You put quotes around a specific string of text. You hit search. You get back a page of results that don’t contain the string. We’ve all been there, staring at the screen, wondering if we forgot how to read or if the machine forgot how to search.
Search engines used to be tools that did exactly what you told them. Now they are assistants that think they know better than you.
You’ve probably noticed it. You search for an exact error message, a specific line of code, or a niche product review. Instead of giving you the exact match, Google—or even Discord’s internal search—tries to “understand your intent.” It strips your quotes, ignores your AND operators, and serves up a generic, SEO-optimized pile of garbage. It feels like you’re arguing with a toddler who keeps handing you a toy truck when you asked for a screwdriver.
Most people blame this decline on advertising bloat and SEO spam. They aren’t entirely wrong, but they’re missing the deeper, more insidious shift. The real problem is that search engines have fundamentally changed their underlying mental model. They’ve abandoned the reliable “database query” paradigm for a “conversational assistant” model. And we, the power users, never consented to this shift.
When you optimize for the lowest common denominator of human intent, you betray the power users who built the internet’s infrastructure.
It’s a paradox. By making search “smarter” through semantic interpretation and AI, they’ve made it profoundly dumber for precise, literal queries. A developer searching for a specific stack trace doesn’t want the engine to guess they meant “how to fix a computer.” They want a literal string match. As one frustrated user put it: if I quote a search term, every returned result must contain that string. Even if that means showing zero results. Anything else is madness.
This isn’t just a Google problem. Discord search is suffering the same fate. People are literally begging for a basic SQL command: SELECT FROM 'posts' WHERE 'text' LIKE '%search%'. We don’t need natural language processing to find a specific string in a database. We just need the database to do its job. The only lifeboat in sight is niche players like Kagi, which proves that a clean, literal-first search experience is still possible—if someone actually cares to build it.
Making a tool “smarter” by removing its ability to follow precise instructions isn’t innovation. It’s a hostile takeover of user agency.
The visceral frustration we feel isn’t just nostalgia for the early 2000s internet. It’s the feeling of losing control. We are researchers, journalists, and developers. Our work depends on precision. When search engines decide that semantic guessing is more profitable than literal retrieval, they break the foundational trust that made the internet useful.
The next time a search engine ignores your quotes and serves you an AI-generated summary of what it thinks you meant, remember: it’s not a bug. It’s a deliberate choice to abandon you in favor of the masses. Stop guessing what I mean. Just give me the search.
FAQ
Q: Isn't semantic search better for the average person who doesn't know how to use operators?
A: Yes, but optimizing for the average person by breaking exact-match functionality for power users is a false dichotomy. You can have a conversational assistant without destroying database query functionality. They just don't want to pay for the compute to maintain both.
Q: How do I get exact search results back?
A: Use alternative search engines like Kagi that respect user intent, or append `&tbs=li%3A1` to Google URLs to force verbatim mode—though even that is increasingly unreliable against their semantic algorithms.
Q: Should search engines just go back to being dumb databases?
A: No, they should offer both. The future isn't choosing between a smart assistant and a dumb database—it's having a tool that knows the difference between 'tell me about X' and 'find exactly this string'.