Tuesday 15 of July, 2025

Having had a nice day out yesterday, despite over 7 hours in the car (more in a rant), I sat down this morning to check a few facts. I was looking to check up a little more of my own time line, but first hit was my involvement with Firebird. I had been using it since the fort from Interbase, so when Mozilla tried to mussel in on the name I was quite vocal that it WOULD cause a conflict. I tried using Mistral to establish a few facts, but it quickly became obvious that facts are something that it's training model seems a little short of. All of the stuff I am looking at happened in the early 2000's so should be well covered, but apparently not. A bit of firebird history is what I ended up with, but I had to correct numerous mistakes in getting most of the facts right. I still don't have references for some of the claims which is proving an area where Mistral avoids helping with. It originally claimed that the hijack of firebird was sorted in a few days in 2004 and gave a date of 14th Feb as the date, following it's original date of 9th Feb for the announcement of the use of Firebird by Mozilla. The write-up now shows the right dates but I still have had no answer for Mistral as to where it created the 14th Feb date from! I will rework the write-up in my own language, but I had hoped it would have at least given me an accurate date framework. It's not the first instance of this either, if asking for dates one HAS to validate them, so is Mistral actually being productive at all? Not convinced at all as other reports are also confirming. Call center ai assistants on The Register for example ( There was abetter one I can't immediately see )