I think the ultimate answer is that you can’t overcome physics. A heavier kart/ driver will be slower.
The question that is more interesting is “but by how much?” This is where it seems we are all unclear, as a lot of it depends on factors beyond this, namely things that try as we might, we can only make somewhat consistent - the track conditions, parity among engines, and of course ‘skill’ of a driver.
To really understand anything meaningful about weight vs. pace, you’d ideally want the largest data set possible. A couple people have reference relative pace between different categories, but is this enough to make a full determination? Perhaps at one track the larger set of data gives us one conclusion, at another it yields a different result. Another factor to consider is the disparity between two different engine package situations - would time difference be the same gap between a light group and heavy group of 206’ers, and similarly weight difference Rotax racers or Shifter kart racers? Just looking at your local series’ lap times could give you a hint, but it’s still a smaller data set as there are so many variables that are hard to eliminate.
It’s a frustrating question to really nail down, but one id be very curious if with a nationwide data set a more meaningful conclusion could be met.