Starting from suitable large cardinals, we force the failure of (weak) diamond at the least inaccessible cardinal. The result improves an unpublished theorem of Woodin and a recent result of Ben-Neria, Garti and Hayut.1. Introduction. We study the combinatorial principles diamond, introduced by Jensen [9], and weak diamond, introduced by Devlin-Shelah [5], and prove the consistency of their failure at the least inaccessible cardinal.Suppose κ is an uncountable regular cardinal. Recall that diamond at κ, denoted ♦ κ , asserts the existence of a sequence S α | α < κ such that for each α < κ, S α ⊆ α, and if X ⊆ κ, then {α < κ | X ∩ α = S α } is stationary in κ.Also, weak diamond at κ, denoted Φ κ , is the assertion "For every c : 2 <κ → 2, there exists g : κ → 2 such that for all f : κ → 2, the setIt is easily seen that ♦ κ + implies 2 κ = κ + , and in fact by a celebrated theorem of Shelah [12], for all uncountable cardinals κ, ♦ κ + is equivalent to 2 κ = κ + . It follows that it is easy to force the failure of diamond at successor cardinals.By [5], 2 κ < 2 κ + implies Φ κ + . It was later observed by Abraham and Baumgartner that Φ κ + implies 2 κ < 2 κ + : see for example [7], [11] where the proof is attributed to Abraham, or [13] where the proof is attributed to Baumgartner. It follows that Φ κ + and 2 κ < 2 κ + are equivalent, and ♦ κ + implies Φ κ + . It also shows that it is easy to force Φ κ + to fail.Unlike the case of successor cardinals, it is difficult to force the failure of diamond or weak diamond at an inaccessible cardinal.