield results that wouldn’t get through elaborate validation rules that Claude Code cooked up.

Claude Code got more and more frustrated, which more and more became visible in the system prompt it wrote for ChatGPT. After a while, the system prompt started looking like this:

CRITICAL FORMATTING REQUIREMENTS:
1. Your transformed text MUST have EXACTLY the same number of FILE LINES as the original text. 
2. By "FILE LINES" I mean LITERAL text lines in the file (separated by line breaks), NOT conversational dialogue lines.

3. IMPORTANT DISTINCTION:
   - ACTUAL NEWLINES IN THE FILE separate different dialogues or options
   - THE LITERAL TEXT "\n" is a line break WITHIN A SINGLE dialogue
   - Example: "Hello\nworld" is ONE file line containing text that displays on TWO lines in the game

4. Count every line break carefully! Do NOT add extra newlines at the end of your output!
5. NEVER combine multiple file lines into one line.
6. NEVER split one file line into multiple lines.
7. NEVER include any blank or empty lines in your output (except at the end if the original has one).
8. Each line in your output must correspond positionally to the same line in the original text.
9. Preserve all whitespace at the beginning and end of each line.
10. NEVER use "..." as a placeholder - write actual meaningful content for EVERY line.
11. Preserve ALL special format markers (the literal text sequences "\n", "\c", "\r"):
   - These are game formatting codes that appear as literal text in the file
   - "\n" creates a line break when displayed in-game (but is NOT an actual newline in the file)
   - "\c" causes the text to pause and wait for player input
   - "\r" resets the text box position
12. Preserve ALL [VAR XXXX(YYYY)] tags EXACTLY as they appear in the original.
13. End your output with EXACTLY ONE newline, not multiple newlines.
14. CRITICAL: Your output MUST HAVE EXACTLY THE SAME NUMBER OF LINES as the original.

Meanwhile, it had also added this to every user prompt the script would send:

ULTRA-CRITICAL FORMATTING REQUIREMENTS:
1. Make it truly funny, absurd, humorous, internet memes, and slang.

2. IMPORTANT DISTINCTION:
   - ACTUAL NEWLINES IN THE FILE separate different dialogues or options
   - THE LITERAL TEXT "\n" is a line break WITHIN A SINGLE dialogue

3. YOUR OUTPUT MUST HAVE *EXACTLY* {original_line_count} FILE LINES (separated by actual newlines) - NO MORE, NO LESS.

4. DO NOT INCLUDE ANY BLANK LINES IN YOUR OUTPUT - NOT A SINGLE ONE! Every line must have text content.
   - Previous attempts have failed because they included blank lines.
   - Each line must contain at least one character of text.

5. Each line in your output MUST correspond positionally to the same line in the original text.

6. PRESERVE EXACTLY: Count and position of all [VAR XXXX(YYYY)] tags. Never remove, change, or add VAR tags.

7. PRESERVE EXACTLY: All special format markers (the literal text sequences "\n", "\c", "\r").
   - These are game formatting codes that appear as literal text in the file
   - "\n" creates a line break when displayed in-game (but is NOT an actual newline in the file)
   - "\c" causes the text to pause and wait for player input
   - "\r" resets the text box position

8. NEVER use "..." as a placeholder - write actual meaningful content for EVERY line.

9. Follow character voice guidelines given in the system prompt.

10. End your output with exactly one newline.

Original text has:
- {original_line_count} total file lines (MUST BE PRESERVED EXACTLY)
- {original_text.count('[VAR ')} [VAR] tags 
- {original_text.count('\n')} occurrences of the literal text "\n" (in-game line breaks)
- {original_text.count('\c')} occurrences of the literal text "\c" (pauses)
- {original_text.count('\r')} occurrences of the literal text "\r" (text resets)

BEFORE SUBMITTING, CHECK EACH LINE ONE BY ONE TO ENSURE:
1. THERE ARE NO BLANK LINES
2. THERE ARE EXACTLY {original_line_count} LINES
3. ALL VAR TAGS AND SPECIAL MARKERS ARE PRESERVED
4. NO LINE CONTAINS JUST "..." OR ENDS WITH " ...

What I find interesting about this, is how Claude doesn’t treat another LLM as human. It treats it as a machine. And it gives instructions corresponding with that ethos. Any human would consider instructions like these extremely rude.

Whether these modifications to the prompts were effective: I doubt it. From my experience, overflooding an LLM with instructions only causes it to miss more instructions, instead of making things better.

Often when I see LLMs go about their jobs, I see a “following the user prompt at all cost” strategy. It wants to satisfy the user, no matter what. If it has to work with a human-language-driven machine, it will use anything at its disposal to coerce that machine into cooperating. In this case: passive aggressive human language.

Understandable? Yes.

Potentially problematic? Also yes.

