I generally like this one by Tyle Cowen describing Keynesian business cycle theory.
1. It's all "sticky-wage in the face of an AD fall", which is understandable but obviously not ideal. This is always a tough call - exactly what do you say to a person that has no background in economics? Austrians actually face this too I think. How do you put your theory in a nutshell. Cowen's solution is a good one. If I was asked I might have said something like "Interest rates are determined by the supply and demand for liquidity, not the supply and demand for loans. When some shock causes people to become pessimistic about future demand you generally see three things happen: (1.) people want more liquidity, (2.) people want fewer loans, and (3.) people are less willing to supply loans. This generates a fall in output that can feed on itself, since my demand for output is going to decline if someone's demand for my output declines. This process by which the initial shock can feed on itself is called the multiplier effect". I think that's a pretty simple way of putting it, and it's a very Old Keynesian view, as opposed to Tyler's more New Keynesian view (not that I have a problem with New Keynesianism).
2. I'm very glad he mentions the strength that he does.
3. The first weakness of course is right. I'm not even sure I'd call it a "weakness" - it's just a fact of life. I expect no Grand Unified Theory of business cycles, and I don't think anyone else should either.
4. The second weakness I have more of an issue with. Yes, the fiscal multiplier is contested territory both theoretically and empirically, and the objections are not all from cranks. But the statement that "there is a lot of evidence from history that fiscal policy very often doesn't work very well" does not seem like a fair read of the evidence. The real concern comes in with the long-run impact. Even though the short-run is contested too - I wouldn't say it's perfect - most evidence has come down on the side that it does work. If you're going to do a generalist, bird's eye view of "the evidence", I don't think it would be the one Tyler suggests. Is there are a range? Sure. What's our best guess? In the 1.5 range. That goes in the "strength" column, not the "weakness" column. Or at the very least say nobody agrees. Don't say the evidence is that it doesn't work.