I mean, the second point is that it's asking the audience to derive a story from raw data. Again, this implies a lack of respect for the audience. They're really only willing to take 7 seconds to look at the slides and understand what's going on, so make it count. The second chart (the bar chart) tells the story that, in no uncertain terms, that the children were more excited overall after the pilot program. What story do the 2 pie charts tell? I dunno, you figure it out yourself, dummy. People understand stories better than just raw data. Make use of that. (At this point, you really should be reading that book to get what I mean; I found my copy quite easily online)
I went over the pie charts myself and generally agree with you here. It took quite a bit of back and forth between the charts, looking at the numbers, the legend, and the size of the chunks to work out what the story was. For example the first thing I noticed was the largest green chunk on the left, checked the legend and it was "OK", then my eyes went over to the right pie chart and noticed that the green chunk had significantly shrunk, implying less people were "OK" now. So, that initial information was sending a mixed message: why was the largest "OK" chunk now smaller? Less people were "OK" afterwards? Then it took a bunch more flicking back and forth between the pie charts to work out that less were "OK" because they'd shifted into Interested and Excited categories. All of that is clearly wasted time that could be avoided with a more appropriate choice.
It's not really the sizes however, it's the extra work connecting the dots. The choice of chart type just wasn't appropriate. For example, the categories are ordered from best to worst, whereas a pie chart removes the ordering, thus it's lost information by default: best and worst are now right next to each other. So yeah, I'm not sure the main problem here is the size judgements, it's the unordering of the information and the extra legwork you need to do to connect related information since laying it out like that scatters relevant info to all corners of the infographic.
Really the only time you want to use a pie graph is when you're trying to point out how one thing, the biggest, is disproportionately large: making a visceral point, such as doing a pie chart showing federal spending and the chunk that's for the military or something. The point isn't to accurately convey the sizes of all chunks, the point is to highlight something being out of whack, and in that case a pie chart makes the point more cleanly than a bar graph, since some of the bars will be very small, so most of your screen would be wasted whitespace. For example if you wanted to highlight how massive the sun is compared to the planets you could do a bar graph, but the sun would be one huge line and there would be empty space on most of the screen, or you could do a pie chart representing each body's proportion of the mass of the solar system. The pie chart would be the punchier choice here.
Using dual pie graphs for a before and after like that is clearly an engineered example that's using the tools incorrectly. That's not what pie charts are good for. The information that matters in that example is how the values changed over time, and the pie charts don't convey that well.
