Sorry — Can you hold, please? One quick second while I dust off this here blog, writing something for the first time in a very, very long time…
[He takes a big breath and exhales onto his keyboard. Nothing happens at all, yet the blog allows him to pretend he is clever, not wanting to make him cry.]
Cool. Okay. I think I’m ready now.
Ahem.
I am a man. Yes, I am a middle-aged man. A middle-aged man who is slowly coming to accept how his body may or may not be in better shape than it was when he was younger. A middle-aged man who is looking to get back into fighting shape and who has made a hobby out of avoiding self-loathing by denying how old he really is. A man who is enduring this long-simmering, unceasing battle between how old he feels versus how old he is.
“You sure you can’t just stay young?" I ask of my little body, more frequently than I’d care to admit. My body says nothing in return because it’s a body, not something you can speak with, and because talking to your body like this is weird.
This is where bicycles enter the picture.
Not to brag, but ever since COVID I have become highly adept at sitting. I can sit for work, sit for gaming, sit for weightlifting, sit for eating, and even sit for watching TV; that’s just a fraction of my sitting power. I am, like, so good at sitting (insert hair flip). Basically, I sit too much since starting to work from home in 2020 and I’m deciding to do something about it. This summer I worked up the gumption to invade my local bike shop, buy a bike, and crank out some cardio while the weather’s warm.
My first visit to the bike store was an interesting one. At this bike shop there was a friendly chap named John who asked what I was looking for and helped show me around the selection of bikes. John was not only willing to talk shop and share his knowledge of the biking world but also great at helping me pick out a bike that met my needs: I’m looking to stay in shape, will mostly hit paved or prepared bike paths, and, while I think mountain biking is neat, I don’t really need a bike that capable right now. I told him I was looking for a great starter bike that’s built for—ahem—humans of increased heft and a bike that would last a long time.
John, the friendly bike shop guy, talked me through the ins and outs of the bike, what features I’d likely appreciate based upon my goals, what I wouldn’t need, and what elements are important to consider. One of those important elements was the bicycle seat. Our chat went something like this:
John: “…and when it comes to the seat, you’re going to need to be careful about getting a seat that’s too wide and comfy.”
Me (who knows nothing): “Wait, for real? Why wouldn’t I want a more comfortable seat?”
John (who knows things): “Yeah, so when you have a modern bike seat like [the one I’m pointing to], it’s narrow because it allows your butt to more naturally situate over the seat. It’s less comfortable, yes, but it’s better for your back, hips, and buttocks overall.”
Me (still knows nothing): “No kidding. I always thought a wider, more cushion-y seat was just a preference thing. That’s interesting…”
John: “For sure. A lot of people think that way so no worries. The reality is that a wide, thickly cushioned seat forces the lower back into a different posture than the standard, more narrow seat does. Over the long term, this hurts posture in general and can contribute to lower back pain/issues…”
John went on to explain how experienced riders and professional riders use the narrow seat exclusively, relying instead on special, padded bicycle shorts where the padding is formed to your backside instead of being flat on the seat. “It’ll take time to condition your butt to this kind of seat if you’re not used to it, but it’s the best route to take.”
First of all, bless this man for having to put up with me in face-to-face conversation. Second, I remember this conversation because it completely reshaped how I view modern bicycle seats:
Recumbent bikes at the gym? Yup. Big, wide, and thick with dense foam.
Many of the bikes I see folks riding in my neighborhood? Big, wide, and thick with dense foam.
As I started riding my new bike, I do remember cursing John’s name a few times. I hadn’t ridden a bike in 10+ years, so that first 6-7 mile ride involved a lot of popping joints, creaky knee sounds, my quads screaming at me as they fought to survive the horror of supporting my body weight, and my butt; my poor little butt, saddled across this torture device that called itself a bicycle seat. It seriously felt as if I was straddling one of those support poles in the middle of a stairwell.
Needless to say, the lingering soreness for the next few days was, shall we say, interesting? I must say that I was not a fan at all.
