Questions for Will MacAskill's fireside chat in EAGxAPAC this weekend

It looks like I'm  too late. But here's something I've been wanting to ask. 

In your paper "The Definition of Effective Altruism," you distinguish effective altruism from utilitarianism on various grounds, including that:

• EA does not claim that a person must sacrifice their personal interests (e.g. having children) when doing so would bring about greater good; and
• EA does not claim that a person must violate non-consequentialist constraints in the rare situations when doing so might bring about greater good.

For me, this points to a broader principle that EA does not require a person to sacrifice something "morally major" to bring about greater good. This would imply that an EA can choose to prioritize things like a duty to contribute their fair share, a duty to family members, and a duty to rescue those they are uniquely positioned to rescue over bringing about greater good. 

However, in a 2015 debate, you argued (scenario; response) that a person alone in a burning building should choose to rescue a Picasso painting (assuming they can keep it) over a child since the money from selling the painting could be used to save thousands of children. 

Do you think effective altruism necessarily entails that position or were you just speaking to what is morally better?