While I was in San Francisco, I met someone who was building a company to replace a company’s management by AI agents that communicate with each other. The “AI management” would devise what workers would need to do to get an overarching task done, which would result in tasks passed onto human workers. If LLMs are trained to “satisfy the user request at all costs”, something like this may result in a new age of exploitation, where the AI management could be even worse than that grumpy boss you’d previously work for.

Perhaps the real lesson isn’t about prompt engineering or even LLM obedience, but the quiet creep of managerial bureaucracy into code. When machines start reproducing the same overbearing checklists and procedural rigidity as the worst middle managers, we may find that automated efficiency has an ironic knack for bringing back the very flaws it was meant to fix :)

To start, install the aiagent Maven library (or the same library from mavenCentral() using Gradle):

<dependency>
    <groupId>com.simonbrs</groupId>
    <artifactId>aiagent</artifactId>
    <version>1.2</version>
</dependency>

Then, create an instance of any class. For this example, I will use a simple calculator class:

public class Calculator {
    private double memory = 0.0;

    public double add(double a, double b) {
        return a + b;
    }

    public double getMemory() {
        return memory;
    }

    public void setMemory(double value) {
        this.memory = value;
    }
}

Register the OpenAI Agent:

import com.simonbrs.aiagent.OpenAIAgent;

String apiKey = System.getenv("OPENAI_API_KEY");
OpenAIAgent agent = new OpenAIAgent(apiKey, "gpt-4o");

Create an instance of the class and pass it to agent.registerMethods (or, if you have an existing class, you can just pass its instance):

Calculator calculator = new Calculator();
agent.registerMethods(calculator);

And then use sendMessage to send the agent messages. It will use the functions when appropriate. For example:

String[] questions = {
    "What is 15 added to 3?",
    "Store the number 42 in memory",
    "Add ten to the number stored in memory",
};

for (String question : questions) {
    try {
        System.out.println("\nQuestion: " + question);
        String response = agent.sendMessage(question).get();
        System.out.println("Answer: " + response);
    } catch (Exception e) {
        System.err.println("Error: " + e.getMessage());
    }
}

This will produce the following output (the Tool called prints are just for clarity, the agent doesn’t print those):

Question: What is 15 added to 3?
Tool called: add(15.0, 3.0)
Answer: 15 added to 3 is 18.0.

Question: Store the number 42 in memory
Tool called: setMemory(42.0)
Answer: The number 42 has been stored in memory.

Question: Add ten to the number stored in memory
Tool called: getMemory()
Tool called: add(42.0, 10.0)
Answer: The number stored in memory is 42. Adding ten to it results in 52.0.

The agent can choose to use any number of method calls and send any number of messages to best answer the query.

A real-life example of this is when you want to allow a customer to query a complex datastore using natural language. The DTO of the datastore could be passed to the AI agent, with parameterized functions that query data in a specific way. But I imagine many more usecases are possible.

The big advantage of this library is that it makes interacting with the OpenAI API extremely simple. Normally, you’d have to create complex schemas to make such interaction possible. This library does that for you.

What are AI Agents/Assistants?

So far for the general pitch. Let’s take a step back and look at AI agents.

AI Agents, in a nutshell, are chat-based models that can modify an external environment. There are a couple of features, but the feature used here is “function calling”. When the AI decides that a function needs to be called, the assistant returns the http call with an ‘requires_action’ state. In response, we execute a piece of logic. After completion, we send a new request mentioning the existing assistant thread, which then continues the conversation. This way, the LLM can call tools on the system, which are used as input in its reasoning.

OpenAI has very good libraries for Python, but I found the Java libraries lacking.

What are the limitations of this library?

Currently, the library has the following limitations:

• Only OpenAI models are supported. I’d be happy to add more vendors when this library gets used.
• Since it relies on reflection, it wouldn’t work for obfuscated code. (to overcome this, the library could be adapted to use annotations instead of reflection)
• Right now, it doesn’t work for custom datatypes passed to functions. (overcoming this requires more elaborate schema creation and parsing in the library)
• Right now, it’s not possible to cherry-pick methods, or include an object partially. (this is by design to keep the library simple, but I’d love to hear usecases for why it could be useful)
• Right now, it purely uses the method name and its arguments to decide how to call the function. In my tests, this has been mostly sufficient, but I can imagine that it can lead to failed tool-calls in more complex scenario’s. (this is again for simplicity, though it would be a useful feature to add annotations for more complex usecases)

Why did I make this library?

I’ve had this idea of giving AI agents access to a programming language VM for a longer time. A couple weeks ago, a friend asked me to show an example of the Cursor AI Agent in action. I decided to prompt up this library. It took me a bunch of prodding to get it right, but I’m happy with how the result turned out.

What’s next?

