Apple's New Mac Line is Built Around Local AI. The Bet Is You'd Rather Own Than Rent.

Comments · 853

• DM

  @dmdunn11d · 4 weeks ago (edited)

  I bought my first McIntosh in the mid 90s. For the last 31 years I’ve been surfing the computer tech wave as a desktop publisher. My learning curve began with Aldus PageMaker and continued into the present with Adobe Creative Suite. The energy of this evolving technology has kept me on the board. In 2025; My last paid gig was designing and desktop publishing a memoir for a longtime partner, client, and friend. <br><br>I’ll turn 84 this fall. AI offers the energy of a new wave, a new learning curve, and new creative-ventures-in-retirement. The ability to ask an AI assistant “how do I do this,” gives me the safety of a smart outrigger on a new board.

  97

• PM

  @pmbrig · 1 month ago (edited)

  I&apos;m not a &quot;prosumer,&quot; I&apos;m an ordinary bloke embedded in the Apple ecosystem. I think there are a lot of us who want AI to streamline daily tasks and have no need to bespoke crunch huge databases. I&apos;m solidly in the local-AI/cloud-when-needed camp. I&apos;m still suspicious about cloud models storing and accessing all my data. Now, I know that there are also lots of folks who need something more for sophisticated and more power to them, but I don&apos;t need to pay an enterprise model for monthly tokens with questionable privacy. I think you&apos;re right on target about both models doing well for different user bases.

  105

• PE

  @peterscheepers5780 · 3 weeks ago

  I’m a retired industrial control engineer and in the controls industry we have seen the shift from centralized to distributed to the extend that sensors are even highly intelligent and still getting more affordable. Keeping the intelligence as close as possible to the place it is needed is a fail soft benefit, will be &nbsp;cheaper in the long run and gives you control over your own data. You don’t need a Bentley to drive to the supermarket, good enough will do. Yes I think there will be a future for frontier ai but i expect the majority of ai to be distributed locally.

  5

• ZE

  @zebratroopfan6695 · 3 weeks ago

  I use my computer to organize my life. &nbsp;I track my health, finances and schedule my daily tasks. &nbsp;I don&apos;t code, edit or design. I check my email. I watch a bit of YouTube. I hope I&apos;m reminded what I&apos;m supposed to do or be for the day. &nbsp;I’m in the Apple ecosystem and i like it. The new siri is exciting because it is using my personal data on my phone to make my life function a little easier. &nbsp;It’s that simple for me.

  4

• KE

  @keme8541 · 1 month ago

  The slow down in frontier releases means open weights have nearly caught up. I vote for local AI, as privacy is vital for proprietary data.

  9

• BA

  @BavariaR · 1 month ago

  I think one of the main motivations for local AI models is confidentiality, sovereignty, and independence from large frontier models that could disappear anytime at the whim of others. If we had a setup with cloud models not owned by a single greedy company but owned jointly by humanity, and that could ensure our data is safe and the models are not biased, then many individuals would no longer care about purchasing hardware to run models locally.

  33

• JC

  @jc2044 · 1 month ago

  The problem with cloud services is the uncertainty they introduce: service availability and reliability, data security and privacy, and the constant risk of unexpected changes to policies, pricing, or terms of service.<br><br>The number of horror stories just keeps piling up. It is difficult to ignore the risks of putting critical data and operations in someone else’s hands.

  13

• CO

  @coonhound_pharaoh · 1 month ago

  Apple is on the record stating their excitement about technology introduced in the M7 made investment in M6 get cut short. The reason the M6 is only in cheap units is because Apple is gearing up for the M7.

  65

• RA

  @RaymondSiu_Love_God · 4 weeks ago

  I think you’re missing one major factor: data privacy. <br> <br>For non-technical users, data privacy is often a bigger concern than performance. Running an LLM on our own hardware can significantly reduce the risk of data leakage. <br> <br>Our team is currently piloting Apple hardware to see whether it can run LLMs locally while still delivering results comparable to frontier models.

  2

• CH

  @Christ-on-a-Bike · 1 month ago

  Local AIs for image and 3D model generation are going to be a really big thing. Because in a strange way there is kind of a finite amount of AI you need there, you are already seeing it in some of the most advanced image and even video generation models.

  5

• ST

  @StephenFlournoy · 1 month ago

  100% worth it for local models. &nbsp;Easy to solve the routing issue, just use frontier model to build the router &quot;you need&quot;. &nbsp;I&apos;m already doing this sharing my compute with a friend.

  20

• BA

  @Barry-z7e · 3 weeks ago

  Thank you.... everything you mentioned makes sense to me, and has increased my knowledge about how AI is evolving one way or another. I will add, that local AI for me is a must for privacy and data security. Using the best Mac I can afford and running AI locally while allowing me to use all my Mac-like tasks is the combo I will embrace.

  2