Oliver, you are comparing an individual with a group. Usually this is misleading.
Expanding the graph.
This is the EPICA-Dome-C-CO2-Composite time series from my database.
For the current level, here is monthly data from Alert Point.
Notice the nearly 20ppm swing down and up each year that is lost when showing yearly or coarser data. What sort of statement can be made if sampling is only every 6,000 years? Following is the sampling intervals in the EPICA-Dome-C-CO2-Composite data.
For much of the last 100,000 years the sampling interval was over 2,000 years, sometimes reaching nearly 6,000 years. Some of the earlier data was better but much was still coarser than 1,000 years. The single 1,000 yr value gives no indication how high or low the values went during the 1,000 years.
What type of scientist would compare monthly data with a single value every 1,000 years?
CO2 concentrations vary hugely over the regions of the earth from hour to hour. See this earlier post.
The Law Dome ice core. shows ice core CO2 going up while ice core temperature is going down.