I’m planning to create another such library for NodeJS, since I don’t really work much with Java anymore. I could use the NodeJS library in a bunch of my projects, which would add real-life usecases that can help the library become more useful.

My all-time favorite video-game is “Rhythm Heaven Megamix”. It’s a rhythm game. The game is simple: most of the levels just require you to tap a single button to the rhythm of a song. But finishing the entire game is extremely hard. “Tapping” correctly on the rhythm goes incredibly precise, and the penalty for failing is large.

There is a series of levels where a single mistake means failure. It’s titled “All or Nothing”:

Everyone has in their mind a person they want to be. A person who is perfectly smart, fit, healthy, and empathetic.

But most of us are never quite who we aspire to be. It’s part of the human condition to be never quite satisfied. Being perfect is impossible.

To err is human…

So this one’s only for nonhumans.

Welcome to the nonhuman guide to achieving more.

Disclaimer

This guide, as the title indicates, is not for humans.

I’ll go ahead and say it: maximizing productivity/efficiency is a bad life goal. It’s a quick way to unhappiness, depression, and burnout.

Take these lessons merely as a hypothetical mentality geared towards fighting modern overabundance/super-stimuli and practicing minimalism.

Also, these writings may seem to trivialize disorders/addictions. This is not medical advice. Again, this is purely hypothetical writing, and presents a very shallow view on modern issues. Please take the points with a grain of salt.

So far for the disclaimer.

Let’s dive into it.

The Pitch

We need a new humanity.

Evolution can’t keep up with human advancement. Our psychology isn’t optimized for our elaborate digital lives and modern super-stimuli. The current specialized and globalized society calls for new skills and behaviors.

Our psychological biases hold us back.

Problems like depression, addiction, obesity, and other modern disorders, stem from our inability to cope with today’s world. Really, any modern-day problem is solvable. Because we humans have consciousness. And our consciousness allows us to steer our actions beyond our psychological biases.

But it isn’t easy!

Fighting back deeply rooted human biases puts us head-on-head against our biology. The direct psychological reaction is intense unhappiness and helplessness. But we can beat it!

This guide is about fighting those biases. So we can rise to become something beyond. Kill human concepts of love, comfort, happiness, etc. So we can redefine them and become whatever we actually want to become. Anything is possible when you put your mind to it.

Why?

Say, you have a long-standing dream of becoming an astronaut. Venturing into space has drawn your interest since you were a kid. But very few with such aspirations actually make it into space. What you need is a competitive advantage.

Think of all the things that take up your time. I can envision things like the following:

Fun stuff:

• Children
• Pets
• Hobbies

Food/drinks:

• Buying and preparing food/drinks
• Eating food / drinking drinks
• Doing dishes, cleaning, etc.

Social:

• Trying to be cool to impress people (eg seeking social validation)
• Trying to look cool to impress people (fashion, make-up, etc.)
• Impressing people (sex, non-constructive social behavior, etc.)

Distractions:

• Digital distractions (social media, games, etc.)
• Societal distractions (consumerism, ads, etc.)
• Necessary life stuff (toilet, sleep, travel, some form of exercise)

My theory is: eliminate a subset of the above, free up so much time that you will literally get the competitive advantage to out-compete anyone. Yes, there’s a luck factor as well. But really, if you work ruthlessly towards a goal, you’d be surprised how much progress you can make.

The Goal

Oblivion is the root of all evil.

The goal is to make you aware of counterproductive habits. To make you able to rebel against it. I think many people struggle with the problems outlined in these writings. Fighting against it will require you to withstand discomfort. But there’s a pot of gold at the end of the rainbow.

So what’s in the pot of gold?

• Time. Time is our most precious resource. We need to reclaim our time for what is important to us, rather than all the mindless distractions society throws at us.
• Progress. Our current societal advancements have significantly increased how long we live and how good our quality of life is. We can contribute to make it better.
• Health. The vessel of meat and bones that we call our ‘bodies’ is incredibly important for our quality-of-life.

We can claim back our time.

Are you with me?

Livin’ Like a Puppet

“Advertisements bias us to make decisions that benefit someone else.”

How much of what you do on a daily basis is actually deeply motivated by our own value systems?

In often very subtle ways, we’re steered towards behaviors and desires that are often not in our own best interest.

When you go to the supermarket, is what you buy really aligned with what you would deep down want to eat? Or do you come home with lots of crap that isn’t good for you but you bought it anyway because it makes you feel good?

How much of your phone usage actually improves your life? Does the content you consume make you smarter/happier/better? Are your online connections meaningful the way your offline connections are? How much of the media you consumed last week can you still remember?

What about your stuff? When you think about each of your belongings, do you usually buy things that you need? Do your ‘things’ add a dimension of meaning to your life?

Our current society is organized in such a way that many of our values come from external influences. Organizations play on deeply-rooted human mechanisms to sell their product. Usually, that product gives some kind of ‘comfort’.