However, John The Wise had told me to anticipate this. As he told me:
“It’ll take time to get used to riding on a standard bicycle seat again. Once you hit 10+ miles at a time, or if you start riding for extended periods, you’ll want to start looking into bicycle shorts with the padding. Until then, treat bike rides like you would weightlifting. You don’t want to work the same muscles every day and likewise you won’t want to ride your bike every day, either. Give your body time to recover and you’ll condition yourself before you know it.”
I have long enjoyed lifting weights, so this kind of description made sense. Whatever your goals are, when it comes to strength training or endurance, you almost always need to work your way up to that goal, whatever it may be.
So, if heavily cushioned bicycle seats aren’t good for you in the long term, why are they so popular?
Connecting cushy bicycle seats to AI
This wasn’t as odd of a connection as it might seem. “AI” is everything, everywhere, all at once. In 2026, I can’t open my phone, watch Netflix, or check my work email without a barrage of “AI”-this or “AI”-that. The saturation of AI into everything extends across our personal lives and our work lives.
Lately I have been thinking about the extreme amount of cognitive offloading that is occuring as we use AI more and more across various kinds of forms.
Generative AI continues to be ubiquitous (ChatGPT and Claude chief among them), trendy personal AI assistants are popping up more and more, and the injection of “AI” across the ecosystem of business software continues. Whether you are seeking out AI or not, AI will find you.
I’ve been thinking about these questions (and I know that I am far and away from being the first, here):
Why does everyone seem to gravitate so much to AI/LLMs?
What’s happening to our brains as we engage with these AI tools?
Are we more productive than we are without the AI? If we are more productive, do those productivity boosts come at a cost?
As we prompt instead of doing something ourselves, do our skills atrophy over time?
Why on Earth wouldn’t someone use AI?
I’ll dive into these questions as we move along.
Why does everyone seem to gravitate so much to AI/LLMs?
As human beings, we have an inherent need to preserve energy, since there was a time where we needed to preserve our energy in order to survive. That biological, innate magnetism towards lower-effort, physical tasks is why diet fads, work out supplements, and weight loss medications are as popular as ever. As some studies put it, we are just lazy creatures in general.
Tools and machines allow us to burn less calories while increasing production. Health supplements, caffeine included, give us the ability to manipulate internal systems instead of doing so through diverse diets or getting enough rest. Weight loss medications have flown off the shelves thanks to semaglutide solutions, allowing us to feel satiated for longer and after eating less food.
What’s the connection to AI, then?
I think we’re drawn to AI systems, particularly LLMs and their chatbots, in the same way. The perceived benefit may be cognitive in this case instead of physical, but the attraction remains the same. That attraction is less mental effort + AI system/tool = desired product.
When in our lives has less effort been used or encouraged in order to produce a high quality result? I’d say never!
It’s fascinating, isn’t it? We are living in an age where capable professionals are directed to use AI—or already drawn to them—instead of doing things “the hard way”. And by “the hard way” I am referring to how we used to do things way back in 2022 before ChatGPT’s release. “The hard way” is simply without using AI. If I’m a software engineer who used to spend 8 hours on developing a feature, yet Claude Code can achieve that same feature in less than 1 hour, you can bet I’m going to try and use Claude Code.
Likewise, students are using these tools in lieu of writing or doing everything themselves. Why would I spend an entire night reading something at the last minute and drafting a research paper when ChatGPT can knock that out in a couple hours, freeing me up for something more fun? It’s easy to see why. And I don’t know about you, but when I was in school and college I would have been all over something like ChatGPT and Claude. It’s such an attractive option when you have a lot of homework and projects to deliver.
What’s happening to our brains as we engage with these AI tools?
I mentioned cognitive offloading earlier. In my own words, carefully crafted through years of laymanship and whatever the opposite of “expertise” is, cognitive offloading is when you shift the burden of information processing from your brain to an external system.
