ARM has a high probability of blowing a tire.
They have a complex relationship with their licensees which may try to cause self-sabotage trying to pull more of the money home. See the various licensing fights.
If you don’t want or need x86, what does ARM have to offer-- in the long term-- over RISC-V, which is much less coupled to a single firm’s caprice? We can assume the gap in performance will continue to shrink ovrr time.
My objections:
Instead of writing the code now, you end up having to review and debug it, which is more work IMO.
Not to mention the Xbox Box, and the shipping cintainer full of 'em, the Xbox Box Box
The experience could be somewhat tamed by a lottery process.
Accept a token deposit for a week or two, and then draw from people contending for a given seat, then give them another week to pay the balance. Any unclaimed seats are put up at will call night-of-the-show. Limit the number of deposits taken from any given card to prevent “I’ll claim 30 seats and only buy 1” gaming of the lottery.
There’s probably some more complexity about it (if you want N seats together), but I think that would dramatically cut back on the frustration for “the tickets were only available for 14 seconds and the server was being DDOSed by scalper bots.”
Having to put down a deposit with no guarantee of a ticket also makes “buy All The Seats” scalping theoretically impossible and economically riskier. If there’s 5/1 contention for a ticket, you’d have to find a way to get 3 lottery slots for a better than even chance of getting it. If the deposit was $10, you’re spending $30 for the chance to buy a $50 ticket-- so if you can’t resell the ticket for at least $80, you lose. Under current policies, if you can sell that $50 ticket for $51, you’re ahead.
Or that there’s a huge amount of legit demand for mature node chips and it makes sense to own the supply for it.
The 5000 microcontrollers you inyeract with each day, by and large, do not need 5nm processes.
We saw a few years ago how relatively cheap, commodity-grade, low-complexity chips suddenly become vital when you can’t get them and they have unfinished cars piling up at the assembly plant.
I suspect Intel has a broader product range than AMD to justify the headcount, but I’m not sure where the extra resources should go.
Their networking chipsets were gold-standard in the 100M and Gigabit era, but their 2.5G stuff is spotty to the point Realtek is considered legit.
They’ve pulled back from flash, SSDs and Optane.
There must be some other rich product lines that they do and AMD doesn’t
It can also throw things against the wall with no concern for fitness-to=purpose. See “None pizza, left beef”.
Also on modern firebreathers.
I like runit better than systemd, the packages are current, and it has most of what I want in the main repos.
I also found the documentation excellent in thst it’s a cohesive list of real-world topics rather than a 500-km-deep wiki or forum archive.
I should try a modern Slackware one day. I loved it back before I had broadband and just ordered a burned CD for each new release, but I should try following -current and the Slackbuilds stuff.
We get, for some reason, a huge number of window replacement contractors coming door-to-door. Because I really want to be high-pressure sold $10k worth of low quality glass from the people who are running big enough marhins to put a full page colour ad in the local newspaper every day to go with their 6 hours a day of local TV spots.
I actually said to one “We’re a Linux household. Not interested in Windows” and slammed the door on them.
I now realize cocking a rifle would have made the effect even better.
They’ve got a guy at work whose job title is basically AI Evangelist. This is terrifying in that it’s a financial tech firm handling twelve figures a year of business-- the last place where people will put up with “plausible bullshit” in their products.
I grudgingly installed the Copilot plugin, but I’m not sure what it can do for me better than a snippet library.
I asked it to generate a test suite for a function, as a rudimentary exercise, so it was able to identify “yes, there are n return values, so write n test cases” and “You’re going to actually have to CALL the function under test”, but was unable to figure out how to build the object being fed in to trigger any of those cases; to do so would require grokking much of the code base. I didn’t need to burn half a barrel of oil for that.
I’d be hesitant to trust it with “summarize this obtuse spec document” when half the time said documents are self-contradictory or downright wrong. Again, plausible bullshit isn’t suitable.