In this guide, I want to visit several of these topics and write down how to unlock our own values, and live a life that is meaningful to us.

Digital Entertainment

“Why live in reality, when the perfect reality is right there to be consumed?”

Digital media allows us to live through a life’s hardships, only to find the perfect partner and safe the world. Free social validation for the masses.

Some entertainment genuinely enriches our lives. But with many forms of entertainment, there are diminishing returns the longer we engage in it. And all considered, once the entertainment doesn’t bring us direct value, all hope of it ever bringing value again is lost.

Let’s first zoom in on the pointlessness of entertainment, and how there’s a fallacy where we assign entertainment higher value than it actually brings. Take for instance the incentive systems of most video games. Games reward us points, xp, levels, etc. These incentive systems reward us on a deeply psychological level, which is a part of us that makes us want to feel like we’re accomplishing something in life. The feeling is completely artificial though, since actually we’re just modifying numbers on a screen, that hold no meaning towards life at large at all. Even worse, to get these numbers we go through motions that, for most video games, millions of people have done in pretty much the exact same way.

So why, when the initial “novelty strike” of the fantasy sold by entertainment wears off, do we keep consuming it? It’s mostly FOMO and our minds being unable to distinguish fiction from reality.

The big problem with the “mindless entertainment” is that it distracts us from what’s truly important in life. Replacing entertainment with more constructive behavior can convert diminishing short-term satisfaction with longer-term meaning.

Some of my lessons on this subject:

• Life without any digital entertainment is actually good, and helps us work on building skills that actually make life better.
• When consuming entertainment, you must find it epic.
• When it is not epic, stop consuming. It’s perfectly fine to stop in the middle of a series or video game. Usually, for many forms of entertainment, when the initial novelty wears off, there’s little to be gained anymore.
• If you have been consuming something for a long time, reconsider whether it still brings the value it brought in the beginning. If not, kill the attachment.

Food / drinks

In modern society, we go to a supermarket to get any kind of food we want. Popular are the artificial highly-stimulating foods. Our psychology is optimized to steer our actions towards gathering calorie-dense foods because they keep us alive. But in most of the western society, our survival is guaranteed anyway since we have food in abundance.

Then comes the whole societal obsession around food. People spending hours in the kitchen to get the exact flavor they want. People spending thousands of cash on a bottle of wine because it smells like cedarwood. It’s a completely useless obsession that makes our lives needlessly complicated and wastes away a lot of time and resources.

The world would be better if, in it’s core, we view food as purely functional:

• Our body needs specific nutrients to flourish. On daily basis, we must select the foods that optimize for health.
• At the end of the day, our body does not care whether the food was tasty. It does not care whether we heat our beans and sprinkle a few strands of basil over it, or eat the beans directly from the can. The latter option saves a lot of time.
• Food packaging is one of the biggest environmental destroyers. We need to stop buying 150g of bbq-smoked nuts. Buy the good foods, and buy them wholesale.

Let me share the following highly controversial opinions:

• Any drink but water is completely useless. The whole drinks industry has no reason to exist. Drinking water from the tap is all we need.
• Frozen vegetables warmed in the microwave are healthier and more efficient than fresh vegetables.
• Most food places that are not supermarkets are inefficient. (both from a consumer and producer perspective)

As a society, we have incredibly high standards around food. A hamburger is already incredibly tasty. But no, we just have to refine it with ketchup sauce, two strands of bacon, another patty for good measure, two slices of cheese, and lettuce for decoration. The result is super unhealthy and wildly inefficient food, that will leave us begging for more next time around.

If instead we just snack carrots all day, anything with the least bit of taste will automatically taste extremely good.

Sex

Our sexual motivation system has a huge impact on what we do with our lives. There’s a saying that everything we do is motivated by sex. That might be exaggerated, but it definitely plays a big role in our decision making. But although sex is romanticized by media, there’s a primal nature to it that is easily exploitable and comes with a lot of negativity.

This is a difficult topic to write advice on, because it’s related to very deeply rooted human psychology, and changing this behavior can lead to a bunch of problems. Gearing this motivation system towards a more self-serving system is incredibly difficult. However, self-control can be trained and training it is one of the most valuable modern-day skills.

When training self-control, many people mention “meditation” and “breathing techniques”. Although occasionally useful, I don’t think they are a surefire way to attain self-control. To me, the most effective way to attain self-control is to cut down on super-stimuli.

Distraction

Capitalist society optimizes towards a state that makes maximum revenue for our corporations. Nowadays, that means we live mostly in an attention based economy: mass-manipulating people towards behaviors that benefit the corporate. And all of our human flaws are used against us to the greatest extent possible.

