Anchored in the Chaos: Managing Cognitive Load, AI Uncertainty, and Well-Being as a Developer

Introduction: The growing pressure on developers

If you feel like you’re perpetually sprinting just to stay in the same place, you are not alone. Across the industry, from junior developers pushing their first commits to seasoned seniors leading large teams, the professional landscape feels less like a steady ascent and more like a chaotic treadmill set to maximum speed.

AI has certainly contributed to this whiplash-inducing landscape: one day we hear proclamations about the end of our profession, the next we’re expected to deliver 10x improvements. For leaders, like Chief Information Security Officers, this is compounded by the tangible challenges of AI’s (mis)use.”

The ever-evolving nature of our position cannot be denied either. Today’s full-stack developer isn’t just expected to handle the front-end and back-end. Nowadays, we are often also juggling infrastructure-as-code, complex testing methodologies, cloud deployment, and continuous security monitoring. This widening scope means that being a software engineer today involves much more than just writing clean code.

This has led to a rampant increase in three key areas that I’d like to tackle in this article:

• Cognitive overload:
  The relentless pace of change, the continuous influx of tools, and the pressure of context-switching across an ever-expanding technical domain can feel utterly exhausting and chaotic.
• AI uncertainty:
  The rise of powerful AI tools, while exciting, has created palpable anxiety about job security and the future role of human programmers. Expectations are also continually shifting, with some viewing AI as the latest silver bullet.
• Personal Wellbeing:
  We often treat self-care as a luxury or an afterthought, yet neglecting our mental and physical health makes sustainable, fulfilling work impossible.

I have written this piece to shed light on how this constant pressure manifests, drawing on my own experiences and conversations I’ve had with others in our community. I aim to move beyond mere recognition and discuss practical, accessible strategies we can use to manage our mental load, reframe our relationship with AI (viewing it as another potential tool in our belt, rather than a source of dread), and firmly establish wellbeing as an essential, non-negotiable component of our professional life.

1. Cognitive Overload

Let us start with something we’ve likely all experienced, but maybe haven’t quite managed to put a name to.

If you’ve ever ended a workday feeling utterly exhausted, not because you solved a tough challenge, but because you spent all day juggling Slack threads, tracking down documentation, and switching between three different codebases, that’s cognitive overload. It’s the feeling of your mental hard drive running critically low on space.

Now, a cognitive load itself is not bad; we just need to be able to properly manage it. It is important to differentiate between the three kinds as each can be tackled in its own manner, and not everyone is as susceptible as certain types.

• Intrinsic load (The hard/fun challenge):
  This is the good kind of challenge. It’s the effort needed to solve a genuinely tricky technical issue, designing a complex microservice or figuring out a tough algorithm. You feel focused, maybe frustrated, but ultimately, you are productive.
• Extraneous load (The friction):
  This is the one that saps our battery. It’s the wasted mental energy spent battling the environment: poorly organised code, confusing meetings, documentation that’s always outdated, random build failures, or relentessly being pinged with notifications. This is the load we have to fight, because it burns energy without delivering value.
• Germane load (The learning):
  This is the effort that actually builds skill. It’s the time spent integrating an architectural pattern or connecting a new framework to an old concept. This is where real growth happens.

Right now, all too frequently, the Extraneous Load is dominating our days. We are continuously swimming in complexity.

For me, this quickly became clear when I became a consultant. When you land on a new client site, the challenge isn’t just coding (Intrinsic Load), it’s having to rapidly absorb their entire world: their acronyms, their specific environment setup, their internal politics, their application stack with its legacy, good and bad. A lot will, of course, depend on the quality of the onboarding procedure and documentation, if present.

You might also experience a certain “context whiplash” when changing clients, which is pure Extraneous Load. After all, how many organizations operate in the same manner?

This makes it vital to manage this mental friction to ensure your mental well-being.

The Hurdle

The continuous influx of new tools, the unrelenting pace of information, and the pressure of constantly shifting focus between unrelated tasks leave us mentally depleted. Our brain is forced into shallow multitasking rather than deep work.

The Impact

What does this look like in practice?

• Reduced productivity: We feel busy, but our output is low. We spend an hour regaining context for every hour we code.
• More mistakes: When our working memory is full of clutter, such as Did I check that email? What was that meeting about? It is impossible to hold all the necessary variables in our mind. Our mental buffer overflows. Mistakes are inevitable.
• Feeling Perpetually Behind: At the end of a full day, we feel like we’ve achieved nothing, which leads to anxiety and stress.

