@itsallykrinsky pls don’t do this #ai #vibecoding #techcareer ♬ original sound – ally krinsky
Shout out to Ally Krinsky for a great story that made me laugh and inspired me to write this article.
For those that can’t see or watch the TikTok… first off, congrats for not being on TikTok, but here is the recap:
Ally worked at a company with unlimited access to AI, and employees were encouraged to build whatever they wanted. This one guy gets an idea and is motivated to turn his PowerPoint slides that he made into a website. So he works with AI to convert these slides into content, then HTML, then creates a website, and finally asks AI to deploy the site, all so he can share it with his co-workers.
When he could have just attached the slides as a part of an email.
This is an over-engineered state where people are wasting time with AI, and costing money, and it doesn’t add any value.
The Space Pen Problem All Over Again: Why Using AI Doesn’t Always Mean You’re Moving Forward
There’s a story that gets told in engineering and business circles all the time. To be honest, for years I thought it was made up, but during the space race competition between the Cold War rivals, the United States and the Soviet Union, NASA realized that ballpoint pens wouldn’t work in zero gravity, as ink needs gravity to flow. So they spent years and a significant budget developing a pressurized pen that could write upside down, underwater, in extreme temperatures, and in the vacuum of space.
The Soviets used a pencil, [insert facepalm emoji]
The punchline lands every time because it captures something deeply human: the tendency to reach for the most complex, sophisticated solution available, even when a simpler one is sitting right in front of you. We confuse effort with progress. We mistake complexity for intelligence. And we burn resources, time, money, energy, building something impressive when what was needed was something useful.
Now fast-forward to 2026, and that same instinct is showing up everywhere. Only this time, the pen has a chatbot attached.
Complexity Is Not Value
The space pen story has a twist, by the way. NASA’s pressurized pen wasn’t actually wasteful; graphite from pencils in zero gravity could float into electronics or eyes, creating genuine safety hazards. There was a real reason the simple solution didn’t work. The complexity was earned.
That’s the question worth asking every time you reach for a sophisticated tool: Is this complexity earned? Is there a real problem that the simple solution can’t solve?
In the PowerPoint-to-website story, the answer is almost certainly no. PowerPoint slides are attached to emails. Emails get opened. Teams read them. Problem solved, zero infrastructure required.
The AI workflow didn’t remove friction… it added it.
Time spent prompting, iterating, debugging, and deploying. Compute costs. The cognitive overhead of managing a multi-step process. And for what? A link instead of an attachment.
Why This Keeps Happening… And My God, When Will We Learn
It would be easy to mock this kind of over-engineering, but it’s worth understanding why it happens, because it’s not stupidity, it’s a very human response to a genuinely exciting moment.
AI tools feel like superpowers right now. And like anyone who’s just discovered a superpower, there’s a temptation to use it for everything. To demonstrate mastery. To be seen as someone who “gets” the new technology. There’s social capital in being the person who built the AI-powered workflow, even if the workflow is solving a problem that didn’t need solving.
There’s also a subtler trap: AI makes complex things easy enough to attempt. Tasks that used to require weeks of development now take an afternoon. That’s genuinely incredible. But “easier to attempt” doesn’t mean “worth attempting.” It just lowers the barrier to building things that don’t need to exist.
The Well Was Right There The Whole Time
In parts of rural Africa, access to clean water is a genuine, life-or-death problem. It draws attention from governments, NGOs, tech companies, and well-meaning engineers from around the world, all eager to help, and all arriving with ideas.
Some of those ideas are extraordinary. Solar-powered pumps, filtration systems, and IoT sensors that monitor water quality in real time and send alerts to a central dashboard. Impressive technology, built by smart people with good intentions.
The problem is maintenance. When a solar pump breaks, you need replacement parts, technical knowledge, and supply chains that often don’t exist in remote communities. When the sensor network goes down, nobody local knows how to fix it. The solution that looked modern and scalable on a whiteboard becomes a liability in the field.
Meanwhile, the hand-dug well that has served the village for generations continues to work. Community members know how to maintain it. It requires no electricity, no internet connection, and no spare parts from a warehouse three countries away. It just works.
The intention behind the complex solutions was never wrong. But complexity has a cost, and that cost is often paid by the people the solution was meant to serve.
Sometimes the oldest answer is still the best one.
The Real Measure
Here’s a simple test worth applying before any AI-assisted project: what problem am I actually solving? And does this solution solve it better than the simplest alternative?
Not more impressively. Not more technically. Better. For the person at the end of it.
AI is a remarkable resource — in terms of time, money (API costs add up quickly), and the opportunity cost of skilled people’s attention. Burning those resources to produce outcomes that a two-minute task could have achieved isn’t innovation. It’s the space pen. Expensive, impressive, and entirely unnecessary when a pencil was right there.
The teams getting real value from AI right now are the ones asking “what can’t we do without this?” not “what can we do with this?” The first question leads to genuine leverage. The second leads to websites that should have been emails.
Don’t mistake the tool for the outcome. The goal was never to use AI. The goal was to share some slides.
A pencil would have been fine.