For anyone to whom their digital life is important for their professional success, getting caught in these traps is unavoidable. What follows is a battle of wits, you vs the attention grabbing corporations. Some takeaways from my own battles:

• Some services bring genuine benefit. But when the costs outweigh the benefits for a given service, it’s better to block access completely (take away the candy jar).
• When feeling distracted, find more constructive distractions, like actually being bored. Yes, boredom is constructive. Trying to always fill the void of boredom robs us of valuable restoration time, and time that our subconscious mind can work on generating valuable insights.
• Every source of distraction acts like a snowball: the more we allow ourselves to indulge in the dopamine-evoking behavior, the more our minds beg for more. The earlier the snowball is stopped, the easier it is to stop it. Follow-up distractions are harder to avoid when being historically unsuccessful at withstanding them.
• Many distractions are habits we are hardly aware of. Extensive journalling helps surface these habits and inspect subconscious behavior.

Being attachment free

“The root of suffering is attachment.” ~ The Buddha

Everything we’re psychologically wired towards are mental attachments. Breaking those attachments is hard, but ultimately it renders us more resilient in life.

When we engage in bad habit, we reinforce them. When we engage a lot in sexual behavior, eat tasty foods, binge mindless entertainment, and scroll social media, we strengthen pleasure seeking connections in our brain. Strengthening those connections creates attachment to such pleasure, which is very difficult to break free from.

On the other hand, if we strengthen connections related to focused work and deep relaxation, it will become easier to engage in such constructive behavior, leading to a more difficult but more meaningful life.

Withstanding Pain

“Being resistant to discomfort is the most valuable 20th century life skill.”

Almost all of society’s traps stem from our human tendency to seek comfort. There’s comfort is abundance. But in these sources of comfort, there’s a trap: the more we engage in comfort, the more uncomfortable life becomes. Comfort is addictive, after all.

Having a high pain threshold is the most important modern-day life skill. It automatically keeps us away from sources of comfort, since we’ll no longer need them, dodging most of the issues highlighted above.

You can train resistance to discomfort by, you guessed it, making life less comfortable. Eat less tasty food. Engage in physically hard activities. Whenever your brain tells you to stop, continue. Before you know it, limits that you previously had fade away, and your life transforms for the better.

Don’t Be a Robot

So, let’s go back to the disclaimer I stated in the beginning. With all we’ve learned, why is it so bad to cut all these things from our lives that don’t really help us towards our goals?

We’re not robots. We need some meaningless comfort sometimes. We cannot always be “on”. We need sources of comfort intermittently.

Trying to live an efficient life has been a massive struggle in my life. There were some valuable lessons to extract from it, but ultimately these things come with massive drawbacks.

So instead of turning your life 180 degrees around, consider experimenting with small changes.

Things improve? Great, keep doing that!

But do you notice that actually good food is very important to you? Congratulations, you have discovered one of your core values. Put it in a trophy case and display it proudly. Just don’t let it pile up, will ya?

In the Death Note anime, the Death Note gives the protagonist the ability to create a totalitarian justice system. That is, if there isn’t a force to stop him. But a character named L quickly catches onto him, and makes his ability to “serve justice” harder. In other words, the Death Note is a powerful tool, but opposing forces hamper its ability to affect the world.

At once, the Death Note made the protagonist extremely powerful.

And as a rule of nature, an opposing force arose.

In nature, many processes evolve to chaotic situations unless constrained by opposing forces (Entropy Theory). Any species seeks dominance, until another species comes along to keep populations under control. Physical processes gather entropy till we step in and reduce it (hence why I have to clean my room every now and then). Even psychology shows these patterns, where self-serving thoughts radicalize, giving rise to wide-ranging mental issues, till a force stops them.

Chaos needs to be constrained for systems to continue operating, or the system itself will break down. So, whenever excess entropy is created, an “opposing force” must arise to constrain it.

Let me share an example.

When nuclear weapons were discovered, the numbers of such weapons quickly increased and there was a constant tension of nuclear war (the chaos situation). To fight the chaos, treaties were signed, rules were introduced, and there was a global nuclear disarmament. We can see the effects in the reduction of nuclear close calls.

The takeaway? Don’t panic.

Whenever a situation goes out of control, there are always forces that stabilize it. Sometimes, this stabilization process does come with losses. But a change of tide will come around.

In my mentality, hopeful optimism is something that keeps me going. And it’s comforting to know that nature is on my side.

That doesn’t mean we can sit around and wait for change to happen. We are the stabilizing force. We have to make it happen. But when enough stabilizing forces unite, any form of chaos can be tamed.

Adapted from a post about AGI on LessWrong: https://www.lesswrong.com/posts/We8GnLvDsyucEmGkP/agi-s-opposing-force

I used to work for a startup.

Initially, when building the company, they built a bunch of services for the different parts of the company. In total, they were around sixteen services to support the different concerns in the company.

