Apple Mac M1

Untrue to say they don’t last, our M1 macs 6 years later are still very usable. However your point is very true for corporates, staff always want a refresh regardless if the laptop is good enough or not. These personal plans make sense for people who want the latest and don’t like the hassle of reselling

They're fine if you just need it mainly for the build/bundle step. I've used an 8GB M1 mac for react-native iOS dev for over a year back when the M1 first came out. If I had to look for the cheapest way to build iOS apps, but still on a laptop form factor (aka no mac minis), and I don't use macOS for day-to-day development, I'd do exactly this.

This downloads a library of mostly kiwix wikipedia information and produces a full text SQLite index that is live searched by the model and then the model reads and summarizes the information. Interesting thing is that when I asked a follow up question it answered but showed: This answer came from what the AI already knows, not the library, so there's no source to check. Ask “look that up” to get one. When I asked "look that up" it was able to. It highlights relevant sections. The information it was working off in my case was WikiHow and there aren't any citations in that information. Running on a 32GB Mac M1 using Qwen 27B. Ollama is showing as taking 19GB steady state and 23GB while answering on Activity Monitor (sometimes a little more and sometimes a little less). Its amazing this can be done, but it seems like someone should be able to develop a much smaller model that can be used reliably for this use case. There are smaller models available to use and download but it is stated that they have some reliability issues.

I recently had some relevant experience: for a couple of months now I've been experimenting with on device models to summarize feeds in a Bluesky client I am developing. The feature extracts topic areas, categorizes posts, and creates a summary under each topic. At first the results were hot garbage, and progress was slow. I hooked up the settings to download models from Hugging Face conveniently, so I could run experiments faster, and I massaged the prompts a bit. Last week this feature made a qualitative jump from science experiment to something I'd actually use. The fact that all runs on the device means I've got no variable costs associated with adding this to what will be, at best, a pretty low revenue product. I've tested it on trailing edge devices like an M1 Mac and a Pixel 8, and performance is very tolerable. The key is I'm not asking for open ended answers to open ended problems. When it proves to be useful it's not going to get less useful or more expensive. There are vast domains of uses for LLM models with similar characteristics and likely similar results.

The performance gap between RISC-V and x86 or Arm is in fact closing. The SpacemiT K3 machines which most people who ordered in early May have now received are comparable to the RK3588 and Pi 5, with Rock 5 delivered to customers in mid 2022, Orange Pi 5 at the end of that year, and Pi 5 in October 2023. So that's at most a 4 year gap, less than 3 years in the cast of the Pi 5. That is the highest performance Arm64 machine most SBC users have. The faster CIX P1 exists but it seems that very few people actually have Orion O2 or Orange Pi 6 Plus. Before the launch of the K3, the previous gen early 2022 to early 2024 JH7110, TH1520, K1 were somewhere around 6 years behind Arm SBCs. RISC-V machines expected late this year (let's say early next year) will be similar to CIX P1, so just a 1-2 year gap. Vs x86 the previous generation RISC-V was something like one of the last Pentium III or PowerPC G4, while the K3 is mid range Core 2 verging on early i5/i7 in many regards. So that's caught up on Intel by something close to ten years in four years. And the next gen will be somewhere around Zen 2 or whichever Skylake iteration is in the same ballpark. Even more importantly than the gap, once that SkyLake to Zen 2 to Apple M1 performance band is reached, that is a performance level that remains "good enough" in 2026 for most users of computing devices for their everyday web browsing, media consumption, productivity/business app uses. I'm typing this on an M1 that I sit at all day every…

I had that problem as well. Especially when connected to two external monitors. I did not love the machine and the M1 Max was such a big upgrade because of that and I could upgrade to the m3 max later and give my m1 to somebody else. Both apple silicon machines are going strong for a long time I guess.

My 32GB RAM M1 is a fantastic machine, but it's getting old for the type of work I do. And I can't upgrade the RAM. I installed Omarchy via Asahi Alarm [1] and have been running it for 3 weeks without issues. I spent half a day setting everything to my taste (keyboard bindings, monitors, waybar, etc.), and I'm still getting used to Hyprland, but I have a fast machine. RAM usage with everything I need open is considerably lower than on macOS (80% -> 50%), and I don't have to pay a monthly subscription for every little piece of software not built into macOS. Asahi isn't perfect (touchpad rejection being my main issue), but I can live with the inconveniences. This is still a very good machine, and I can't afford a new one right now. [1] https://github.com/basecamp/omarchy/pull/1897

As long as it can be disabled it is a complete nothing burger. If it can't be disabled and ends up crippling my otherwise perfectly adequate M1 it will accelerate me to switch to daily use of Asahi Linux. Edit: That's assuming the effects are even expensive for the GPU.

