> Decrease in emissions is not really possible as of now is it? CO2 is not going anywhere. We would have to start "harvesting" it.

I think you're confusing a decrease in emissions with a decrease in levels.

CO2 emissions are the amount the world pumps out every year. If we replaced all coal-fired power with non-emitting power, our emissions would fall dramatically.

CO2 levels are the fraction of the atmosphere that is CO2 molecules. If we replaced all coal-fired power with non-emitting power, but kept using gas-fired power and ICE vehicles, CO2 levels would still increase.

CO2 emissions are highly likely to start decreasing this decade. That means CO2 levels will still be increasing, but -- for the first time -- more slowly. That change -- yearly emissions flipping from increasing to decreasing -- is a necessary step towards net zero emissions.

> % decrease in emissions is incredibly flawed metric. If you have 1000 imaginary units of CO2 and increase it by 100 then it was increase of 10%. Next year you have 1100. You now increase it by 105. Increase of only 9.5%. So growth slowed down right?

I think you're confusing a decrease in emissions with a decrease in growth of emissions.

Fundamentally:

• CO2 levels drive warming.
• Emissions are the first derivative of CO2 levels.
• Growth of emissions are the second derivative of CO2 levels.

If you imagine a curve, the second derivative declining and falling below zero is a necessary precondition for the first derivative to decline and fall below zero, which is itself a necessary precondition for the curve to start dropping to lower values.

Growth of emissions is the second derivative; it has been declining for years, and will drop below zero around 2025.
Emissions is the first derivative; it still start declining around 2025, and will drop below zero if/when we reach net zero.
CO2 level is the curve itself; it will not start declining until after net zero.

Clear?