Pros
Amazing scientific resources as far as equipment, space, colleagues, access to reagents/materials, etc. NIH attracts talks/visits from the best scientists world-wide. The library and journal subscription list are phenomenal. The folks at the Office of Intramural Training & Education are great and are continuously developing/adding to a first class program - especially the career counseling program. There are other institutions that likely match these resources depending heavily on the lab you are in (e.g. labs at Harvard, Stanford, Janelia, etc), but I can't imagine many places exceeding the NIH in this regard. The size and diversity of world-class research at NIH is breathtaking, and you'll never have the time to fully appreciate its scope. This can be overwhelming, but the ability to find what you need (or create the resource yourself) is there if you take the initiative.
Cons
For a post-doc, there are some important limitations on your path to an independent career that are unique to the NIH (except #5): 1. There's generally a 5 year limit that no matter how much your PI likes you, can be hard to overcome. Some projects just take a while, and the 5 year limit can be a big problem if you're slow to get publications out. The likelihood of moving up in the intramural system is very small. There's not the equivalent of a research assoc professor - the closest thing is a staff scientist position, and the probability of finding an staff scientist position that you are qualified for (i.e. possess necessary skills/expertise) is vanishingly small. Thus, after your 5 years are up, you'd better have a plan to move to something outside the NIH. 2. You are not eligible for some grant mechanisms (e.g. K01) and the K22 has become rare as some institutes have done away with this mechanism. Because there's no equivalent of a Res Asst Prof, you're not eligible to apply for R22 or R01 grants. This means the K99/R00 mechanism is really your only practical bridge to obtaining funding that a large an increasing number of departments now require (de facto if not official policy) to obtain faculty position offers. K99/R00s themselves are becoming exceptionally competitive and are no guarantee of a job offer. Furthermore, the 4 year limit on applying for a K99 puts you on a tight time table (see #3 and #4 below), and I think many intramural PIs are slow/unable to appreciate this fact. Consequently, having fewer options than extramural post-docs is deeply problematic. 3. Because intramural funding is more stable than most extramural "grants-based" funding (I'm leaving the HHMI and other major foundation funding sources out as they comprise a relatively small percentage of labs), NIH PIs, in my experience, are more willing to take risks with their laboratories projects (i.e. assign/allow post-docs to risky projects). As a post-doc coming into this situation, you need to be vigilant and understand that it is you who is assuming the risk (and a lot of risk vis-a-vis your future career!). You would be wise to develop (and make sure you have time for, see #4 below) a second, less impactful, but safer project. 4. While many university labs are more heirarchically structured (i.e. you've got undergrad workers, technicians, grad students, post-docs, and maybe a Res Asst Prof or two), NIH labs tend to be flat and heavily post-doc driven. There is variability from lab-to-lab regarding how projects are assigned, but frequently, everyone works on their stuff. This means you will likely not have team members underneath you, and participation in collaborative projects with other post-docs is potentially limited (both you and the other post-docs have to concentrate on your own projects). Consequently, NIH post-docs frequently work on a project by themselves and that limits throughput. If you're competing against extramural labs that can't afford to structure their labs this way, then you're at a disadvantage. On the other hand, you probably have more scientific resources (and money) at your disposal, so you've got to do the math on what you need to hit a home run. 5. What post-docs get paid is an embarrassment for the amount of time they work, the creativity they bring to the project, and the number of years of experience in the laboratory most already have coming into the job. The NIH is probably on the higher end of the spectrum in post-doc salaries/benefits, and this is an industry-wide practice. Still, I personally believe that the NIH sets the standard (or at least has large influence on this) based on its practices.
