I agree, but my M1 MacBook work laptop is by far the fastest dev machine I've ever laid hands on. It struggles a bit from the UX standpoint, two things:
1. I have the same desktop layout every time, from left to right: slack desktop app, a two column wide emacs window, a 90col wide terminal. I also have two chrome windows--1 which is the same width as the slack window and overlays it, and another the same width as the emacs window which overlays that one. The problem is every single time I wake my laptop from sleep the terminal window has shrunk to fewer columns and I have to drag it back to full width.
2. Sometimes the external monitor support bugs out. I don't know if that's my hub ("pluggable" something or other) or the OS or both.
Then of course there's all the warts of homebrew, and the fact that it's not easy to build some software..
However, the performance of the Apple silicon is nothing short of astonishing. I'm curious about the AMD chips that ship in the new Framework as I look towards an upgrade to my personal laptop, but it's basically between that and a new M4 Max Macbook. Never thought I'd see the day.. will probably wait a year or so before deciding but it's interesting that Apple is even a contender.
For the 1st point, I struggled with the same, then a short applescript thanks to claude, attached to a keyboard shortcut on karabiner solved the painpoint. There's also an app called Stay which should do something similar with a ui, but my solution is good enough for me at this point.
Reading back I see I came across as kind of an asshole, I'm sorry for that. I genuinely thought at the time you might be kidding but I wasn't sure--pretty sure I got it wrong. I've grown (probably too) cynical about ux tweak over the years, in the past I did it a lot (even ran stumpwm and i3wm for a while) but over time I kept getting burned by the effort it takes to maintain nonstandard stuff and gave up. I appreciate it's fun, I'll try to remember that.
Thanks for your 2nd reply :)
I'd also prefer everything worked out-of-the-box, but where the world fails, you unfortunately need to patch it. If it would have taken me more than 30 minutes I'd probably skip and learn to live with it, but it was kind of easy and i find it useful every day.. for the cynical part, i think you can't beat me on that, I'm in my late 40s and seen too much :)
No, it's just that I don't customize systems anymore. It's not worth the hassle of maintaining a bunch of extra stuff just to fix issues that shouldn't be there in the first place. All those extra ux tweak programs just end up costing time to keep them working across updates, etc. because they're not part of the system.
That would've been a much better comment in the first place. I do understand where you're coming from, though. In my case, however, a single Brewfile is enough.
My favorite is how when I close my macbook with an external display plugged in, the laptop screen remains on (and lit up!) with seemingly no way to configure this behavior. Sometimes a window will end up on that (non-visible) screen which can be very confusing.
That seems like a misconfig or a broken lid sensor or something. I’ve been using MacBooks with a single external monitor as my only display (MacBook closed) for over a decade and I’ve never had the laptop display stay on when closed with an external display. Maybe time to visit the Genius Bar?
I'm curious about your dock/hub. There's a good chance I just ordered the same one as I try to build a more sophisticated home station. Which one and how does it bug out?
Pluggable TBT3-UDC3, input freezes and screen goes all pschedelic with greens and violets and then patterns and lines as everything melts to white noise. Flipping the lid open and closed and/or unplugging the thunderbolt cable and plugging it back in again repeatedly seems to cool the vibes
That sounds prudent. There's no reason for them to continue this embarrassment of cramming the product full of worse-than-useless "AI" features. Wait until you can separate the good from the bad from the ugly and choose to just do it good.
The best thing that JSON is good at is human readability. I would say human composibility as well, but YAML and others are better for that. As a machine data exchange and storage format, JSON is extremely inefficient.
I really dislike writing YAML. It might be a skill issue (as the kids say), but it just never landed. I'd rather write JSON, even with all its verbosity.
It never ceases to amaze me how good computer people are at solving problems which are not problems (readability of serialized representation) at the expense of multiple orders of magnitude more compute and storage.
> problems which are not problems (readability of serialized representation)
I don't know that I fully agree. A format that is both human- and machine-readable is very useful, for configuration and debuggability for example. It's indeed more expensive in compute and storage, but that's additional work _for machines_ rather than _for humans_ (having to compile/encode/decode data at the human/machine interface).
Certainly doesn't mean it's always the right trade-off, but it's an appealing one, as the popularity of these formats shows, in particular JSON which isn't even that good for human-readability (eg no comments in standard JSON).
But that's the thing: if we want to optimize both machine and human experience, we should create a 1:1 dual format, one which is machine-readable and one which is human-readable. A standard tool easily converts one to the other. This should not still be a problem. Yes, we have computers and should offload work onto them. But exactly what that work is is still important. Two approaches aren't equal because they both put the burden on the machine.
> But that's the thing: if we want to optimize both machine and human experience, we should create a 1:1 dual format, one which is machine-readable and one which is human-readable. A standard tool easily converts one to the other.
There's things like the Anduril Bolt. They cost like 100x as much as devices the Ukranians are building from cardboard. Another major innovation is TOW style fiber optic control which is immune to electronic countermeasures. There's definitely a lot to learn, but sadly necessity is the mother of invention and not since WW2 has manufacturing cost really been a serious concern for US defense production. Seems like we'll need a really big war to make that essential, and then there's an open question whether we'd actually be able to do it given that nobody knows how to do anything anymore.