Practical Techniques

While we can’t single-handedly alter how our industry works, we can adapt our workflow to try rebalance things. We can influence our extraneous load to a certain extent to better achieve deep work (also known as flow-state) where time seems to fly, and we feel and are productive.

1. Owning Our Time

In the past, I have all too frequently had the feeling of: What did I actually achieve today? Where did the time go? There are certain things that have helped me out, such as:

• Batching communication: Do not live in your inbox and Slack. Designate specific, short windows, say, 10:00 AM and 2:00 PM, solely for communication. Outside of those times, everything is silenced outside certain channels to keep abreast of time-sensitive matters.
• Use the pomodoro technique: It’s 25 minutes of intense, single-task focus, followed by a 5-minute break. It trains your brain to sustain attention and gives you built-in rest periods.
• Timeboxing obstacles: If you hit a tricky bug, set a timer (e.g., 45 minutes). When the timer goes off, you must either ask for help or document the issue and move to a different task. This prevents one small hurdle from eating your whole day and draining your mental reserves.
• Bullet list: Starting the day off by creating a bullet list of key goals for the day helps you structure your day and refreshes your memory.

2. Prioritising Deep Work and Setting Boundaries

Deep work is where we create real value, which is key for both our organization and our own mental well-being. It’s where we’re at our most productive.

• Schedule mon-negotiable deep work: Look at your calendar and block out 2-3 hours every day specifically for coding, design, or architecture. Treat this block like a high-stakes client meeting; do not let it be interrupted.
• Learn to say ‘No’ gracefully: While it is certainly a good thing to help out when you can, make sure that this does not interfere with your key tasks. Every time you agree to an optional meeting or take on a tangential task, you are saying “no” to your ability to focus on your core priorities.
• Be selective with your learning: Stop trying to read every tech article or follow every new framework announcement. Focus on becoming an expert in the tools you use right now or see yourself using soon, rather than chasing everything new. Just imagine trying to keep up with all the latest Linux variants!

3. Systemising Knowledge Management (The ‘Second Brain’)

Your brain is a processor, not a file cabinet. Let it process and let an external system handle the storage.

• Build your second brain: This could be a bullet journal, a Notion page, or even just a well-organised folder of text files. The point is to have one single, trusted place where you capture all meeting notes, to-dos, code snippets, and random ideas. This is helpful for both you and the organization. Too many of us have joined an organization and received an awkward chuckle when asking about where we can find the documentation.
• Immediate offload: The moment a distracting thought pops into your head (e.g., “I need to order more coffee”), immediately capture it in your Second Brain and then return your focus to your task. This allows your working memory to release that thought. I can highly recommend a Fisher Space Pen Bullet and a Field Note notebook.
• Automate the mundane: If a task is repeatable, automate it. Use linters, robust testing suites, make, and so on. The less mental energy you spend performing and verifying trivial steps, the more you have for creative problem-solving.

2. Navigating Uncertainty Around AI

If managing cognitive load is about fighting the present chaos, navigating AI is about managing the anxiety regarding the future.

The moment large language models (LLMs) like GPT-4 hit the mainstream, two extreme, conflicting narratives took hold in the developer community:

1. The Doomsday Scenario: AI will replace programmers entirely within five years. The underlying fear is that if your worth is defined solely by writing lines of code, you are effectively obsolete because the machine can now do it faster and doesn’t fall ill.
2. The Silver Bullet: AI will resolve all our obstacles instantly. We should now be 10x more productive and deliver impossible features overnight.

Both of these fuel anxiety and extraneous load. They force us into a state of comparison and panic, making us believe we must either master every prompt engineering technique immediately or face redundancy. The reality, as always, is far more nuanced and, thankfully, less terrifying.

For many of us, this anxiety and pressure to keep up isn’t just some niggling feeling, but rather a continuous drain. From my own experiences and interactions with others in the development community, it’s clear that the pressure is being manufactured and amplified by external forces. You have IT Managers and clients who, after reading a short article or looking at the latest Gartner chart, suddenly see AI as a magic cost-cutting bullet. AI is showcased as something that can generate applications from scratch, without ignoring the prototypical nature of these, and ignoring the domain knowledge aspects. This often forces us to integrate unstable solutions purely because of pressure from above, just because it’s the latest buzzword. Many will be reminded of the low-code/no-code wave a few years ago, and well, here we are once again.

