ianhowson

I think the proposal suggested an orbit where the solar panels are always in sun and always properly aligned and always clean due to space gophers.

But more seriously, GPU loads are super spiky. Ground-based power grids and generators and batteries have trouble keeping up with them. You can go from 1MW idle to 50MW full power in 10ms. Unbuffered solar cells are right out.

This is the best idea to come out of this whole scheme. Space solar panels are super cheap and efficient? Prove it! Launch them and transmit the energy down.

This is orders of magnitude easier than the original proposal -- and yet still nonsensical.

I had a good laugh.

- You can't build 40MW of solar panels for $2M, even with theoretical maximum efficiency. You can't even build the cabling and regulators at that price.

- You need battery storage -- not as your backup -- but as primary source. It is going to cost more than $2M. Batteries are heavy. They are going to cost a lot to launch. This is not even solved on the ground yet.

- You need a heat transport medium to move heat into your massive radiator. Either you use water or you use air or you use heatpipes (metal). You have to pay for the cost and weight and launch expense. This is probably half the weight of the rack and I haven't bothered to do the math about how you transport heat into a 500 foot solar sail.

- Let's not even talk about how you need to colocate multiple other racks for compute and storage. There aren't any 1TBps orbital link technologies.

- Rad shielding? It doesn't work, but I'll let this slide; it seems like the least problematic part of the proposal.

- 15 year lifetime? GPUs are obsolete after 12 months.

I don't want to be the guy who shoots stuff down just for fun, but this doesn't even pass the sniff test. Maybe you can get 10x cheaper power and cooling in space. Still doesn't work.

1. They're overpriced. A quick Amazon search will show comparable kits for half of the price. You can buy a whole prebuilt ebike for the same price.

2. They're dangerous. FWD ebikes have a tendency to break traction on corners.

3. They tend to destroy the dropouts in your fork.

4. Installation is much more complex than "pop off your bike's front wheel". I can't even find the axle compatibility specs on their website. You know that there are many standards for front axles, right?

Do not recommend.

> Maybe the device was churning on error re-transmissions from the brownout

I think this is more likely. Two different routers impacted. Crappy grounding or induced noise causing high BER on the links.

Irrelevant.

1. Downstream of the mains power supply are DC-DC converters that run the router hardware. Those contain the filters and capacitance you think you're fixing. Nothing in that router actually cares about mains power quality. They absolutely do not care about perfect sinusoids.

2. If you were seeing insufficient power to the router, you would observe crashes and faults -- not slowdowns.

3. Two different routers showed the same behavior, which suggests that the fault lies outside the router+power supply and more to do with something common (e.g. network, laptop).

The dip shows a reduction in voltage, and a larger one than I would like, but without a scale on either time or voltage, it's difficult to guess if it actually matters. I would suspect not, since the device does boot successfully. Again, the voltage doesn't matter, since the router runs off its internal DC-DC supplies, not the external power supply.

I'm happy that the capacitor and new supply has fixed the issue, but I'm unconvinced by the explanation. Check grounding between inverter and laptop.

Well, there are encrypted CNIs like Weave. I've used Calico over ZeroTier to similar effect. The network is 'encrypted' and there isn't much effort required past initial configuration.

But that's not really the issue. You still have a big plaintext network with a bunch of random stuff talking, no mutual auth and no security controls other than segmentation. That's the tricky problem that mTLS and service meshes attempt to solve.

There's now FDA guidance on cybersecurity risks, but in a nutshell, it says "you must consider cybersecurity risks in your risk management plan." Manufacturers were (usually) doing this anyway but the probability and impact have been reassessed since the various lawsuits.

I don't work on pacemakers, specifically, but I'm confident you can no longer say "it has RF remote and nobody can buy the programming interface, therefore it's secure", because that's what went wrong before.

Instead, the risk management plan will look something like:

1. "magnet only with big disclaimers" (because you need the emergency cut-out switch; 'patient died because malfunction' is a worse outcome than 'maybe someone put a big-ass magnet on someone's chest and they got mostly non-life-threatening symptoms')

2. optionally, some form of 'secured' RF interface. Don't presume SSL. This is risk management, remember, so it's probably enough to say "hold a magnet on and then we enable unencrypted unauthenticated comms". Or maybe they do use SSL and just put way bigger micros in there. I don't know.

(edit: someone below who knows more than I do says "use Bluetooth", possibly triggered by magnet taps. Makes sense to me.)

Further, it's not that the magnet 'hurts' the pacemaker or device in any way. The presence of a magnetic field is used as a control signal to change the operating parameters of the device.

This is critically important for a device that may not have any remote control functionality or easy way to reprogram it other than 'cut up the patient'.

(Keep in mind that these things were designed decades ago, and 'RF remote control of pacemaker' probably introduces more problems than it solves, even today.)

