Apple’s much-touted “Intelligence” features and revamped Siri launch have so far fallen flat—and even prompted legal threats. Now, the company appears to be doubling down on AI by building a fresh, LLM-driven Siri from the ground up.
Bloomberg’s Mark Gurman reports that Apple’s AI ambitions lagged behind rivals, hampered by skepticism and underinvestment. The “Apple Intelligence” effort didn’t materialize until late 2022 after ChatGPT’s debut, and the company was hesitant to purchase the GPU power needed for large-scale models. John Giannandrea, Apple’s AI lead, reportedly doubted consumers wanted chatbots on their phones, and continually bolting new AI tricks onto classic Siri only introduced more problems.
As one insider put it to Gurman, “It’s whack-a-mole—fix one bug and three more pop up.” To address this, Apple is now creating an entirely new Siri built around a large language model, aiming for more natural conversation and superior information synthesis. The update will also give Siri live internet access—allowing it to pull together data from various sources in real time, much like Perplexity does. Clearly, Apple is rethinking its approach to AI and gearing up to finally deliver a more capable Siri.