Adding significantly to this Extraneous Load are the Tech Influencers and self-proclaimed gurus on social media. They preach that if you aren’t using specific technique X or methodology Y, you are falling behind your peers. Luckily, they are offering their $499 ‘AI Prompt Engineering Masterclass’ at a promotional discount to help uplift others. Their agenda is often transparently self-serving—selling courses and generating clicks. The effect on us is the same: intense, constant stress that we are failing if we aren’t chasing the absolute bleeding edge, and prompting managers to ask: shouldn’t you follow some of those?.

However, this discourse overlooks certain key aspects: AI is only as good as the context and verification we provide. Our enduring value lies in the hard-won skills that AI lacks, namely judgment, architectural insight, understanding ambiguous business context, and taking legal/ethical ownership. If you jot down what your average day looks like, how much do you spend on actual coding versus design, alignment with other parties, analysing what is actually needed?

It became clear to me and many others that the real challenge is not mastering the technology, but mastering our response to the sheer volume of external pressure. It requires a fundamental mindset shift to filter out this noise and use AI responsibly and sustainably without losing our heads.

The Core Issue

AI advancements generate palpable anxiety about relevance, job security, and an overwhelming “need to know everything” right now. This pressure often comes from leadership, who may see AI as a simple productivity multiplier without understanding the nuances of implementation, security, or verification.

Reframing the Story

We need to shift from a position of anxiety to one of agency. AI is not a singularity; it is a complex, often flawed, and fundamentally non-human tool. By viewing it as a collaborative partner, a super-smart, tireless, but frequently clueless copilot we can redefine our role from ‘code generator’ to ‘system architect and editor.’. As someone who’s a keen proponent of rubber-duck debugging, I view it as having a cyberduck next to me.

• You are the pilot, AI is the copilot: The AI can fly the plane on autopilot and handle the routine checks, but the human pilot is still the one setting the destination, checking the calculations, and taking control during turbulence. Your value lies in judgement, context, and verification.
• The AI produces drafts, you produce quality: Think of an LLM as a brilliant intern who never sleeps but sometimes hallucinates facts or introduces security flaws. It can give you a starting point in seconds, but your expertise is required to check the logic, ensure best practices, integrate it correctly, and take ownership of the final result.

Practical Strategies

By framing AI as a tool, we can focus on where our unique human skills provide the most worth, essentially protecting our professional ‘Intrinsic Load.’

1. Focus on Augmentation, Not Replacement

Spot where AI can take over the low-value, repetitive tasks, freeing you up for deep work.

• High-Quality code yields better AI results: AI tools operate much more effectively when a codebase is well-documented, follows clear conventions, and has comprehensive tests. This isn’t just a kindness to the AI, it benefits us all as well. Good documentation reduces your Extraneous Load and allows the AI to provide more accurate, context-aware suggestions, accelerating your productivity without introducing risk.
• Automate the mundane: Focus on AI tasks that help your workflow without touching code directly. For example, generating boilerplate README drafts, summarizing long code comments, or translating a snippet from one language to another to quickly grasp what’s going on. This reduces mental friction and time spent on administrative tasks.
• Use it as a rubber duck: Instead of relying solely on it for solutions, use it to explore constraints, test assumptions, or rapidly prototype alternative approaches. It’s an interactive, tireless sounding board. It is a convenient way to help tick off your mental checklist of edge cases.
• Verification: While AI can indeed generate code, there are two critical responsibilities it is relatively poor at, namely verifying that code, checking edge cases, and ensuring it adheres to organisational standards and that it actually matches what your business users have a mental picture of. These are high-value skills that AI cannot replicate.

2. Double Down on Your Core Human Skills

The value of a human programmer increasingly lies in the areas that AI cannot do. You’ll also notice that these happen to be the key skills that are expected once someone wants to move beyond the junior level.

• Architecture and system design: AI struggles with abstract, long-term system integrity and cross-domain dependencies. Our ability to design robust, scalable systems that solve business demands remains essential.
• Communication and empathy: Explaining complex technical concepts clearly to non-technical stakeholders, translating business requirements into technical specifications, extrapolating what someone actually wants and needs rather than what they state, and collaborating with a team are uniquely human skills.
• Debugging and problem-solving: While AI is good at suggesting solutions, the core skill of tracing a bug through a complex, unfamiliar system, understanding root cause, and applying lateral thinking remains an area it struggles with.

3. Adopt a Growth Mindset

View engagement with AI as essential upskilling, not a competitive threat.

