Let me assure you, I am no luddite. I use electricity, computers, the internet, and I develop websites as a professional web developer. I use AI when it makes sense, I use Google and StackOverflow, GitHub and digital notes online.
However, many fellow developers, customers, and users seem to fall for a misunderstood promise of the current tech and AI hype, failing to understand marketing patterns, hype cycles and psychological bias.
Constructive Criticism like 8 Alternatives to AI for Coding and Creativity, Google Alternatives and StackOverflow alternatives for web developers already show that it's no solution to fall back for yesterday's tool to avoid today's problems.
Deceptive patterns are a thing: misguided KPIs and online marketing insights reinforce bad practice, like making users spend much time and interact on a website as a positive signal only serves ad pusblishers and content creators promoting quantity as consistency. A clear and focus experience allowing people to reach their goals quickly falls below the radar and gets shadow-banned for not creating enough engangement, and this cycle keeps boosting bad UI and pushing poor search results to page one top places - so users would spend more time browsing search results while seeing more paid ads. Part of generative AI's success is that we'd rather read a flattering slop post than bother with any more SEO blogs bloated with ad banners or risk that StackOverflow gatekeepers mark our carefully crafted question for deletion.
A Healthy Digital Information Retrieval Pyramid
We could compare third-party online services with buying groceries. Maybe we have some home-grown specialitities or in a small neighborhood garden. We could go to a local food market, fill our cart in a supermarket or order something delivered to our door. Food quality and sustainbility differ a lot. A single avocado shipped, stored, packaged and delivered has a higher carbon footprint than the one that grows in a local greenhouse. An organic fruit bowl with overnight oats has a better nutritional value than a toast sandwich wrapped in plastic since leaving a factory oven a week before.
Offline-first may be a new trend, rediscovering autonomy with local language models, localhost development servers, and regional data centers independent of centralized cloud services like CloudFlare whose recent outage took down a great part of today's mainstream internet. Offline-first is like a sunny kitchen window or a small greenhouse in grandma's garden: nice, useful, but it doesn't scale well. Still, that aligns with the IndieWeb POSSE principle: start at your own place, keep and grow a local core of your business and content resources that you control completely.
Offline First: Local Resources
Local resources also include my own analytic thinking and creativity! Before turning to AI or web search, I can ask myself if I look at a problem from a helpful angle, if I already checked typical points of failure, read error message details and log files, and use local code analysis like linters, auto completion suggestions or even a local LM. Local resources also include pen an paper: standing up and drawing a sketch of an information flow, design detail, or software architecture can help a lot! Also, sketching and painting can inspire and empower humans much more than crafting the right prompt to make an AI generate a picture.
Any third-party service online is a commodity that may or may not be available at a certain moment in time. I wouldn't go so far and keep a local copy of every website and it's recent IP address. But offline documenatation and notes on paper won't hurt, and I even buy or borrow a printed book occasionally. I have a link list of websites, formerly known as digital bookmarks, so I don't need to google or waste Perplexity tokens to land on an autoritative page like Mozilla's MDN web development reference.
Bookmarks, Ecosia, Google, AI
I use Ecosia as my default search engine, to save energy and improve privacy. Only when the search results don't seem helpful or I suspect that they don't based on prior experience (like specific coding error messages), I use Google instead. Within the search results, I favor well-known sites like MDN, DEV or StackOverflow (which still has not the best internal search function). When the search results don't help, or when the problem seems too complex to phrase it as a short sentence that a classic search engine might understand, I turn to AI. Search engines sometimes run an AI-mode query already anyway, otherwise I'd prefer Perplexity when I want to check information sources, or Claude when it comes to coding.
Creative Surprises and Intutition
Creativity often comes with suprise and intuition! When sketching my pen-and-paper version of an information retrieval pyramid, the diagonal walls didn't quite conntect, leaving the top open. That made me think of an erupting volcano as an apt analogy for AI's verbosity and energy consumption.
Conclusion
Don't be a fan of technology, and don't reject it either, but use it wisely and sparingly! I work in tech, I use it every day, and I have used AI for fun and professional reasons. However, in my upcoming posts and projects, I'll stick to more traditional techniques and culture when I can and try to prove that makes me more productive and stand out as a creative web developer!
Top comments (0)