Even Apple cannot explain why we need AI in our lives

Original Article ↗ Hacker News Discussion ↗

Apple Intelligence: Substance vs. Hype

  • Many see Apple’s AI reveal as underwhelming and derivative, more like catching up to others than defining a new category.
  • Some criticize the use of the old “for the rest of us” tagline given the modest, partly third‑party feature set and lack of live demos.
  • Others argue the on‑device / Private Cloud Compute approach and UI integration are meaningful, especially for text summarization, proofreading, and notification/email triage.

OpenAI Integration, Privacy, and Branding

  • Dispute over how much Apple depends on OpenAI:
    • One side says Apple Intelligence itself runs on Apple’s stack, with ChatGPT only invoked explicitly for certain tasks and behind a permission prompt.
    • Others counter that from a user perspective, Siri using ChatGPT is Apple using OpenAI and that the “Apple Intelligence” branding may overstate how private/local everything is.
  • Commenters note Apple makes ChatGPT involvement clearly labeled, with indications it will ask permission each time.

Do Consumers “Need” AI?

  • Strong view that everyday consumers don’t need AI; code completion and writing tools help professionals more than typical phone users.
  • Another camp says most people find current UIs like “casting spells,” and natural‑language interfaces could finally make complex tasks approachable “for the rest of us.”
  • Some phone owners report barely using existing AI features like Circle to Search or enhanced assistants.

AI Hype Cycle, Bubble Talk, and Nvidia

  • Several commenters argue current AI products are underwhelming relative to 2023 hype (AGI, mass job loss, “insane value”), leading to a sense the bubble is nearing a pop.
  • Others say there’s no objective sign of a pop yet; valuations and GPU demand keep rising.
  • Debate on whether Nvidia’s dominance is justified:
    • One side likens it to being the sole locomotive manufacturer in a railroad boom, with cloud providers profiting by renting GPUs.
    • Skeptics question whether end customers are yet making enough money to justify the massive infrastructure spend.

Use Cases, Limits, and Risks

  • Recognized strong uses: code completion, debugging help, drafting/corporate writing, search‑as‑conversation, article summaries, basic image generation.
  • Serious concerns about hallucinations/bullshitting and the danger of over‑trust, especially in high‑stakes areas like banking.
  • Debate over LLMs in education: some see “cheating,” others see tools akin to calculators or encyclopedias, useful if paired with critical thinking.
  • Broader view: LLMs likely won’t be AGI soon but may become the primary interface to services, displacing many traditional apps and web UIs.