> Seems like we'll need a really big war to make that essential, and then there's an open question whether we'd actually be able to do it given that nobody knows how to do anything anymore.
I'm not sure if I'd be rooting for this eventuality.
I think there's another really important set of lessons available from basic tool use which translates directly to the software industry--intuitive understanding of what makes a tool good. Tools (as opposed to appliances) scale with the user's ability. A good lathe in good condition does better and better work as you learn its behavior and capabilities. You could spend 5 decades with one hammer incrementally improving your forging technique day by day and week by week. Your dishwasher, however, just always does the same thing.
Knowing the difference between a dishwasher and a hammer is something it seems like many of the engineers, designers, and product managers in the software business are completely incapable of.
> If you want shop to be relevant to modern industry you would need to teach robotics, CAD/CAM, CNC, 3D printing, etc.
Uhh.. what? A CNC mill is, fundamentally, a mill. A CNC lathe is a lathe. You're not absolved from knowing how to use a manual lathe or a manual mill even if you use CNC machines all day long. Where do you develop an intuition for feeds, speeds, finishes, tolerances, etc if not by spending hours and hours doing it by hand?
A classic machinist apprentice task is to take a rough steel cube and make the sides flat and smooth using a hand file. That teaches you how steel behaves and can be worked. Life was slower a century ago.
High school students can learn CNC with smaller machine tools. There are little desktop CNC machines in the US$1000 range. You can cut aluminum, brass, and plastics, but not steel.
They talk the same G-code as the big machines. You design jobs for them the same way you do for the big machines. At that scale you can usually avoid coolant, oily rags, oily chip disposal, and the general mess of a real machine shop.
Taking that lesson to its extreme you realize there's a rounding effect where seemingly no matter how carefully you hold the file, the edges of the face are cut down more than the center. To combat this you might discover you can use the end of the file (like the last tooth) as a scraper. You can use a gauge block and some blue ink to discover the high spots and repeatedly work them until it's flat. Congratulations! You've just discovered how to make a lathe bed.
I tried to learn machining on a CNC mill. it went ok, I guess. but I ruined a lot of tools, had to do a lot of polishing passes, and didn't really understand why my tolerances were so off. things like the flycutter mystified me, and using the boring head was deeply frustrating since it didn't have a tiny little servo. aligning the vice was something I dreaded, took me a good hour for some reason. 4 jaw chuck? maybe you use that if you need to drill off center or work on square stock.
couple of decades later I only use manual machines (with power feeds and DROs), and I'm really a lot faster and more consistent.
CNC is cool, and you can do some pretty amazing things e.g. with a 4 or 5 axis machine, but ultimately the value is in repeatably churning out the same part over and over again. The industrial utility is obvious, but there's not a lot of pedagogical utility and to a hobbyist or prototyper it's generally just not worth the overhead.
> and roofing.. well, let's be fair, most people won't be doing that themselves
Do you have any idea how much it costs to get a roofer out to fix your roof? In my area if the job is less than $50k you can't find anyone who is interested. Not doing it yourself is an incredibly expensive luxury. Best know how to do it properly and safely.
Learning to fix a roof is like learning how to build a boat hull upside down: if you can do it properly then you can make lots of money doing it. That said,
" if the job is less than $50k you can't find anyone who is interested"
is difficult to believe. I only use experienced roofers and make sure my and their insurance coverage is good. I've always felt roof repairs were well worth the price.
Experienced, good roofers and builders around here have enough new construction and full roof replacements that scheduling a small repair just isn't feasible. Those are not jobs they want, because they already have too many big jobs--some are fully booked 1+yr out.
A metal lathe is the fundamental machine tool. You learn how to calculate feeds and speeds, plan depth of cut, thread cutting, parting, etc. You learn about surface finish, chatter, cutter shape... Why rob students of this? Should nobody learn basic machinist skills? It's no more or less dangerous than any other shop tool.
The same lack of context is what (I think, if I'm feeling charitable) makes these people think AI is real:
If you spend your entire day in meetings, you might reasonably think that you'd be better off if all your meetings were face-to-face.
If you only touch a computer to write and respond to emails, the email summary parrot might reasonably seem like some omniscient god.
The trouble is I don't feel charitable. These people got to where they are by behaving like narcissists and sociopaths. That's because they are narcissists and sociopaths. It's about controlling other people and hurting them. Full stop.
1. I have the same desktop layout every time, from left to right: slack desktop app, a two column wide emacs window, a 90col wide terminal. I also have two chrome windows--1 which is the same width as the slack window and overlays it, and another the same width as the emacs window which overlays that one. The problem is every single time I wake my laptop from sleep the terminal window has shrunk to fewer columns and I have to drag it back to full width.
2. Sometimes the external monitor support bugs out. I don't know if that's my hub ("pluggable" something or other) or the OS or both.
Then of course there's all the warts of homebrew, and the fact that it's not easy to build some software..
However, the performance of the Apple silicon is nothing short of astonishing. I'm curious about the AMD chips that ship in the new Framework as I look towards an upgrade to my personal laptop, but it's basically between that and a new M4 Max Macbook. Never thought I'd see the day.. will probably wait a year or so before deciding but it's interesting that Apple is even a contender.
reply