100 grams a second is about 0.4 tons per hour. That's 9.6 tons a day, or 3,504 tons per year. This is roughly as much as the City of Milan's domestic CO2 emissions from 2005. The Scherer plant, the most pollutant individual power plant in the USA, produced 25.3 million tons of CO2 alone every year. We're really good at pumping shit into atmospheres, in fact. Just so we can compare the numbers in numbers, Martian air loss looks like this:
And Scherer's CO2 output per year looks like this:
Mars presently has an atmospheric mass of 25 tera tonnes. Earth has one of 5148 tera tonnes. A Teratonne looks like this:
Mars is about half the size of Earth, so we'll say we need to go from 25 teratonnes to 2500 teratonnes to get a similar atmopsheric density. In tons, those look like this:
Current atmosphere: 25,000,000,000,000
target atmosphere: 2,500,000,000,000,000
Loss rate: 3504
Now look at martian air loss again in comparison to those, but in teratonnes:
Current atmosphere: 25
Target atmosphere: 2,500
Loss rate: 0.000000003504
We need to increase the Martian atmosphere by around 10,000%. To do this in a human time scale, were looking at doing that in around 100-200 years. We need to be increasing the atmosphere by something in the region of 12,500,000,000,000 tons a year. Countering that 100 grams of atmospheric loss a second increases that to 12,500,000,003,504 per year. It's a rounding error. A barely-relevant 9th decimal place loss rate. For it to become relevant to our calculations, it would need to increase by billions of percent as the atmosphere thickens.
TL;DR: Merely talking about terraforming Mars requires us to assume a rate of transformation that makes atmospheric erosion irrelevant.