This behaviour is normal because a title element is bound to the video frame rate. All visual elements must start/end in "between" frames. In your example GIF you zoomed enough in and even can see the markers. Note the line just below the timecode.
The reason sound is not bound to this is because sound has a much greater Samplerate. If you think of Samplerate and Framerate as Data points over time then you understand video has commonly 24 or 25 (higher fps becomes more common) data points in a second, sound usually has 48000 data points per second.
Have your title element start at a frame what feels the best.