Sage advice from the sage himself.  In my opinion he saved his best advice for last:

1. The investment management business creates no value, but it costs, in round numbers, 1% a year to play the game.  In total, we are the market, and given the costs, we collectively must underperform.  It is like a poker game in which the good player must inflict his costs and his profits onto a loser.  To win by 2%, you must find a volunteer to lose by 4%. Every year.
2. In a zero sum world, hedge funds in total merely increase investment fees.
3. Most stock markets are approximately efficient at the stock selection level and probably getting more so.
4. Transaction costs and management costs are certain, but anticipated outperformance is problematical.
5. Given the above, within single asset classes indexing is hard to beat, and relative passivity is not a vice.
6. Therefore, indexing must surely squeeze out active managers until it represents a substantial majority of the business.  Remember, it is the worst players who drop out of the poker game to index.  The standard of the remaining players, therefore, rises … and rises … but, fortunately, for us, beginners continue to join the game.
7. Indexing is held at bay only by the self-interest of the players or agents, as opposed to the real investors.  The outside managers want fees, and the hired guns want a job that looks demanding.
8. More recently much of what passes for outperformance or alpha in hedge funds (and private equity for that matter) is merely leveraged market exposure.
9. Asset allocation is intellectually easy to get right because mean reversion is a reality, and new paradigms almost always an illusion.  Asset class mispricing is sometimes so large it simply cannot be missed.  (35 P/E in March 2000).
10. However, in asset allocation timing uncertainties can be longer than clients’ patience, introducing large career and business risk.
11. Historically, equity investors have overpaid for excitement or sex appeal:  growth, profitability, management skills, technological change, and, most of all, acceleration in the above.
12. Bodies in motion tend to stay in motion (Newton’s First Law).  Earnings, and stock prices with great yearly momentum, tend to keep moving in the same direction for a while.
13. Everything concerning markets and economies regresses from extremes towards normal faster than people think.  Factors that regress include sales growth, profitability, management skill, investment styles, and good fortune.
14. One of the keys to investment management is reducing risk by balancing Newton (Momentum and Growth) and regression (Value).
15. Growth companies seem impressive as well as exciting.  They seem so reasonable to own that they carry little career risk.  Accordingly, they have underperformed for the last 50 years by about 1½% a year.
16. Value stocks, in contrast, belong to either boring, struggling, or sub-average firms.  Their continued poor performance seems, with hindsight, to have been predictable, and, therefore, when it happens, it carries serious career risk.  To compensate for this career risk and lower fundamental quality, value stocks have outperformed by 1½% a year.
17. Real risk is  not accurately measured by beta or volatility, which is compromised by a positive correlation with other characteristics, such as growth, excitement, liquidity, and analyst coverage, which are valued as ‘goods’ and reduce career risk.  The good news is that they don’t take Nobel Prizes back.
18. Real risk is mainly career and business risk, which together shape our industry.  Efforts to reduce career risk – “never, ever be wrong on your own” – create herding, momentum, and extrapolation, which together are the main causes of mispricing.
19. There is no small cap effect, price/book effect, or stock vs. bond effect, only a cheap effect.  The current price tag is always more important than historical averages.  (Stocks don’t beat bonds because it is divinely ordained – Jeremy Seigel’s Stocks for the Long Run – but because they are usually priced to outperform.  Today, for example, they are not.)
20. The stock market fluctuates many times more than would be suggested by its future stream of earnings and dividends or by the GNP, both of which are historically remarkably stable: i.e., the market is driven by greed, fear, and career risk, not economics.
21. Inflation is the primary influence on P/E levels in the equity markets however illogical that may be for a real asset.  The correlation coefficient is -.73 in the U.S.:  low inflation ‘explains’ or is coincident with high P/Es.
22. But since inflation is probably mean reverting, and certainly unstable, buying when inflation and interest rates are low and P/Es are high will mostly be painful.
23. Size of assets under management is the ultimate barrier to successful investing.  As assets grow, you are forced either to pick increasing numbers of decreasingly good stocks or to buy larger, indigestible positions of your original holdings.  The investment business is the perfect example of the Peter Principle: do well with $500 million, and they’ll give you $5 billion.
24. In the good old days, little talent came into the business as belief in efficient markets discouraged serious quants in particular.  Now finance professors run quant shops and vastly more talent is drawn into the business, painfully increasing competition.
25. Quantitative investing is to traditional investing as the written word is to the spoken: you believe it more and can march confidently off the cliff.
26. Quants also find it irresistible to put in just one more variable and risk drowning in data mining.
27. Quants naturally prefer the mathematically neat to the rugged and simple.  A sign on every quant’s wall should read:  “There are no points for elegance!”
28. For quants, the advantage lies in their ability to handle complexity with speed and consistency. Quants also never fall in love with a stock – just methodologies.
29. The most critical advantage for quants, though, is that they can build on the past, remember mistakes, and pass on all their accumulated knowledge.
30. 90% of what passes for brilliance or incompetence in investing is the ebb and flow of investment style (growth, value, small, quality).
31. Since opportunities by style regress, past performance tends to be negatively correlated with future relative performance.
32. Therefore, managers are harder to pick than stocks.  Clients have to choose between facts (past performance) and the conflicting marketing claims of several potential managers.  Practical clients will usually feel they have to go with the past facts.  They therefore rotate into previously strong styles that regress, dooming most manager selections to failure.
33. Getting the big picture right is everything.  One or two good ideas a year are enough.  Very hard work gets in the way of thinking.