It has been that or a tie for months now. The error bars have overlapped this entire time as well.
I don't have a lot of questions about the cumulative polling average. If the election were held today I'd generally think Trump is more likely to win.
I do take issue though with breathless reporting on subgroup polling as though Trump is doing well with young people and Black people. People shouldn't be writing stories on a poll's subgroup that has a sample size of 50 and is unweighted.
When we actually do have robust polling on say young people (the Harvard Youth poll), Biden is doing quite well with them unsurprisingly.
And I do take issue with people acting like polling error and bias fit a pattern across cycles. They don't. Not even from the same pollster!
It is entirely possible the 2024 election polls are underestimating Biden's chances. And it is entirely possible they are underestimating Trump's chances. We won't know till the election. The historical average polling bias is close to zero- mainly because polling bias changes across election cycles.
I will say, I think a lot of Dem voters are "expressive" polling right now in regards to the incumbent president. And they will come home to the top of the ticket on election day. I assume that based on how the generic ballot is and on the improvement in enthusiasm among Democrats. But I could be wrong.