Think of it like this: You’re on one side of the room and have an idea that needs to get from your side of the room to the other side where the end product will be. You can take that idea from start to finish yourself, sure, but you know that there is a high amount of effort and strain involved. You are capable of shouldering that cognitive load yourself. However, there is a conveyor belt and machine next to you. You notice how this machine extends all the way to the finish line. With just a little effort, you telling it what you need, you feed your idea into the machine and the rest happens auto-magically. In a fraction of the time—Poof! Your idea is now the desired end product.
With AI available, why on Earth would you carry that cognitive load all the way from start to finish?
As a burgeoning armchair psychologist and neuroscientist myself, I liken it to carrying something from one place to another. I can pick up a lot of things and move them to another place, but as that load gets heavier it becomes harder and harder, eventually impossible, for me to lift it. As that load gets heavier and heavier, I know that there are fewer and fewer people out there in the world who can pick it up and move it.
If I pick up a heavy dumbbell and put it back down, then I have only experienced what it’s like to lift it. I haven’t moved it anywhere. If I pick up a heavy dumbbell and put it back down, yet this time I’ve placed it on a conveyor belt that moves it somewhere for me, even though the dumbbell has moved I have still only experienced the picking up and putting down of the dumbbell. My effort level is the same in terms of lifting the dumbbell up and down, yet with the machine available it went across the room with zero effort on my part.
As the wise Ronnie Coleman, professional bodybuilder and 8x champion of the Mr. Olympia contest, once said, “Everybody wants to be a bodybuilder, but don’t nobody want to lift no heavy-ass weight.” I’m being funny here, but he’s right. Many people desire to look like a bodybuilder looks without realizing how much effort is involved in achieving that level of muscle.
We all experience the cognitive equivalent of “heavy-ass weights”. I hit this all the time at work, where I’m fortunate to work for a company that’s moving quickly and affording me the chance to tackle interesting challenges. For me, LLMs alone have been a force multiplier, though I feel this increased effectiveness comes at a cost. The more that I rely on AI to power my work, the less in the weeds I am overall. I’m not reading through logs or Frankenstein-ing together scripts to do one thing or another. I’m no longer picking up the dumbbell and carrying it across the room; I’m just lifting the dumbbell, setting it into the machine, and guiding the machine while it does all the heavy lifting.
Now, what if that machine were to let me work with dumbbells much, much heavier than I could ever lift on my own, even to get it started? AI and LLMs provide that capability for us in all manner of disciplines, like coding and analysis, on through image and music creation.
As I work out those muscles less because I simply don’t need them anymore, what happens to those muscles? If AI means that I don’t have to think as critically, what happens to my critical thinking abilities over time?
Lifting heavy-ass weights is hard. Thinking through tough issues and building solutions is also really hard. Yet, just like with lifting a weight consistently over time, thinking through problems builds the mental muscles necessary to think critically with more ease.
As we all engage with AI, we should be conscious of when the system is doing something with us versus doing something for us. For the lion’s share of AI-led activities, I fear it’s mostly the latter.
Are we more productive than we are without the AI?
The cushy bicycle seat has been waiting patiently for me to refer back to it, so I’ll do so here.
Think about taking a bike ride, once with a standard, unpadded seat, and once with a cushy, comfy, padded seat that just hugs your backend ever so beautifully. Let’s say you’re like me and haven’t ridden a bike in a million years. You would likely find the standard bike seat to be uncomfortable (I most certainly did) and the cushy, padded seat much more enjoyable.
So if my goal is to ride a bike for 10 miles, who cares which bike seat I choose?
The comfy seat will allow you to hit that 10-mile goal much more quickly, since you wouldn’t have to worry near as much about conditioning your tailbone as you would about your leg muscles and overall cardio. The standard bicycle seat will prove challenging overall, since the discomfort of the seat overpowers any other benefits that a narrow seat provides.
If you choose the standard seat, you would have to condition your body to tolerate that discomfort over time. That 10-mile goal would take you much longer since you have to incorporate rest and recovery into your timeline. If you choose the cushy seat, boom—you might be able to power through those 10 miles before you know it.