Maybe the problem is that I’m too close to the specific problem. AI tooling might be better for open-ended or free-association “why not try glue on pizza” type discussions, but when you already know “send exactly 4-7-Q-unicorn emoji in this field or the transaction is converted from USD to KPW” having to coax the machine to come to that conclusion 100% of the time is harder than just doing it yourself.
I can see the marketing and sales people love it, maybe customer service too, click one button and take one coherent “here’s why it’s broken” sentence and turn it into 500 words of flowery says-nothing prose, but I demand better from my machine overlords.
Tell me when Stable Diffusion figures out that “Carrying battleaxe” doesn’t mean “katana randomly jutting out from forearms”, maybe at that point AI will be good enough for code.
Source?
I’m more willing to forgive not getting Baizhu for the promise of unlimited cheap energy…
Is United Express actually United? I thought those tended to be a regional carrier using the name under license.
I’d expect the labour friction to be still worse; I was peripherally involved with such a firm 20 years ago and know they had terrible problems with staff retention, mostly because they wouldn’t pay enough to retain people after they got fed up with the free-standby-flight privileges.
I think the appeal is that you probably don’t need a huge CPU for a lot of workloads-- just something to run an OS, handle talking to the outside world, and configure the GPU/NPU complexes.
I could imagine a something like a Quadro card that had a small RISC-V core built in as a freestanding device, no motherboard needed. Even if the CPU ran like a Core 2 Duo, that would be sufficient for purpose, but it will be a lot easier to license an appropriate RISC-V core than an x86 one.
I wonder if it might be an unreproducible moment in history.
I suspect the Cambrian explosion of X11 window managers came from two things:
Propriatery and former-propriatery systems with unique look and feel (see, for example, Open Look/olvwm) There was also a tendency to copy any style you could (WindowMaker copied NeXTStep, IceWM mocked OS/2, and when those cute QNX demo discs came out, within days there were lookalike themes). It feels like the last major outside inspirations, MacOS and Win1,1 are converging on almost intrrchangeable insipidness.
The 1990s/2000s customization era. Machines were finally powerful enough to do mildly nifty things, but still attainable by hobbyists gluing together pixmaps and this bred stuff like Enlightnment E16 or Afterstep
Do these forces still exist in 2024? It seems like Unix Porn today is a bunch of neokvetch windows without even a titlebar to provide a personal statement.
I think I’d be a lot more excited about Wayland if I felt like I can get a compositor that matches my tastes.
I want to iconify things to the desktop, not relying on a taskbar-alike. Nothing seems to offer that. Hell, the taskbar is often a third party program.
I want to double-click to shade. Labwc just added this, a feature that X11 window managers have been offering since the 90s.
I want an aesthetic that’s got real depth and skeumorphism, rather that flat and featureless. Maybe something offers that, but there are plenty of X11 choices that have beveled buttons out of the box.
The charm of Unix systems used to be flexibility, buy Wayland seems to be an extinction-level event for traditional window management. Nothing fills the gap of FVWM or WindowMaker. But gosh, I can get 92 flavours of tiling compositor and windows that ripple when dragged.
I had a similar positive experience with Gamescope, which tamed a game that freaked out every time I moved the moude onto the other monitor.
Maybe Wayland’s healthy place is as a secondary window system you launch inside your normal X11 session.
So next they’ll use AI-generated infantry?
I liked ASrock when they were in the ECS tier of quirky and weird. Got a Socket 939 board with the ULi M1695 chipset that was really nifty.
Then I had an awful experience with an AM3 board that claimed to run a FX-8350, until they edited their support list.
I grudgingly chose them for AM5 because it was $50 cheaper for the featured I wanted, and it’s been okay, aside from me breaking the x16 slot clip due to hamfistedly removing a shipping-container sized GPU.
There are some canned choices like “50 newest tracks”.
Fujifilm successfully repositioned towards other chemistry. I know there’s that Eastman spinoff but why wasn’t it as successful?