I lived in the Gold Coast for a year while my job was in Sydney. I commuted by air once a week. Total commute time and housing expenses decreased.

I generally agree with you. I don't personally find eMTBs to be bothersome, and fully expect to be riding one when I'm not physically able to ride an acoustic MTB. And I definitely agree with you re. bystanders, where regular MTBs are fast enough to be dangerous.

The only point where I disagree is:

> There's nothing about eMTBs allowing people to venture in more difficult trails than with normal MTBs

If you want to go down the hill, you need to climb up the hill. This is a skill and fitness gate, normally; you're not going to tackle a difficult or extended descent if you don't think you can climb out again.

Soquel Demo Forest, one of the very popular Bay Area spots, bans eMTBs. It might be reasonable in this case due to the elevation profile. From the car park you need to climb for 20-40 minutes. You descend. You then have another long climb back to the carpark. It's a worst-case scenario for batteries, and because the area is so popular, it tends to attract people riding beyond their limits already.

I don't personally care, and I don't want to tell people not to enjoy an activity, but I can understand the reasoning in some cases.

Powertrains are much, much better.

- You can shift under load (usually)

- Shifts are a little faster

- Clutches give less chainslap

- Narrow-wide rings hold the chain freakishly well

At least here in NorCal, it's not trail damage that's the issue. It's:

* people riding beyond their ability (too fast, too technical) and getting injured

* batteries overheating and stranding the rider somewhere they can't climb out of, requiring heli rescue and risking wildfires

I also don't like to be overtaken by someone who isn't suffering as much as I am [1] and it makes me feel bad.

[1] Not really, I also ride a road e-bike which strokes my fragile ego.

It won't work with the LG Ultrafine 5k, which requires a Thunderbolt connection to run at full resolution, and behind-the-scenes is a tiled pair of DP1.2 displays.

It ought to work with a DisplayPort 1.4 5k, though there aren't many of those.

I do misunderstand. Please educate me! What makes these better than other headphones or the AirPods Pro, which are looking downright good value right now?

I think we're in vigorous agreement.

I use the AirPods Pro not because they sound the best, but because they sound good enough and the convenience factors make them worthwhile.

I dump on the HomePod because it lacks any convenience factors that are meaningful to me [1] and it doesn't even serve the "speaker that sounds good" purpose [2].

So I really want to know what the value prop for AirPods Max is. They're not convenient or useful for travel because they're too big. They're not "best audio quality", because that's been done at lower price points. Spatial audio and ANC? Already solved, better, by AirPods Pro, at half the price. They're not even usable for critical listening or gaming because of Bluetooth.

So what are they for? Fashion? (Nothing wrong with that, but I'm sure as hell not going to spend $550 for it.)

[1] Siri can't understand me and it false triggers constantly.

[2] It wasn't even like, "hey its good but there are better speakers". They sounded like a cheap plastic box. They were better than my $100 Google Home. They're worse than the $120 soundbar I put on a TV. It's a low bar.

I honestly think that audio quality is consistently poor on Apple stuff and they try to patch over it with EQ boosts and marketing.

AirPods are better than EarPods, but they're still in $25 earphone range.

AirPods Pro are nice, but you can get wired IEMs for <$100 that sound just as good. In that price range you can get Pinnacle P1 ($200) or ER4XR ($250) which dump all over them. I use AirPods Pro daily, not for quality, but for convenience.

HomePod is probably the biggest disappointment I've ever heard. $300 and it sounds like a plastic box, despite "computational audio".

At $550, you're solidly in headphone big leagues. Beyer DT770 or DT990 are close to perfect and they're <$200. Beyond that point you're hitting diminishing returns in audio quality; double price is going to get tiny marginal improvements in quality.

I'm eager to hear them but I can't imagine them outperforming DT990s, despite costing twice as much.

(Yes, none of these options have bluetooth or ANC. Get an ES100 for BT. If noise is a problem, get IEMs or AirPods Pro.)

> Did you miss the part where they

I don't think that's notable, sorry. I would expect that of any modern CPU.

> it’s not common for general purpose laptops and desktops

Well, yeah, because "memory on package" has major disadvantages. You (laptop/desktop manufacturer) are making minor gains in performance and power and need to buy a CPU which doesn't exist. Apple can do it, but they were already doing it for iPhone, and they must do it for iPhone to meet space constraints.

I think unified memory is the right way to go, long term, and that's a meaningful improvement. But as you point out, there is plenty of prior work there.

> they don’t have everything that’s part of the M1 system on a chip

They actually do! The 'CPU' part of an Intel CPU is vanishingly small these days. Most area is taken up with cache, GPU and hardware accelerators, such as... hardware video encode and decode, image processing, security and NN acceleration.

Most high-end Android cellphone SoCs have the same blocks. NVIDIA's SoCs have been shipping the same hardware blocks, with the same unified memory architecture, for at least four years. They all boot Ubuntu and give a desktop-like experience on a modern ARM ISA.

