Based on the number of comments, I guess I should have qualified recent memory to be the last 2 years - here's my reasoning. In January of 2007, Intel was just ramping up the core-based Xeon, which proved to be not only a solid x64 product but dominated the aging Opteron products. Newly acquired ATI was also not competitive with nVidia. Fast forward a year, and we see the Barcelona debacle, with ATI still not showing much promise. During that time, AMD did not seem to have a clear direction or strategy, was spending well ahead of prudence on assets with limited near-term potential, stumbling badly on base execution, and generally seemed to be on their heels and disorganized.
Since summer of 2008, we saw much of AMD's senior management replaced, an ATI product line which beats nVidia soundly at nearly every price point, some strong chipsets, and a very solid 45nm launch. If, as seems likely, AMD completes the fab spin-off, they should be on a more solid financial footing, with good products for many segments and winners in some segments.
I admit that I have not followed AMD closely until very recently, but my general impression prior to that was they were an also-ran until the Athlon, failed to capitalize on that (despite brilliant incorporation of advanced Alpha features that gave them a performance lead in desktop for several years), managed to take advantage of one of Intel's few mis-steps (Itanium and Netburst strategy providing a big window for x64) to launch Opteron, failed to use that brief surge to clean up their act (and obviously underestimated Intel's ability to change course), and only in the last year have shown the kind of focus that can keep them in the game long term.
Core is a great performance architecture and Nehalem is even better. But AMD has several areas where their products can have broad appeal, especially in today's market - performance is not the whole story. 2009 looks to be as interesting a year for servers as 2005 was.
Culture, can-do attitude and compelling leadership can change the landscape pretty quickly - only a few years ago, Apple was considered a footnote headed for the slag heap. IBM pre-Gerstner was given little hope. I can think of other examples where a near-death experience, the right market conditions, and the right leadership blew away previous notions like smoke on a windy day.
The real question in my mind is whether the trial by fire has really caused a deep shift in thinking and execution at AMD. Time will tell - but I don't see them as close to collapse as many here seem to believe, and I hear a lot of good things about clarity of vision and execution focus at AMD.