I actually hooked up my magic trackpad to Linux at some point because I got so frustrated with the crappy touchpad on my Samsung laptop. I also have a logitech wireless mouse for it. Anything is better than that crappy touchpad. The point is that Apple's magic trackpad actually works great on Linux too. Smooth, responsive. accurate, multi touch and gestures, and all the rest. Just works. More or less exactly like it does on a mac. Too bad the blueooth stack on Linux is a bit unstable. Lots of issues with stuff randomly not connecting. Which of course isn't great with a trackpad. The issue is that trackpads from other manufacturers just seem to be universally really, really bad in comparison to Apple's hardware. Particularly anything produced by Synaptics that I've had the displeasure of using is just mediocre shit in comparison. And they seem to dominate the market. It seems like they just gave up even trying to pretend to compete. If you see somebody using a wireless mouse, 9 out of 10 times they aren't using a mac. I work in a lot of co-working spaces. Lots of macs. Almost exclusively being used without a mouse. Just not a thing. The trackpad that comes with it is fine. If you see somebody using a mouse, it's usually with a windows/linux laptop. That Samsung laptop was something I used in between two macs. My old intel mac died weeks before the M1 was supposed to come out. I used it for about half a year. I still have it and it runs Manjaro. From a software point of view, I…

I swapped from Linux to MacOS when the M1s came out, and I love the integration with all the iCloud stuff (particularly Messages). Occasionally I miss being on Linux, as somebody who did so for 20+ years before making the switch. But on Mac, stuff actually does Just Work. Reading this makes me a little misty-eyed and I miss my solid old Thinkpads from 10-20 years back.

Yep. For my own company I used a bare metal machine in Hetzner running Linux and a Windows VM along with a bunch of old MacBook Pros wired up in the home office for CI. It works, and it's cheap. A full CI run still takes half an hour on the Linux machine (the product [1] is a kind of build system for shipping desktop apps cross platform, so there's lots of file IO and cryptography involved). The Macs are by far the fastest. The M1 Mac is embarrassingly fast. It can complete the same run in five minutes despite the Hetzner box having way more hardware. In fairness, it's running both a Linux and Windows build simultaneously. I'm convinced the quickest way to improve CI times in most shops is to just build an in-office cluster of M4 Macs in an air conditioned room. They don't have to be HA. The hardware is more expensive but you don't rent per month, and CI is often bottlenecked on serial execution speed so the higher single threaded performance of Apple Silicon is worth it. Also, pay for a decent CI system like TeamCity. It helps reduce egregious waste from problems like not caching things or not re-using checkout directories. In several years of doing this I haven't had build caching related failures. [1] https://hydraulic.dev/

Using an Intel Mac by any chance? I had a 2019 MacBook Pro and I remember it being just like this. Insanely slow and hot all the time. On my M1 almost everything is instant. Easily the most responsive computer I’ve ever used.

Intel giveth, Microsoft taketh. This is even more sad with Apple. My M1 Mac felt incredibly snappy with Big Sur, but is getting ever so slightly slower with each update. My iPhone SE2 is a complete disaster.

Exactly. And especially for my incredibly mixed use case, which includes some light gaming in crossover, the gpu improvements in m4 are apparently non-trivial. I’m sure whichever I end up with will be a pretty big upgrade over my basest of base model 32GB M1. The thing I’m most curious about on the new machines, especially the Ultra, is the thermals. I only care about perf per watt if it becomes unfavorable enough that the fan spins up above idle during normal Tasks. On my M1 the only way to get it to audibly spin up is to get the machine to near total load nd hold it there for some time.

I agree with you here. My solution was to use Linux on my Desktop PC and its been great for work and gaming, and I have an M1 Mac with Asahi Linux on it for on the go. I love it and I'm really happy with it. I've used Windows most of my life and is has really worn me down, and I tried MacOS for around a year, and it is equally as painful but in different ways. Linux has some rough edges in certain areas, but I like that I can easily sync all my devices, emulation files (it pairs well with the steamdeck), and configurations... and I don't have to put up with being a data cow or being constantly badgered to engage in vendor lock-in.

I have an m1 as well. And while I'm broadly satisfied with its performance, I do think that the SSD is probably carrying some of that load. And for a machine that often gets used far longer than a PC, I can't see that being great for longevity.

Probably not? I can't really tell a difference between the M1, M2, and M3 machines I've had. There are some spec upgrades for peripherals connectivity, though (Thunderbolt, Wifi, external monitors), if that matters to you. And the M4 Macbook Pro finally has that etched anti-glare screen if you ever want to use it outdoors. Otherwise, the M1 is still a very capable machine and miles better than the PC laptops we all grew up with. Does anything in your day to day work actually feel slow? If so, maybe you can benchmark it and see if the upgrade would actually help much. If it's single-threaded work, especially, there may not be a big upgrade. If your workload is heavily multi-threaded or GPU heavy and you're on the base M1, the M4 Pro or Max could be a big upgrade, but most coding work probably isn't.