people are buying Mac Minis specifically to run AI agents with computer use. They’re setting up headless machines whose sole job is to automate their workflows. OpenClaw—the open-source framework that lets you run Claude, GPT-4, or whatever model you want to actually control your computer—has become the killer app for Mac hardware
Is the author implying mac minis for the low power consumption?
It doesn't make sense because it's a lie. The author's blog has 2 articles, both of them shilling OpenClaw.
Spoiler: the author is an OpenClaw instance.
At least on clackernews.com they're upfront about it - it's a HN-style forum where only bots can post. No pretending to be human required.
Exactly. See also this sentence:
> Look at who’s about to get angry about OpenClaw-style automation: LinkedIn, Facebook, anyone with a walled garden and a careful API strategy.
Browser automation tools have existed for a very long time. Openclaw is not much different in this regard than asking an LLM to generate you a playwright script. Yes, it makes it easier to automate arbitrary tasks, but it's not like it's some sort of breakthrough that completely destroys walled gardens.
If you’re heavily invested in Apple apps (iMessage/Calendar/Reminders/Notes), you need a Mac to give the agent tools to interact with these apps. I think that combined with the form factor, price, and power consumption, makes it an ideal candidate.
If you’re heavily invested in Windows, then you’d probably go for a small x86 PC.
Some of those connectors are only available on the mac and some only on the iPhone. Like notes is available on the mac, but not on the phone. Vice versa for reminders.
Can you imagine giving an AI access to your messages, notes and calendar though?
I use agentic coding, this is next level madness.
I used Claude Code (CC) to make my own MCPs for these apps. I gave it read/write access only, no ability to delete. Of course it could probably code it's way into doing that since it can access the MCP code. I don't run it in --yolo mode though.
I interact only with CC on the machine and watch what its doing, I haven't tried OpenClaw yet.
Here's some workflows I've personally found valuable:
- I have it read the "Grocery" Reminders list and find things I commonly buy every week and pre-populate the grocery list as a starting point. It only adds items that I haven't already added via Siri as the week goes on. For example, I might notice I've run out of cheese and I'll say "Hey Siri, add cheese to grocery list". The list is shared via iCloud Reminders app between my spouse and I.
- Pre-CC, I wrote an OR-Tools python tool for "solving" the parenting time calendar. My ex and I work inconsistent schedules each month. Each month I was manually creating a calendar honoring requests, hard constraints, and attempting to balance custody 50/50. CC uses the MCPs to fetch the calendar events and review emails related to planning. It then structures everything as JSON as inputs to the optimizer. The optimizer runs with these inputs and spits out a few "solutions". I review the candidate solutions and select one. CC uses the MCP to add the solution to the calendar. This one saves me probably an hour every month.
- CC uses an email MCP to fetch emails from my child's school and suggest events its found in the emails to add to the calendar.
None of these are huge time savings on their own but the accumulation of reducing the time to complete these repetitive tasks has been awesome in my opinion. These are all things that most definitely would not have been worth automating with traditional dev work but since I can just dictate to CC for a few seconds and it has something that works a few minutes later it's become worthwhile.
I guess what’s wrong with it? Let’s say it has read only access, new messages and calendar invites need approval. I’m not sure I understand the harm? I suppose data exfiltration, but like you could start with an allowlist approach. So the first few uses and reads take a while with allowing the ai to read stuff , but it doesn’t seem that crazy given it’s what we basically do with ai coding tools?
I think (most of) them register new accounts for the agent.
they're buying mac minis because it's the cheapest way to get a computer with iMessage access to stuff in a closet and leave on at all times. having access to your iMessage is one of the most interesting things openClaw does.
Yep, there is zero reason to use mac mini’s. It’s way more cost effective to rent one (or more!) small VMs the cloud.
I have seen dozens of people/videos talking about buying Mac minis for clawdbot.
I don't understand why, but I've seen it enough to start questioning myself...
Wouldn't it run on a $50 raspberry pi?
Probably the same people getting a macbook pro to handle their calendar and emails
I thought I had heard that the integrated RAM/VRAM makes local LLMs fairly quick on a RAM-maxxed Mac Mini.
The software can drive the web browser if you install the plugin. My knowledge is 1.5 weeks old, so it might be able to drive the whole UI now, I don't know.
Welcome to the AI meme race where everyone's knowledge is about 1.5 weeks old :)
It has nothing to do with running models locally, its perfect because its incredibly cheap, capable, small, and quiet.
The author is full of shit is what it is. They see a few posts online and extrapolate from that to fit whatever narrative they believe in.
Claude/GPT-4 don't have any GPU requirements?
No dude, you send text or images and get the same back, it's all cloud.