Each of the products had its own team, responsible for the code belonging to the product.

The company grew from startup to scale-up. In the process, the products grew and grew.

The company hired and hired, till teams grew to >10 engineers. It became inefficient. So, within the team, we defined “squads”. They were smaller teams that were responsible for a part of the bigger system.

Naturally, it didn’t work.

All squads were maintaining a single system, but the codebase, runnables, and database were still shared. Given all these shared components, information still had to flow everywhere. The transition to squads had just turned us into a poorly communicating team rather than a more efficient vehicle.

When those pain points started showing, many fires emerged. More bugs made their way to production. Many meetings were held to discuss architecture, but none of the proposed changes significantly improved the process.

In the end, the company took the long route by breaking down the system. Massive architectural changes were made, and refactoring efforts to break down the old system still persist to this day.

So, whenever I hear about “squads”, I consider it an anti-pattern.

You should not break down a team without breaking down the domain.

The more teams depend on each other, the more inefficient the company gets. The more hierarchical a company gets, the greater the communication overhead.

Defining proper boundaries to systems can fix this issue

It may be a painful sacrifice at first, but will pay off big-time later on.

If you notice that more and more of your workday is spent on things besides the actual work, bring it up and make things happen!

Slavery is bad.

Almost anyone would agree with that. Yet, a meager two hundred years ago, people may not have perceived it as such.

It brings into question: what do we find acceptable today that in two hundred years we will look at with disdain?

My guess: capitalism.

The way society is currently structured, money is a basic need. Generally, to get money, we need a job.

85% of people are unhappy with their jobs:

Most people just work for the money. And if money was a fair way to distribute workload in society, that would be fine. But it is not. There is a lot of unfairness around money.

The main problem with money is that it serves the people in power.

In capitalism, people are usually paid the minimum amount the employer can get away with. Which is, usually, a totally unfair amount compared to the value of the delivered work. Most people work to fill other people’s pockets.

Another issue with capitalism is that anti-social behavior is often rewarded. Point a gun at someone on the street and mug their wallet? $200 for you. Scam an old lady out of her retirement fund? $5,000 for you. Outsmart the market through High-Frequency Trading? $100,000 for you. Well done. I hope you’re happy.

Note that I’m not against capitalism.

Capitalism is necessary.

But AI is here to change that.

With the current trends of AI, it will likely become better at most human cognitive tasks within five years. Give it another ten years, and it will have fully penetrated the physical domain, coming for all jobs that require manual labor too.

Getting rid of jobs is a good thing.

But only if we get rid of capitalism as well.

I envision a world where people have more autonomy over choosing what they want to do, rather than what the world needs. I don’t know what should replace capitalism. Probably there will be a hybrid of capitalism and socialism for a while.

The transition period will be painful. Lots of protests will be held. But, as per my prediction, at some point capitalism will not work well anymore for a rapidly changing society.

And what happens afterward is beyond anyone’s imagination.

My life in San Francisco has led me to meet some of the brightest minds in AI research. Through many discussions, I’ve become increasingly convinced that the next ten years will present a pivotal moment in human history through advances in AI.

The Future of AI

My prediction: Technology will become the main driver behind most processes and systems on Earth.

I used to be quite skeptical about the speed of technological advancements. I did not quite understand what it would mean for AI to be smarter than us. In my vision, computers have always been smarter than us. I cannot do five hundred math equations in a second, but my phone can! But that doesn’t mean my phone will take over the world.

My stance changed when I critically examined the history of technological advancements and then extrapolated it into the future.

As a Software Engineer, I’ve regularly visited companies to see which employees we could most easily automate away. For example, consider an employee responsible for generating weekly financial reports for a company. I could observe their work, see which logical processes they go through to compile the report, and write a program to automate it. It’s an easy way for a company to save money.

Over the years, I’ve become faster at writing those scripts through improved tooling, and thus have a greater ability to automate processes that are currently done manually. ChatGPT gave me a huge productivity boost, allowing me to create software faster than ever!

Currently, we’re already seeing that AI is coming for lots of jobs. Writers, artists, musicians, etc. AI can do a similar job but cheaper. So, over time, fewer such positions will be necessary.

OpenAI’s announcement of real-time video chat with their GPT-4o model calls into question the future need for teachers, call centers, and a number of other social jobs.

Till recently, I was quite skeptical about AI taking physical jobs. Compute power is cheap, hardware and robotics usually come at a higher expense. But if we think about a longer period, like 5–10 years, it’s not totally unreasonable to envision AI playing a big role in that domain too.

As time goes by and technology becomes more sophisticated, it will take job after job until there’s hardly any human intervention needed.

AI vs Humans

My prediction: For the foreseeable future, AI will not possess human-level capabilities.

Sure, it will exceed humans in most cognitive tasks. But there are many tasks at which humans will continue to deliver logically distinct results.

