Just because they're better at writing CS algorithms doesn't mean they're taking steps closer to anything resembling AGI.
Unless AGI is just a bunch of CS algorithms.
Kinda depends on how much is "a bunch" and how fast that AGI is