A living cell and a working computer both have information [Shannon: reduction of uncertainty in the receiver] They are successfully communicating. A dead cell and a broken computer do not have information, communications has ceased.
An important feature of information theory and molecular biology is that for every bit of information gained in the reduction of uncertainty (the measure of the information gain is the Shannon entropy after state less the before state) releases energy to the local surroundings (thermodynamic entropy, heat).
"I know life when I see it" doesn't satisfy.
What makes it idle speculation is that we don't have any physical evidence about the process, and are unlikely ever to have any. If we had complete information about every step of the process, there'd be no room for speculation of any kind, but we still wouldn't have agreement about the transition point. We'd each choose some distinguishing characteristic or some key event as the changeover point, and all of those choices would be arbitrary.
It all becomes a fuzzy, moving target.
No, it always was a fuzzy, moving target. That's how life is. If we find it easy now to tell life from nonlife, it's because the organisms we see are all the end products of billions of years of optimization. Early on, it probably wouldn't have been so easy.
Here's a different example. It's easy to tell a snake from a lizard. Go back far enough, and it would have been hard. Go back even farther, and there wouldn't have been a difference at all.
Does that mean that the modern difference between snakes and lizards is a matter of speculation or prejudice, or that we can't in principle lay out how they diverged? No. What it means is that even if you had complete information about every reptile that ever existed, you still wouldn't be able to make an objective, complete list of just the snakes. There's no spark of snakiness.