> There’s no other desktop ... at the price point of $699

Literally every modern Intel desktop does this.

Anandtech is comparing M1 vs A14. It's high performance for a cellphone part.

Dual channel DDR3L or DDR4L also has a 128 bit bus. 4200MHz DDR4 is clocked on the high side for most laptops, sure, but it's hardly unusual.

Run the numbers and you get the exact same throughput figure as for M1, which isn't surprising, because we're just taking width * rate = throughput.

So I'll repeat my assertion, downvotes be damned: the memory on the M1 is not special. The packaging and interconnect is interesting. It might reduce latency a little; it probably reduces power consumption a lot. But there's nothing special about it. The computer you're on right now probably has the same memory subsystem with different packaging.

M1 is using standard LPDDRx. It's not "very high performance". It uses a different interconnect -- that's it.

I guarantee that 64GB does not cost anything near EUR300.

You might be thinking of HBM[2] which has a wider I/O path and costs more.

Well... yeah... but they also dropped clock rate from 2.5GHz-ish to 1.7GHz-ish. That could equally well explain the increase in core count at the same TDP. You're gaining about 15% IPC improvement from Kaby->Whiskey [1].

It's an overall improvement, but not as dramatic as "2x cores for 2x perf at the same power"

[1] https://www.anandtech.com/show/14514/examining-intels-ice-la...

They might embarrass today's Threadripper. They don't exist yet, at least not outside an Apple lab. AMD's not sitting on their hands, either.

You've got it backwards. Apple chooses the TDP. Intel provides the CPU to suit. Apple is choosing TDPs which are too small and then providing thermal solutions which only just meet that spec. They could provide better thermals without hurting anything else in the machine and get a higher base clock.

I assume they do this for market segmentation; see 2016 Touch Bar vs. non-Touch-Bar Pro. One fan vs. two.

The TDPs look appropriate for M1 parts. They're too small for Intel. I'm guessing that (a) Apple predicted the M1 transition sooner and (b) Apple designed ahead for Intel's roadmap (perf at reduced TDP) which never eventuated.

So, unfortunately, Apple have shipped a generation of laptops with inadequate cooling.

I hate the "M1 vs Intel MacBook" comparison. Every Intel MacBook back to 2016 has broken thermals. They're all running at maybe half their rated clock speed. 13" MBP is a 4GHz part running at 1.4GHz. 16" MBP is a 4.8GHz part throttled to 2.3GHz. You're comparing M1 vs. a broken design which Apple broke.

Don't congratulate Apple for failing to ship trash.

There's an argument for efficiency on a laptop, no doubt, but that's not what the parent commenter is talking about.

M1 is the highest perf-per-watt CPU today, no question. Ignoring efficiency, there are plenty of faster CPUs both for single-core and multi-core tasks. That's what "my Hackintosh did the build in 5 minutes" is showing.

I'm happy to pay for the electricity if I get faster builds.

My daily driver is a Hackintosh and with the CPUs pegged pulls about 70W from the wall. The two displays add another 90W.

M1 is impressively efficient but there's still a gap for fast, no-compromise workstations.

If my partner leaves, there's more room for bikes!

Or an aluminum frame that can deal with being banged around a little.

Under some constraints [1] [2] if you halve the frequency and double the number of execution units, the overall power consumption drops.

Therefore, you could gain performance or reduce power by stacking layers of silicon.

This is also true for things like power LEDs; within a certain range of their operating curve (current vs. output) you can reduce current by X% and lose less than X% of output. Put down two LEDs, then, and you get more output at the same current.

[1] architectures that scale efficiently to more execution units, like GPUs

[2] you're in a suitable region of the frequency-power curve

Yep, agreed.

If you can admin your own Jira instance and keep the fields and plugins to exactly what you need, it's pretty nice. I keep going back to it for the customizable workflows and fine-grained access control.

Unfortunately, every corporate Jira instance I've ever worked with has been overrun with every possible field possible, an unnecessary and constantly shifting mishmash of plugins and horribly slow access. You're paying the price for all of that cruft that you don't need and which can't be removed because "maybe someone wants it."

As a human, when I have many tabs open, I observe that everything gets really slow. All applications get slow, but especially the browser.

So I put on my engineering hat and pull up Activity Monitor and further observe (a) high memory pressure, (b) high memory consumption attributed to Chrome or Firefox, (c) high levels of swap usage, (d) high levels of disk I/O attributed to kerneltask or nothing, depending on macOS version, which is the swapper task.

I close some tabs. I then observe that the problems go away.

Swap isn't a silver bullet, not even at 3Gbytes/sec. It is slow. I haven't even touched on GPU memory pressure which swaps back to sysram, which puts further pressure on disk swap.

It's slow.