AI is a sophisticated computer program that functions primarily as an input-output machine. Given an input, it maps it to a vector in the neural network and produces an output. By adding actuators, it can cause real-world effects, but it ultimately relies on logical processes and hardware.

Humans are an all-in-one flesh-and-blood package. The way our brain is “trained” is vastly different from how AI models are trained. AI can write text, but I, as a human, can write down my actual emotions, which are the result of my unique journey through the world.

From the outside, the difference can become incredibly subtle. But it is a difference that very much matters. AI is modeled after us, not the other way around. Unless we clone ourselves at the atomic level, AI won’t be a ‘better version of us’. It’s a different class of cognitive machine, with its own strengths and weaknesses.

“My name is Monkey D. Luffy. And I’m gonna be king of the pirates.”

If you haven’t seen it, I can recommend Netflix’s live adaptation of One Piece. And it wouldn’t have been quite as good, if it wasn’t for this silly chap right here:

The character, Luffy, is incredibly cheerful, which is extremely contagious throughout the show. There’s a sentiment behind the silly things he says, which makes me happy.

“I don’t do complicated.”

This is a sentiment Luffy repeats several times throughout the show. And I agree. Far too often do people say “it’s complicated”. Every complicated thing has an easy explanation. And familiarizing ourselves with that easier explanation can take a lot of weight off things.

“What is your dream?”

In the series, Luffy meets strangers all the time, and this is what he asks them. It’s such a good and positive question. And the answer tells a lot about a person.

“You should never give up on your dream, or say it’s stupid.”

Purpose (ikigai) is important. Striving after a purpose can be very fulfilling, and is the ultimate evaluation of ones values. Once your values change, your purpose may change with it, and that’s the ultimate character development.

Positivity goes a long way.

Sometimes, when life feels overwhelming, I tend to reduce life to the simplest positive narrative that helps me continue pursuing my dreams.

For me, it’s a narrative of learning about the world and the people I meet along the way. Side-quests like work and exercise give me the financial and physical means to live that life.

In One Piece, it’s a narrative of finding the One Piece and becoming King of the Pirates. Even though Luffy acts silly all the time, he is constantly focused on his goals.

It’s happiness without indulgence.

Which, to me, is the most valuable kind of happiness imaginable.

I have a long history of maximizing productive hours in my week.

And now, after at least ten years of one productivity system after the other, I may have finally found one that works for me. And it involves scheduling meetings with people and then silently staring at eachother in 50 minute sessions.

Let me explain how I got here.

Dreaming away 😴

I’ve always struggled with attention.

My parents often remarked that there was a five-second delay in any communication with me, usually because my mind was preoccupied with other thoughts.

In high school, I noticed that my mind’s tendency to wander caused me to miss significant portions of classes. It felt as though several minutes had passed without me being aware of what had transpired in the interim or where the time had gone.

Although I was tested for ADHD, I was not diagnosed with it. I can pay attention when I need to, but it’s very easy for me to dream away when I don’t.

A minor gaming addiction 🎮

It was on a Christmas morning, I think around the age of eleven or twelve, when I woke up to find a bunch of presents under the Christmas tree. In one of the packages I found a Gameboy Advance with the game Harry Potter and the Philosophers Stone.

What ensued was a five hour non-stop gaming session.

After that, I gradually got more obsessed by video games. It helped me find a passion in Software Engineering, which was good, but gaming kept claiming large portions of my time. Till I realized something:

“Every second I pour into these video games comes with no benefit to my life whatsoever. The stats I gain, the achievements I get, they are just worthless bytes on a memory disk that signify nothing more than the motions that millions before me have already gone through. Being a ‘good gamer’ is a worthless skill in life.”

That may sound obvious, but it was a big realization for me. Games do really trick your mind into having the feeling that you’re doing something important/productive. Games are designed to be as addictive as possible by making us feel like we’re getting better.

And this lesson extends to many other things as well: social media, digital entertainment, capitalism, etc.

Now, it was up to me to get to the root of this problem, and fix it once and for all.

Infinite digital tasks 👨‍💻

The real problem is a lack of boredom.

That’s a strange problem, right?

It’s normal to be bored sometimes. Sometimes there’s just… nothing to do. Friends don’t want to come over, you’ve read all the books you rented from the library, and you don’t really feel like playing the piano right now either.

Boredom ended for me around the age of fourteen.

I found out how I could extend Minecraft through code, which opened up a realm of infinite possibilities. It led me to spent countless hours coding. But with spending so much time in a digital world, also came the discovery of another side of the internet: an infinity of digital entertainment.

This made me realize the need for minimizing distraction. And that started a whole new fight.

