Hello colleagues,
I would like to ask if any of you are familiar with an FM solution that uses fuzzy logic. I would be interested to know what tasks fuzzy logic performs in this context.
Do you have any suggestions as to where fuzzy logic could be useful in an FM solution?
Wolfgang
For me, fuzzy logic would be useful in searching.
However, after 30+ years, FMP’s “search”, IMHO, is so rudimentary (painfully basic), it should be embarrassing to Claris.
Fuzzy logic would be a dream. Maybe after they let you search all your scripts (a normal version 1.0 kind of thing)?
This may turn out to be an area where leveraging "AI" may be useful in FM, but one will need to create the appropriate tools (scripts/query formations) and prompts. I know, I know... yawn... AI, schmay eye..
Proof+Geist has a new tool called ProofChat (for now) that offers a free level, plus Pro options. Todd Geist did a series of videos with Richard Carlson on the topic of 'ProofChat'. I could see leveraging some of the AI strengths to form 'fuzzier' search queries – at least after some training/tool designing.
Here's a link: ProofChat Is Here » Proof+Geist
No affiliation, but watching the VDOs opened my mind to integrating LLMs with FM, while previously I was a bit bored with it (so far and in my use cases).
I’ve been using totally FREE AI with FMP (Ollama) long before Claris decided to bolt yet another thing on to FMP they could try to make another buck with…
Yes, there are several free models available.
I’ve seen a couple edge use cases for AI in my area of data management, but my clients mostly have no interest in the kinds of stuff being discussed “as the future”. I follow along to see various implementations (of AI), but once we detect a repeated query or parameters for a report (for example) we typically build it into our system (if not already part of the design) and AI would just be clutter for our users. IMO, it’s not the end all, be all for everyone.
Our systems are “closed”. If we needed to scrape LinkedIn or similar resources for data AI would/could surely contribute. Also for example: legal research to prepare for a case it will help a lot, but that’s not my space.
Plus, with AI one has to be careful about potential hallucinations. And there can be privacy and security implications.
I wasn’t pushing AI at all; only responding to the “fuzzy” part of the post and suggesting/agreeing “fuzzy search” could be implemented with various AI models.
Understood. What I do with OLLAMA is to allow the client to set the AI “temperature’.