Now, what happens when you take away that cushy bike seat? What if your buns were no longer nestled into a soft, foamy saddle? That 10 miles suddenly becomes much more difficult. It’d be a shock to the body.
From an outcome-based perspective, you achieved the 10 miles all the same. One method let you achieve that outcome much faster than the other. As a professional in the modern workplace, your leadership needs you to reach an objective. Are you going to tell them to wait until you build up the skills required, then you’ll do what they need? Absolutely not. You might know what brainpower you’re sacrificing and you may be just fine with that.
This is the conundrum of the modern age. “The ends justify the means” (said someone, at some point). Focusing on outcomes alone means that you/your team are responsible for being worried about the path to those outcomes. You will deliver what’s necessary and you will know what shortcuts, tools, AI, whatever you used to get there as efficiently as possible.
The more that AI is relied upon to reach objectives and desired outcomes, the less conditioned an individual or a team is to think critically through those challenges on their own. There’s also something to be said about the perception of AI being more effective than it is, though I may have to save that for another article.
Productivity gains carried by reliance on AI do come at a cost. In the short term, the immediate cost is the maintaining of skills and/or the development of skills (especially for folks entering the workforce). In the long term, I see continued reliance on AI as driving incredible outcomes while leaving the broader workforce less capable than it was before.
As we prompt instead of doing something ourselves, do our skills atrophy over time?
Yes. If you don’t use it, you lose it.
I’ll go back to the weightlifting analogy. If you used to pick up weights in the gym, then stop picking up weights for a while, over time it will become more and more difficult to pick up those same weights. If you don’t consistently train for strength, you will experience a loss in strength. If you don’t walk, run, or ride a bike as much as you used to, you’ll feel tighter muscles and creaky joints because your body has adjusted to less activity.
Those wrinkly masses of gray matter in our heads work in the same way. If you aren’t presented with opportunities to think critically, like when something comes up that you don’t understand, then you become less and less capable of using your brain in that way. Mental tasks that used to feel simple will require more strain, more effort. Cognitive fatigue sets in much more quickly than it did before. Your ability to concentrate on an issue for a duration of time decreases.
I’m burning time because I’m forcing myself to work with the AI, even when I’m not getting the result that I want.
Perhaps you’re like me and find yourself arguing with LLMs (trust me, you’re not alone). I’m not sure how often I’m doing this, but more often than not I spend time trying to get an LLM to fix what it did, do something differently than it did, or clarify something the AI assumed—wrongly—that I thought was straightforward. I’m burning time because I’m forcing myself to work with the AI, even when I’m not getting the result that I want.
The time that I used to burn in learning something, looking up different things, or trying potential solutions is instead dedicated solely to firing off replies to an LLM system that is somehow incredible and frustrating at the same time.
So, yes. The longer we go without doing something, the more effort it takes to do that same thing in the future.
AI isn’t a crutch we are using to walk with; it’s a powered wheelchair that we are riding in, where the system is doing all of the work for us. Step away from AI and you’re forced to figure out how to, once again, walk on your own two feet.
Why on Earth wouldn’t someone use AI?
Saying that AI is like a bicycle seat is a little silly. I recognize that. And I know that my feelings are not everyone’s feelings.
One of my worries in this current moment is that we are redefining what effort feels like; when we forget what effort feels like, it can be scary when you’re forced to apply effort in ways that have gotten rusty.
If you have only ridden a bicycle for 10 miles on a cushy, comfy bike seat, riding 10 miles without that cushy seat will be incredibly painful. If you stopped lifting weights because a machine let you move things without effort, trying to lift those same weights on your own again will be painful, if possible at all. If AI is your go-to means of problem solving, your go-to means of seeking knowledge, your go-to teammate when you want to collaborate on an issue, then trying to think through problems on your own will feel so uncomfortable and working with other humans may feel frustrating.
My call is not to abandon AI. My call is not to avoid the powerful tools that have been made available to us. My call is for us to not forget what it feels like to struggle as we learn, fight through the discomfort of not knowing how to do something, all for a full appreciation of everything these AI systems are doing for us.