The Reddit user who asked for a shorter way to write `=filter(A:A,(A:A<>1)+(A:A<>2)+(A:A<>3))` has put a finger on exactly why traditional spreadsheets feel like a chore. The proposed syntax `=filter(A:A,A:A<>1|2|3)` is cleaner, but it still reveals the deeper problem: users are forced to wrestle with notation instead of focusing on the question they actually want answered. Our take is that this friction is not a minor inconvenience, it is a signal that the tool itself is overdue for a rethink.
The formula the user wrote works, but it requires mental gymnastics. You have to remember that `+` acts as an OR operator inside FILTER, that each condition must be wrapped in parentheses, and that the whole expression must be manually expanded for every value you want to exclude. That is not a skill problem; it is a design problem. When a simple request like "exclude these three numbers" demands a six-part expression, the spreadsheet is asking too much of the person using it. The user is not the one who should adapt.
What this means in practical terms is that millions of people spend time each day translating their intent into spreadsheet syntax rather than getting answers. An AI-native approach can change that. Instead of requiring users to learn the quirks of array operators, a smarter system could interpret `A:A <> 1, 2, 3` as a natural request and handle the logic internally. The user's job is to know what they want to exclude, not to remember how to string together conditions with plus signs and parentheses. That shift from "how do I write this?" to "what do I need?" is exactly where AI can make a tangible difference.
We believe the future of data work lies in tools that meet people where they are, not in forcing them to memorize arcane syntax. A capable user hit a wall not because the task was hard, but because the notation was unnecessarily complex. An AI-driven spreadsheet that accepts plain-language filters and automatically translates them into the correct logic would eliminate that wall entirely. It would let users spend their energy on analysis and decision-making, not on debugging formula structure. That is not a distant promise, it is the logical next step for a tool that should empower, not exhaust.