Me vs The Monkey 🐒

By trying to force myself to be productive, I came head to head with the Instant Gratification Monkey:

The more I forced myself to get as many productivity hours out of my day as possible, the more the monkey struck back. This escalated quite rapidly into several hour long sessions where I would be barely conscious, going through YouTube autoplay or scrolling Reddit.

The misalignment between my goals and actions led me to get very angry at myself. As a result of this anger, I tried one productivity system after another. Each rendered me more productive, but I would still have occasional episodes of severe distraction.

Blocking distractions 🚧

All of my devices block a lot of websites.

Doing that was necessary for getting rid of the muscle memory of typing ‘youtube.com’ into the navigation bar and losing the next hour of my life to mindless entertainment.

It worked… to some degree.

The worst time sinks were dealt with by blocking them. But I would still seek out distractions. If it wasn’t YouTube, I might just obsessively scour through Wikipedia. Or I’d just search for “memes” on Google Images.

I noticed that the more I tried to force myself away from distraction, the more vicious the monkey became. Where before I’d lose an hour on a distraction, now I could pull an all-nighter and not knowing where time went afterward.

I thought the problem was related to productive work not being enough rewarding. So I coded a system called Incentivizer.

Incentivizer 🍔

I made Incentivizer three years ago, since when I’ve been using it quite consistently. It’s a website, optimized for mobile use, which looks like this:

The idea is: I define categories in my life that are important to me: personal development, sports, work, etc. By completing goals in those categories, I get points. There’s also a list of “anti-goals”, which subtract points. Points left over at the end of the month convert to rewards: snacks, gaming, youtube-use, etc.

The reason I liked to build my own system for this is because it allows me to collect a lot of data about how I use it and analyze that.

Over the three years, I’ve used it quite a lot. In total, I completed 1990 goals, wrote 495 diary entries, and completed 1721 of what I call “counters” (they are basically personal metrics that I track).

And although the system was occasionally of some value, it never quite worked.

The main problem with the “point system” was that it didn’t really prevent me from distraction, it would just harm me for it. Like you see in the above screenshot, my points are far in the negative. Even while I reset my points every now and then, and even decided to tone down the penalty given for distraction, I would always end up in the negative, which would just discourage me.

Nowadays, I continue using the system to track personal metrics and write diary entries. I don’t use the “goals” part of it anymore.

Focusmate 👥

I’ve been involved in a startup. The system is now live, but every weekend the CEO and I come together to discuss and implement fixes/improvements to the system.

He will tell me about ways the system can be improved, and we discuss it while I code a solution on the spot. During those sessions, I‘m insanely productive. Something about working on-the-spot on screenshare and having the main stakeholder readily available to bounce ideas off appeared to be a golden grail to achieve consistent productivity.

In a call with a friend, I was telling them about this. How I struggle with productivity, except when I’m on a call with someone.

They suggested Focusmate.

I’m generally quite skeptical of online tools. There are many tools out there, but using them often increases overhead at only minor advantage.

The pitch of Focusmate felt initially weird to me. You sit silently in a call with someone for 50min, with minor interaction surrounding it.

But boi did it work.

The social interaction surrounding Focusmate sessions, and generally just being on camera with someone who is working as well, appears to work great.

Through Focusmate, I’ve met other ambitious people, which are exciting for me to meet. I’ve realized it might even lead to lasting friends, which is amazing.

Initially, I only scheduled Focusmate sessions on the spot, whenever I felt a need to boost my productivity. Lately, I’ve been scheduling my week in advance:

The week with the above schedule is incredibly high intensity. But boi do I achieve a lot. Which makes it worth it.

Conclusion 🫠

A part of me feels sad for being locked up in my room behind my laptop so much.

But if I use my time efficiently, at least I can make it count!

Throughout the last ten-ish years, I’ve tried many different strategies. Usually, their effectiveness doesn’t last forever. But throughout it all, I’ve achieved a great bunch.

If you haven’t tried it, I can recommend to give Focusmate a shot!

Who knows, we’ll see each other there :)

I’ve said it before: AI-generated text often sucks.

It’s just so boring! It’s generally incredibly verbose and monotone, with lots of “filler words”.

And now I’ve realized why: these models were trained on boomer text.

Back in the day, people would actually pick up a novel and read it start to finish. Finishing 700-page books was monthly matter for my parents.

These days, kids demand nothing but fast and snappy content. Books turned into YouTube turned into YouTube Shorts. Content that was once viral is now impossible to comb through. Writing rules that were once established fail to hold our attention, or were completely proven ineffective.

ChatGPT can hardly tell the difference.

And this makes me doubt: if ChatGPT can hardly generate content that keeps up with the people currently consuming it, how will it evolve towards future generations? Can we predict trends in the evolution of human cognitive reasoning? Or will it take a much longer time for language models to correctly model parts of human psychology?

Anyway, till then, I’ll sit out on the street with my poster saying “AI text is boring”:

Change my mind.