One of these has a semantic error that is fixed in the other. Can you find which one??

value_iteration

def value_iteration_A(probs, gamma, rewards, oldV):
   newV = oldV.copy()
   nstates, nactions = probs.shape[0], probs.shape[-1]
   for s in range(nstates):
      weightedV = np.zeros([nactions], dtype='float64')
      for a in range(nactions):
         weightedV[a] = np.dot(probs[s,:,a], oldV)
      newV[s] = rewards[s] + gamma * weightedV.max()
   return newV
   
def value_iteration_B(probs, gamma, rewards, oldV):
   newV = oldV.copy()
   nstates, nactions = probs.shape[0], probs.shape[-1]
   for s in range(nstates):
      weightedV = np.zeros([nactions], dtype='float64')
      for a in range(nactions):
         weightedV = np.dot(probs[s,:,a], oldV)
      newV[s] = rewards[s] + gamma * weightedV.max()
   return newV