← Back to Blog

Technology12 min read

Don't Panic - The Fear of Falling Behind

Three practical steps you can take when technology feels like it's leaving you behind.

An exaggerated version of space, with the text "Don't Panic" in the foreground.
A psychedelic version of the Don't Panic mantra, referenced as a core theme in The Hitchhiker's Guide to the Galaxy.

Like the central theme in The Hitchhiker’s Guide to the Galaxy, let me start by saying: “Don’t Panic.”

Reading tech news can be depressing

It’s become apparent to me how similar the current tech landscape is to the key plot points in that book, and how I find Marvin ever more relatable - I think the overarching feeling of things spiralling out of control is something felt by many in the tech world right now. The pressure to ‘keep up’ with advancements, especially in AI, be it agentic workflows or how to contain potentially rogue AI so it doesn’t turn into Skynet, has become a constant in our lives and, honestly, it’s very tiring.

So, grab a seat, take a ride with me for a moment, and don’t forget your towel.

“It didn’t scare us so much because we were part of it.”

I recently read a conversation in The Pragmatic Engineer (opens in new tab) newsletter between Martin Fowler and Kent Beck, which discussed technology shifts of the past, such as the internet, or growth of object-oriented languages, and how they were similar examples of the rapid rise in technological progress we’re seeing now, albeit with one clear difference: the speed at which it is being adopted. In it, Fowler mentions “[it] scared a lot of people. It didn’t scare us so much because we were part of it.” 1

I found that an interesting idea, and the psychological impact of not understanding something can conjure a level of anxiety that is hard to describe.

If the AI marketing monster is to be believed, it’s become increasingly easy to fall behind and that feeling can be exacerbated if all this started before you had a proper opportunity to be ‘onboarded;’ perhaps you were out of work when things went into overdrive, or the company you work at can’t afford to implement it / doesn’t know how / hates it.

That fear comes from being on the outside. The fix is often to change that, make it less scary by understanding it better, whatever level you’re currently at.

Whatever your reason for potentially being out of that loop, I hope the following steps will help to calm that anxiety and provide some practical ways to overcome the fear that comes with where technology finds itself right now.

Step 1 - Don’t try to keep up

Perhaps this is counter-intuitive and seemingly undoes everything promised in the title of this post. I don’t mean this in the ‘ignorance-is-bliss’ sense, but as a way of how to approach it practically.

I don’t believe it’s possible to stay ahead of all the advancements happening right now; there are too many new tools, processes, theories, and approaches for anyone to keep up with. To try would likely cause more anxiety.

What I mean by this depends on what level you’re already at, and I’ll keep the focus on AI right now because broader tech casts the net even wider.

Before jumping in however, the following assumes you’re not so averse to using AI that you would rather cut off your own foot and throw it at the Altmans, Amodeis and Zuckerbergs of this world. Believe me, I understand that feeling. If you simply feel stuck, however, and don’t know where to begin, this is for you.

“Just play around and assess the output from there.”

Just starting out?

a. If you’ve barely interacted with an AI, then your starting point isn’t chasing the cutting-edge, it’s to just use the technology. Use anything, the starting point doesn’t really matter.

If the fear of the unknown has prevented you from experimenting so far, facing that fear is the first step. The core ‘leaders’ all have their strengths and weaknesses, but starting out, none of it really matters. Get it to smarten up that presentation you’ve been poring over for the past week or ask it to develop a dashboard for the amount of times you get irritated by your colleagues, what your response was, and how to help you manage your anger issues.

Just play around with it and assess the output from there (please don’t actually use it as a therapist though).

b. If you can’t think of anything to ask an AI, ask it to provide some ideas for you.

Sometimes the best way to prompt an AI is to have another model write the prompt for you, with the caveat that you should really read and understand the prompt you’re feeding it, just in case, you know… Skynet - this is a joke, of course, but the advice stands; if you don’t understand why something has been added to the prompt, ask for its reasoning first before using it. Take out anything you’re uncomfortable with or do not fully understand while you get used to the mechanics of it all.

c. You don’t need to pay for subscriptions immediately; if you’re in that experimentation phase, just sign up for a free account and give it a try.

Don’t throw any personal information in there, don’t use anything that is owned by anyone else (yes, I understand the irony of that, considering what most models are trained on, but better to be safe), and try to have fun with it. Just be ready to hit the limits early and don’t expect to be super productive with it early on.

A few good ones to start with include:

These have a low barrier to entry and will generally give you an idea of how things work and what the output can be.

“Understanding how these tools extend what you can do is the most important part.”

Are you mildly comfortable?

If you have used AI to chat through a problem, create documentation, or already understand some of its limitations and haven’t stepped into the realms of MCPs or RAG pipelines, that’s probably your next step. Alternatively, you might have some ideas on a digital product or website you want to build and are unsure where to start.

Again, what company you’re signed up with or what model you use still doesn’t matter massively at this stage. Understanding how these tools extend what you can do is the most important part. The following are still quite ‘black box’ in terms of understanding what’s happening under the hood, but are decent for building out quick prototypes and / or working applications:

It’s also worth looking into the MCP connections for your favourite tools, to see how they can speed up certain processes. Try connecting an LLM to your task management tool of choice, for example, and have it update tickets after fulfilling certain criteria, or adding a bug ticket when it comes across one in the code. Just note that some tools might require a paid plan before you can use them.