• Experiment Safely: Dedicate a small, non-critical block of time each week to intentionally experiment with AI tools (Copilot, ChatGPT, et cetera). Treat it as a sandbox for learning their strengths and, more importantly, their weaknesses.
• Experiment with the interface (prompting): Learn how to interact with the tool effectively. Understanding how to structure a good prompt is rapidly becoming a core competency for maximising AI’s utility.
• Discuss failures, not just successes: Be open within your team about instances where AI gave a misleading or flawed answer. By understanding the failure modes, you build a shared, realistic knowledge base and reduce collective anxiety.

Ultimately, the best defence against AI uncertainty is becoming the individual who is able to skillfully wield the tool, define the problem, and deliver verified, high-quality solutions.

3. Protecting Personal Well-Being as a Necessity

If the first challenge is about managing the input (cognitive load), and the second one is about managing the future (AI), then the third is about maintaining the key piece that processes it all: you.

In our industry, we often obsess over the specs of our laptops, the ergonomics of our mechanical keyboards, and the efficiency of our CI/CD pipelines. Yet we frequently ignore the critical component, the central processing unit, namely ourselves. We frequently run our own biological hardware on poor fuel, insufficient downtime, and ignored error logs (stress signals).

It’s something I’ve seen happen repeatedly, and something I’m guilty of myself. I’ve watched brilliant colleagues who could solve complex architectural problems in their sleep suddenly hit a wall. They didn’t leave the industry because they couldn’t learn the latest framework or because AI took their job. They left because they eventually burned out. They ran out of capacity.

We tend to treat self-care as a luxury, something we’ll get around to “after this release” or “once the crunch is over.” But to quote Da Vinci: “A work of art is never finished, it is merely abandoned.” There is always another release. Neglecting our mental and physical health makes sustainable, fulfilling work impossible.

The Situation

While those of us living in Western Europe may be better protected by labour laws and a culture that values time off than our counterparts in the US or Asia, we face a more subtle, yet equally damaging challenge: intensity and permeability.

It is rarely about working 16-hour days; it is about the sheer density of the hours we do work. We try to squeeze complex coding, architectural decisions, and back-to-back Teams meetings into a condensed window, often skipping breaks to ‘power through’ or to rush to the next meeting.

Furthermore, hybrid work has eroded the boundaries. The physical commute, which used to serve as a decompression buffer and clear delineator, is gone for many. The result is that while we might physically leave our desks at 5:30 PM, the work follows us via notifications on our personal phones, or the mental urge to “check one last thing” before bed.

This leads to Health Debt.

Akin to technical debt, health debt accumulates interest. You might get away with hacking your sleep schedule or skipping movement for a sprint or two, but eventually, the debt comes due. The cost isn’t just physical illness; it’s a degradation of your cognitive abilities—your focus, your patience, and your ability to solve the very challenges you’re paid to solve.

Shifting the Narrative

We need to shift our mindset from “Self-care is a luxury,” and “I will get around to it later” to “Self-care is professional maintenance.”

Think of it this way: You wouldn’t run a critical production server at 100% CPU utilization for months without cooling or patching, and then be surprised when it crashes. You are the most vital piece of infrastructure in your career. Maintaining your well-being is not “quiet quitting” or “being lazy” but rather a requirement for high-performance engineering.

• You are an athlete of the mind: Professional athletes don’t train 24/7. They rest effectively so they are able to perform at their peak. Cognitive work is physically draining (the brain consumes about 20% of the body’s energy). You need recovery protocols just as much as a runner does.
• Resilience requires reserves: When the “chaos” hits—a production outage, a challenging client, a sudden deadline—then you need mental reserves to handle it. If you are constantly running on fumes, you have no buffer for the unexpected.

Practical Strategies

How do we put this into practice? We don’t need a complete lifestyle overhaul; we need patches and configuration changes.

1. Treat Health as a Non-Negotiable Dependency

If a library is deprecated and causing security flaws, you upgrade it. If your sleep is broken, you need to fix it with the same urgency.

• Prioritise Sleep: It is the defragmentation process for your brain. It is when short-term memory is converted to long-term memory and when neurotoxins are cleared. Coding on four hours of sleep is like coding drunk; you might feel productive, but the quality is usually terrible.
• The Physical API: Your brain lives in your body. Simple things like hydration, standing up once an hour, and getting sunlight in the morning directly impact your ability to focus on complex logic.

