3 ms·
i wonder how many of them are like 100% of my requests? - "set a timer for 14 minutes" - "remind me to make a doctor's appointment tomorrow at noon" - "turn
by GuinansEyebrows 14d ago
i wonder how many of them are like 100% of my requests?
- "set a timer for 14 minutes"
- "remind me to make a doctor's appointment tomorrow at noon"
- "turn off the bedroom lights"
that don't require any form of AI and have functioned perfectly well (aside from having a hard time when i mumble) for years on non-AI versions of Siri?
- thewebguyd 14d agoI suspect even some of the more "advanced" queries they demoed off, i.e., "what was that restaurant John and I talked about last week?" also not require LLMs whatsoever and could have been accomplished with a simple search.
- anamexis 14d agoI think you're going to need some kind of ML to turn that question into a structured search, and interpret the results.
- regenschutz 14d agoNot necessarily. NLP is possible to do without any ML. See ELIZA [0] as an example, which used pattern matching and substitution to process natural language. A few simple rules can cover a majority of requests. It's still AI, but not ML. Though I do agree that the ML-based NLP thst we have today is much more usable, albeit more compute-expensive. [0]: https://en.wikipedia.org/wiki/ELIZA https://en.wikipedia.org/wiki/ELIZA
- exographicskip 14d agoThis matches my use cases. Poor speech to text recognition so I don't trust siri to do anything non-trivial
- dreamcompiler 14d agoSiri was always AI, but the popular connotation of "AI" changes all the time. It now means "LLMs built with transformers."