I will survey what (little) is known about the common theory of fields that are algebraic over the field of rational numbers, and I will explain why on the other hand we have a very good understanding of the theory of sentences that are true in "almost all" such fields, in a suitable sense.