We have 866 presidential polls from 96 pollsters in our database as of this morning. The NYT/Siena College poll is one of the better ones. It is being guided by two very experienced people at the Times, Nate Cohn and Ruth Igielnik. Here is their final version for the swing states:
Kamala Harris has small, but essentially meaningless, leads in four states, is tied in two and trails in one. The thing that is important here are the shaded error bars. Harris' vote share could be anywhere in the gray or blue shaded part. Donald Trump's vote share could be anywhere in the red or gray shaded part. There is considerable overlap. In short, it is probably slightly better to be Harris than Trump right now, but other good polls have different results (see the new polls at the bottom of the page). Anything is still possible, including a blowout for either candidate or races that need three recounts before getting an answer.
The article linked to above compares their own polling with 10 other pollsters, two of which we don't trust (AtlasIntel and Susquehanna). They are all over the map. For example, in Wisconsin, it varies from Trump +1 to Harris +6. That's not possible. It can't be both. All we can say is that Harris appears to be doing somewhat better in the Blue Wall than in the Sun Belt, but that might not be true either. If you came here today to learn who is going to win, we are going to have to disappoint you. It's close to a coin flip unless there is something going on that is not visible now. But keep reading. (V)