2. Define “Done” and Disconnect

One of the hardest parts of our job is that the work is never truly “finished”. There is always another ticket.

• The shutdown ritual: Create a distinct boundary between “work mode” and “home mode.” This could be closing all your tabs, writing your to-do list for tomorrow (offloading the cognitive load), and physically leaving your workspace. Personally, I am keen on a forest walk followed by a nice cup of tea.
• The Right to Disconnect: Be clear and firm about your availability. If you answer Slack messages at 9 PM “just this once,” you teach your colleagues and clients that you are available at 9 PM. Protect your downtime; it is where your creativity regenerates.

3. Seek Community and Perspective

Isolation is a major contributor to burnout, especially in remote work.

• Talk about it: In my experience, admitting “I’m struggling with the workload right now” usually opens the doors for others to say, “Me too.” Vulnerability builds trust and reduces the feeling that you are the only one not coping.
• Hobbies outside of tech: Have a hobby that involves zero screens. Whether it’s cooking, hiking, or, as I mentioned earlier, creative writing using fountain pens and paper. Doing something analogue grounds you in the physical world and gives your visual cortex a much-needed break.

Conclusion: Building an Anchor in the Chaos

We started by looking at three challenges that we as developers encounter all too often: the increasingly challenging cognitive load, the uncertainty of AI, and our often-neglected necessity of personal well-being.

It’s important to realise that these challenges are interconnected.

• If you don’t manage your cognitive load, you won’t have the mental space to learn how to leverage new tools such as AI.
• If you fear AI, you increase your stress, further impacting your well-being.
• And if you neglect your well-being, you lack the resilience to handle either the load or the uncertainty.

However, none of these are insurmountable. By systemising how we handle information, reframing AI as a collaborative copilot (pun intended) rather than a replacement, and treating our health as critical infrastructure, we will be able to stop flailing in the current and start swimming.

You have the agency to drop an anchor. It doesn’t mean the storm stops, but it means you stop drifting. It will enable you to weather the storm, regardless of what the next tech cycle brings.

Helpful Resources

Navigating these challenges might seem daunting, but I’ve found that a shift in mental mindset is key. Beyond the earlier recommendations of conferences and meetups, certain books have been instrumental for me, and I hope they will be for you too.

It does behove me to say that this is far from an exhaustive list, as there are so many amazing resources. So please do reach out with recommendations or if you’re looking for something specific.

1. The pragmatic programmer by David Thomas and Andrew Hunt
   To me, this book was a lesson in simplicity. It helped me realise that writing adaptable, resilient code isn’t just about ‘best practice’, it is about reducing the mental friction we face every day. Adhering to the KISS principle (‘Keep It Simple, Stupid’) is an act of kindness to my future self.
2. The programmer’s brain by Felienne Hermans
   Do not let the title of this book deceive you; the content is very much applicable to everyone’s brain. Felienne Hermans dives into the cognitive science of how we process information and how we are able to facilitate this both for ourselves and others. It helped me understand why some code is so taxing to read or debug, and gave me practical techniques to work with my own cognitive processes (and quirks).
3. How to make your brain your best friend by Rachel Barr
   A compassionate and insightful guide to the mind. This book reminded me that sustainable living isn’t about fighting my own psychology, but about understanding and nurturing it. It’s a foundational read for wellbeing. Small acts of kindness truly matter.
4. Four thousand weeks: time management for mortals by Oliver Burkeman
   One of those reads that led to quite a paradigm shift in how I viewed productivity and the value of time. It’s not one of those traditional self-help books, which just try to cram more into one day. Instead, it offers a liberating philosophy on embracing our finite time. It helped me (somewhat) step off the hamster wheel of efficiency and focus on what’s genuinely meaningful.
5. The bullet journal method by Ryder Carroll
   As developers, we spend our lives on screens, but I found that the best way to clear my head was often analogue. This method gave me a structured way to ‘dump’ the mental clutter onto paper. It’s been quite helpful for quieting the noise and helping me focus on what actually matters, rather than what’s just shouting the loudest. My love for fountain pens certainly also helps.
6. Rewiring your mind for AI by David A. Wood
   With so many loud voices debating AI, this book helped me ground the conversation for myself. This book helped me reframe that anxiety into curiosity. It moves the conversation from fear to collaboration, showing how we should view these tools as powerful teammates rather than replacements.

Remember, investing in these areas isn’t a distraction from your usual upskilling. They help make sustained, meaningful progress possible. I wish you the very best on the journey.