So, you’re a subscriber?

This section is targeted at subscribers because the likelihood is, if you’re beyond the previous two categories, you are subscribed to at least one of the major ones and perhaps use a few others on the side. You’re already using it to organise files, create new ones, automate certain workflows around research and development; basically, where letting AI work directly on your files is your comfort zone. Agentic workflows and related tooling for more advanced setups is probably on the agenda for you next.

Some additional tools to consider reviewing could be:

Already using it professionally?

I’m not sure why you’re reading this; maybe get an AI to summarise it as part of your daily digest and move on. You have possibly already considered local models but perhaps that’s something to invest some time into if you haven’t already? Just for the fun of it.

Go and optimise your AI harnesses and get those reinforcement loops working perfectly.

Step 2 - Focus on what is relevant to you

This also might seem obvious but many people I have spoken to, and this applies for both individuals and businesses alike, appear to want to run before they can walk. A reference to the sperm whale falling from the sky (voiced by Bill Bailey in the Hitchhiker’s Guide to the Galaxy film) seems appropriate here: many feel pressured to come up with a long-term AI plan in the shortest time possible, in order to ‘catch up’ with others.


For those unclear about the whale analogy - in The Hitchhiker’s Guide to the Galaxy, there is a passage in the book dedicated to a sperm whale that is suddenly born high in the sky, above the planet. As it falls, it tries to make sense of the world and its own existence, not knowing its demise is coming just as quickly. It’s a commentary on the randomness of life, and the chaos we need to manage in our daily lives.


I think attempting to compress your own timelines for the sake of speed alone is the wrong approach, because it often means you’re looking more at what your peers or competitors are doing; the tools, the processes, the approaches, and not validating whether it’s particularly relevant to your goals. Invariably, that means all you catch up with is the ground, just like the sperm whale, hitting a very immediate stop because the plan you have doesn’t work for you or your business.

The better approach is to have a clear map of where you are, and where you want to get to. What are the goals and milestones that lead you there? Once you have a clear picture, reviewing your current infrastructure and processes is next. Are they fit for the purpose of achieving your milestones and goals, or does something need to change?

“It’s better to take longer and ensure what you’re building is right for you or your business.”

If the answer is something needs to change, it’s possibly a good candidate to see how rebuilding it with AI in mind could help. I don’t want to conflate AI with automation here though; many tasks don’t need an AI in the loop, automation is fine. Where numerous tasks need to be orchestrated simultaneously, or chained together with reasoning, that’s where AI can help with decision-making, so long as it has clear boundaries.

This is all to say: don’t use AI for AI’s sake. Really understand your workflows, your pain-points, and bottlenecks first, and don’t feel like you are time-limited in assessing that. It’s better to take longer and ensure what you’re building is right for you or your business, rather than rushing through it and spending a lot of money on something that brings zero value.

Step 3 - Constantly review the output

That brings me onto the last point: measure, measure, measure.

Measurements, KPIs, metrics; whatever label you want to give it, you need to have a clear idea on which ones are important to you before work even begins. This is especially true in teams and businesses, but even as an individual, it’s useful to have an idea of what you want to get out of it. Even if it’s just to understand the tools better.

Similar to the Hitchhiker’s protagonists visiting the Restaurant at the End of the Universe, this isn’t really the ‘end’ of the process, it’s where it starts; have those measurements in place before it’s in use and review them incessantly.

Just because a workflow or orchestration looked fine on day one, doesn’t mean that it’s still performing well on day 30. The same goes for personal goals with learning modern technology. It’s a part of your daily regime now, just like eating is.

“Nobody expects you to get it right first time.”

You might find the numbers are positive, everything is working perfectly, and you are on track to hit every milestone and, eventually, your core goal, unhindered.

More likely though, something hasn’t been optimised, or the initial implementation of some system or process doesn’t work as expected. Those are the reasons we monitor progress and iterate over time.

Nobody expects you to get it right first time, and the sooner we get comfortable with failing, (or what I prefer to call ‘falling forward’), the quicker positive outcomes will follow.

Don’t be afraid to try something or experiment but do plan for it not to work; it could be a gate that stops code from being deployed if the tests fail, or an easy way to rollback if a build somehow breaks in production.

This is the same order I have worked through most problems in my career, and something I have looked to recreate for the Delivery Diagnostic: reviewing where a team is now, what’s actually slowing it down, and how they’ll know it’s performing better.


Please note: I’m not affiliated with any of the tools listed here, I’m just offering some tips based on my own discovery process. If you have some time and really want to dive in, check out this list (opens in new tab) - it’s a lot to take in!

Footnotes

  1. The original conversation was documented in The Pragmatic Engineer, Cycles of disruption in the tech industry, 7 Apr 2026. It’s a conversation between Kent Beck and Martin Fowler, written up by Gergely Orosz, see source link (opens in new tab). ↩