Keep in mind that this is appearing in a magazine that is successful because of the success of high frequency trading and that the article is written by the member of a company that bases its profits on the ability to conduct high frequency trading.
I do think it's legitimate to look at sources for articles, especially when there are strong conflicts of interest.
Romans 13:1
That being said (disclaimer: I am a HFTer), this article is a pretty weak defense of HFT. Like others have pointed out it makes a lot of assertions not based on data/fact. However, I've yet to read a criticism of HFT that doesn't commit the same mistake (and I've read a lot of them).
For a defense of HFT that uses hard data I'd direct people here: http://papers.ssrn.com/sol3/papers.cfm?abstract_id=1501135
Is it clever arbitrage? is it a massive denial of service so some people can play middle man? I don't know.
However, Cameron Smith has created an excellent straw man of his opponents. This is an ad hominem argument.
The article contains a number of assertions, central to its claim but with little evidential support in the article. (For instance: "No serious market observer disputes the claim that volatility would not be higher without the liquidity provided by high frequency traders" -- has the author really looked hard for serious market observers with a different opinion? -- and "High frequency traders can only trade profitably when their trades push a stock price towards fair value"; what exactly is "fair value" supposed to mean, why shouldn't there be short-term bubbles among HF traders just as there are long-term bubbles among slower traders, and who says HF traders can't all happen to trade unprofitably in some particular case?.) It may very well be that those claims are true, but we basically have to trust the author. And that is what we may quite rightly and rationally be less inclined to do, if we know that the author has a vested interest in persuading us.
The 'fair value' he's talking about us just the supply-meets-demand of that stock for that particular moment-- that moment could definitely take place in a larger bubble of that security. The mini-bubble within high freq trading he refutes by reasoning only ("another hft would detect this"), so I think a more concrete example would've been more convincing. As with all bubbles the question is how far out of line can the price get, and how quickly will the price be brought back to reality