According to that chart the US has had almost 10% inflation for a decade -- it's ridiculous, or at least if it were true, it would call into question the utility of such a measure. If such a measure were accurate, that would mean that real wages have not just been stagnant, but had been declining in excess of 5% a year, and that the median wage-earner's real wealth has been declining precipitously since the 1980s. I guess you can make the numbers look that way, but it's subjectively nuts.
They substitute hamburger for steak because people don't buy enough steak for it to be an adequate model of the consumer basket -- it's a luxury good and it doesn't behave like a proper staple commodity, it's bought when it can be afforded, it's subject to regional and seasonal price changes, and purchasing decisions related to it were not correlated to inflation. From the Boskin commission report:
5. The BLS should study the behavior of the individual components of the index to ascertain which components provide most information on the future longer-term movements in the index and which items have fluctuations which are largely unrelated to the total and emphasize the former in its data collection activities.
This could result in the down-weighting or even elimination of data collection for certain cities and a revision of the commodity structure of the index which would consider some goods as having a national market, sampling a larger number of items but with less regard to geography, focusing on geographical differences only for more "local" commodities, such as fuel costs, rent, personal services, and fresh produce.Currently, the BLS collects a large number of price quotes on bananas, because they are inexpensive to collect and their prices are quite variable, even though these variations are not related systematically to the underlying trend-movements in the CPI. At the same time, less attention is paid to less variable but more likely to change (disappear or be redesigned) and harder to measure commodities, such as surgical treatments, consumer electronics, and communication services.
They changed the CPI calculations in the mid 90's because the CPI's method up to that time was completely unscientific and based on arbitrary, non-evidence based preconceptions about what people shopped for. Meanwhile they eliminated food and energy from the core CPI because they concluded these prices were far too volatile to aid in policy making, and in the end food and energy are just inputs into other processes which are eventually priced by the CPI. They also made the very wise decision of eliminating house prices from the calculation, substituting equivalent rent. The consensus among economists is that the CPI was overestimating inflation throughout the 80s.
I'm not going to say there can't be more room for improvement, but I think you're wrong on this point, and your account of how the CPI works is glib and biased.