Experimenting With Tech: Build Something Just Because You’re Curious
I remember when I was in Class 12, I wanted to build two things that probably made very little sense as “projects.” One was a faster AI suggestion tool for the desktop — something that could understand what I was typing and suggest what I might want to write next. The other was a PDF n-gram counter — a tool that could take a PDF, analyze the text, and tell me which words or sequences of words…
In their Class 12 year, the writer wanted to create two seemingly pointless projects: a faster AI suggestion tool and a PDF n-gram counter. These weren't born out of a need to solve a business problem or a detailed roadmap, but simply out of curiosity. The writer believes this mindset of asking "Can I actually make this?" is what kicks off true experimentation.
Instead of starting with a grand idea or solving a business problem, they suggest beginning with a simpler thought. For example, "What if I had a bot that watched certain websites and gave me a summary of changes?" The outcome could be useful or pointless, but either way, it teaches something valuable.
Curiosity should be the starting point over having strict requirements. The writer proposes creating personal AI information systems that don't aim to turn into products, but are instead technical playgrounds for learning and exploration. For instance, a system could keep track of changes in areas you care about, like AI infrastructure, rather than generating generic news summaries.
They also encourage building somewhat absurd experiments. This could be a bot that argues with your ideas, an AI that groups your bookmarks into topics, or a system that automatically researches a randomly chosen question each day. These aren't necessarily businesses, but rather questions turned into software.
Traditional tutorials provide a destination, while experimentation provides the problem. By turning the question into an experiment, you have a reason to learn the necessary technology, making it a tool for answering your question rather than just following a tutorial.
The writer advises keeping experiments small and specific. This could mean giving yourself constraints like one evening, one weekend, one API, one dataset, or one weird question. Build the smallest version that can answer "Does this idea work?" If it does, extend it; if not, throw it away and start another experiment. This ability to discard failed experiments is crucial.
Lastly, the writer suggests maintaining a personal lab of sorts. After a year, you might have built 50 strange little systems, most of which you'll abandon. However, a few could potentially become something much bigger. Regardless, the process of experimentation is invaluable.
Written by urgent.news from Dev.to's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
