It has been a remarkably stable polling average so far.

I like how FiveThirtyEight has put the uncertainty intervals on this. As you can see the movement has been within the intervals and they overlap, so really statistically it is a tie.

I wish the Times polling didn't get all the breathless media coverage. It is entirely possible that the Times polling is off this year as pollster error is not correlated between election cycles (and entirely possible it is correct). We should be relying on the aggregate of a bunch of polls, rather than a single pollster.

In any case, historically polling today can be substantially different from what the final result will be in November. So we really should discount how much we are focusing on polls right now. They become more relevant in the final months. This is actually what the election models do.

Election models will generally not rely much at all on polls at this stage in the cycle. They'll mostly be relying on fundamentals. Then as the election nears polls will be relied on to a significant amount. And this is because the polls can shift a lot from now to November. To put this in context, the current polling average is entirely consistent with Biden winning the national popular vote by ~5-6 points.