Paternalism is the idea that the government should restrict citizens' freedom for their own good. Examples abound - prohibitions on alcohol/smoking/drugs, prohibitions on risky financial investments etc. Paternalism leads to a special case of government regulation - the kind which is intended to benefit the individual, rather than "society" or nation (examples of the second kind are restrictions on free trade, and other innumerable economic regulations).
Robin Hanson, over at the Overcoming Bias blog, says that paternalism is about bias
... bias claims seem to be central to paternalism; regulators and citizens each think the others are biased.
To evaluate if paternalism is good or bad, we need more than the sort of evidence that would convince regulators that they are less biased than citizens, or that would convince citizens that they are less biased than regulators. After all, we expect each group to be biased in underestimating their own bias.
Without such evidence, paternalism is just arrogance, i.e., an unsupported presumption by regulators of their own superiority.
It's good to see him highlight the bias of regulators, but it's an argument that has been made before by libertarians, albeit in stronger language.
Eliezer Yudkowsky (who's the author of the highly recommended article - Cognitive Biases Potentially Affecting Judgment of Global Risks
(pdf)) makes some excellent observations in a comment
on the post. They say a fool and his money are soon parted - regulation is an uphill battle against this. Regulation happens because society is unwilling to accept that they can't altruistically restrict the choices of fools without doing greater harm to the rest of the society (and even to the the fools, the intended benefit group itself).
The brutal tradeoff: You can open stores for banned products, but people whose sole fault was to be born stupid will shop there and get hurt. There would be social benefits too, like the ability to use important medications that the FDA will take another five years to approve. These benefits will accrue to more people than just the cognitively advantaged. But there are stupid people and they will get hurt and they don't necessarily deserve it. This is what people can't face - they don't wish to think of themselves as uncompassionate - and this is why society tries to impose regulations regardless of the cost-benefit tradeoff.
I don't think all regulation is bad, nor that we should see all regulation through the same lens - i.e., for example, one can't just use the same arguments against alcohol prohibition and against drug prohibition - the usage patterns of drugs and alcohol, have been historically different and are different now also, there is different levels of societal acceptance to alcohol and drug consumption, and they have different effects on the mind and body. I wouldn't be convinced by arguments that treat one as analagous to the other - what I would be convinced by are cost-benefit calculations showing that both prohbitions are damaging to the population at large.
But that's me. What Eliezer is saying is that utilitarian arguments just won't cut it with the rest of the society. Most people wouldn't care to understand the ultimate effects of regulation - they would rather comfort themselves with the thought that they have done *something* to help people.