The AI Panic Attack

Written February 16, 2026

Note: I do not intend to make light of or be glib about mental disorders. Everything in this post is drawn entirely from my own experience, though I expect many readers will relate.

This week I attended The Pragmatic Summit1, which was a wonderful event I will highly recommend to others if it is held again. The thing I most appreciate about it was never feeling like I was being sold something, whether it was a product or a hot new trend. Even more unusual was that much of the discussion wasn't even about the technical aspects of software engineering, but rather understanding the moment we are in with the emergence of AI tools, various predictions for how the future might plausibly play out, and importantly how we are handling the uncertainty and disruption that is already happening today. I left feeling noticeably calmer than I have in months, and brimming with the kind of ideas that come naturally when they are not stifled by the twin devils of anxiety and insecurity. On the long drive back home from San Francisco, I had some time to introspect about why my mood had changed.

On the surface there is a simple, ready answer to this question: humans — even introverts like me! — are social creatures and having engaging conversations with like-minded people is biologically hard-wired to improve our mood. I do miss some of the togetherness that has been lost these last few years working remotely from home. But that isn't all of it; in hindsight, I have felt a heightened level of strain and stress in the last few months that, while not reaching the extremes of a mental health crisis, have similar characteristics.

When your brain turns on you

So, let's digress for a bit and talk about panic attacks. In 2009, for no clearly identifiable reason and with no specific trigger, I had my first panic attack. For anyone who hasn't had this experience, it might help to imagine falling out of a plane without a parachute: your heart rate skyrockets, your palms get sweaty, you start to hyperventilate, your vision narrows, and your brain focuses entirely on grasping for any way out of your desperate situation. Except when you look around, you're in your living room or driving your car, and there is no situation to get out of. Everything external to you being normal, the remaining possibility is that something is wrong with you. The first time this happened I was convinced it was a heart attack, and spent the night in a very expensive bed in a hospital. The eventual diagnosis didn't help at all with future attacks, because as it turns out, it is incredibly hard to convince yourself there is nothing dangerous happening when every organ in your body says otherwise.

My model for understanding the logic of a panic attack is a feedback loop: An unremarkable annoyance that a healthy person might quietly ignore, such as getting a poor night of sleep or sitting in an uncomfortable position for too long, snowballs very quickly in the context of a panic loop until it reaches the End Now, Or Else phase. This creates a sense of urgency, which then heightens sensitivity to any inconvenience or obstacle: "I am tired at work" becomes "I need to get home NOW" becomes "this traffic needs to stop NOW" becomes "I need to get out of the car NOW" and of course each step becomes more difficult because you can't realistically get out of the car in the middle of traffic and at this point the part of your brain yelling "NOW!" and the part yelling "WE CAN'T!" go to nuclear war with one another.

Luckily for me with some medication, the support of my wonderful wife and family, and a lot of time these crises reduced in number and intensity and are no longer a regular part of my life. However, the experience has made me hypervigilant for this kind of self-perpetuating mental loop and how it leads to irrational emotions or thoughts.

AI irrationality?

There is undoubtedly a marked increase in lowercase-A anxiety going around the software industry in response to agentic LLM coding tools since last summer, and it isn't hard to understand why: while proclamations of the death of software engineering as a discipline and as a career strike me as premature, it is impossible to deny the rate of change that is happening almost everywhere at once. There is a palpable sense that those of us who make a living writing code face a threat to either our employability, or our established skillset, or both. I read articles and social media posts that skew either implausibly utopian or unnecessarily fatalistic with regard to where this technology is headed, but not a whole lot of discussion2 about the emotional impact this is having today on those in the industry or those who have invested years training to become a part of it.

I myself feel the same general unease that honestly is expected during any large and unpredictable change, but in addition there is a feeling of having my back against the wall: the claims of 10x (or more!) speedups and the counter-claims that actually teams leaning into AI coding might be delivering more slowly leaves me uncertain how much of my time to devote to using and mastering these tools. Even when the tools work, there are organizational bottlenecks: everyone has a different risk tolerance, there are many opinions about how to ensure code quality, and no one wants to review "AI slop". I don't want to be too far ahead of the curve, recklessly exposing my team to as-yet-unknown risks, but at the same time I am afraid to be left behind and stagnate in my career.

This is exactly the ratcheting effect I believe leads to capital-A Anxiety and panic. I see myself spending more and more of my time trying to see around that corner to anticipate where trends are taking us, while at the same time watching my own behavior carefully to avoid falling into "AI psychosis", where I would fool myself into believing I can somehow foresee the optimal way to use these tools and start trying to convert others.

I am losing sleep.

I am tempted to skip work meetings to work on my own agent orchestrator, to try every new model and tool, to build a piece of this future for myself.

It is very easy to lose sight of other goals as a result of my own internal sense of urgency to get past this uncertain time as quickly as possible. This isn't purely internal: there is a sense of urgency at work and in the tech community at large, and it is in active tension with the fact that we can't change our systems, or workflows, and our team culture overnight no matter how much urgency there may be.

Does the paper bag work for anyone?

I am grateful for my experiences with anxiety disorder because they have taught me two important techniques to break the ratcheting urgency cycle:

  1. Rely on others in my community. The biggest aggravator of panic is being stuck in my own head, feeling my insecurities and inadequacies growing without being able to give them perspective. It may seem obvious, but simply taking the perspective of someone else dealing with the same problem can make it easier to resolve; it is much easier to be dispassionate about other peoples' problems than about my own. This is why being in a roomful of anxious people actually made me feel better - at least I am not alone! We can solve problems together that feel overwhelming for us as individuals.

  2. Give myself the benefit of time. Perhaps this is a learned behavior to combat procrastination, but I've found that nearly any sense of urgency I experience most days is completely self-imposed and not "real". I keep reminding myself that this feeling of not wanting to be left behind or fall behind the curve is not grounded in fact: if I fall into a coma for a year, will I really be completely out of the loop or even unemployable when I wake up? Or will most of the news between now and next February be completely forgotten and irrelevant by then?

The best gift I can give myself these days is time away from the churn and the head-spinning pace of innovation to focus on my family, my friends, my work, and my side projects. Even if the pace of innovation speeds up in 2026 — and I have every expectation it will — I will be making it a point to slow down and prioritize listening and projecting calm over attempting to influence and accelerate.

My mantra is this: We have time. We have community. We are a smart, if sometimes insecure bunch. We can figure out a better future for ourselves and those who come after us. We'll be OK so long as we don't give in to panic.


  1. 1 The Pragmatic Summit 

  2. One exception is this recent piece from Simon Willison: Deep Blue