archi42

Regarding firmware patching, look at InternalBlue [1]. I guess they did it for one IC, but obviously can't do it at will for just any IC and instead hand wave "it can be done".

Overall I don't think it's impossible: Someone stated it would be unidirectional, and the bitrate was not given (read: could even be minutes per bit for all we know).

But yeah, I'm sceptical, too. A gateway would have made more sense to me, because it does not require firmware-level cooperation with chip set manufacturers and you can reuse existing space networks. Especially since they're not aiming to add the feature into future ICs (which could have driven sales across the whole industry), but backport it instead, competing with the manufacturers and voiding existing FCC et al. certification.

[1] https://github.com/seemoo-lab/internalblue

Previous job: "We need some more real & complex test code for that compiler, but it doesn't support all language features" - "okay, Linux is too huge, we already did a few real life libs like compression, encryption and sqlite... I'll port ioquake3 to it". I had to resort to GCC for some parts that could not be ported without disproportionate effort (graphics code iirc?); but once it built, it ran without error and I could play a few matches. At least that's what nostalgia tells me ;-)

Around here people don't buy it because some people successfully campaigned to get the stuff banned. That's all there is to it.

They're pretending that it's good for the kids and prey on the uninformed or those who can't afford better alternatives. That's unethical to me.

Thanks, I wanted to point out the exact same glaring omission. Unbelievable! :')

Uh, yeah, this is all horrible. Even when founding a company makes sense, it's sometimes difficult to justify - e.g. when we were buying a house I was looking at bigger buildings with multiple units, so having a estate holding company would have made sense: We're already taxed over 30% and rents could have pushed us into the highest bracket, while most of the money (rents) in the holding would have gone into financing and maintaining the building anyway.

Don't get me started on founding a company and then getting necessary loans (all while the loan market changes and suitable houses might vanish from the market). Some proper investor is probably now renting out a multi unit house we would have bought, making nice profits while we would have aimed at boosting our pension in a few decades.

But well: That's the way it's meant to be, else the rules would be different, or would they?

(We got an old SFH instead, and slashed energy consumption by ~80-90%. That's not even accounting for PV. It's nice to profit from that now and later in life, but I still believe that we, as a society, would have benefited more from improving a MFH and sharing the lower energy consumption with some tennants)

From the article:

  [...] the town is considering installing a steel beam in front of the bridge to force trucks that are too tall to a full stop.

That sounds a lot like the operating principle of the famous Can Opener :)

(btw, our cats love those videos on 11foot8.com and can watch them for hours)

User Hakkin made an interesting observation: The "consumers" are likely sock puppets controlled by the attacker.

I did not independently verify the claim, but it seems reasonable to me.

Post with Hakkin's reply: https://news.ycombinator.com/item?id=39869703

(But yeah, I've seen some entitled idiots trying to push their demands on FOSS maintainers and being irritated by the harsh reaction...)

Nope. That's all I'm going to spoiler;-)

@Potro: If you liked the movie, read the books. I don't read a lot anymore, but during sick leave I started with the first book. Didn't stop until I finished the main story, including the sequels by Frank Herbert's son about a month later. That's like... uh... nine books?

Absolutely. 116851 long tons deadweight, which the Google info box says includes everything (busy now, so can't check in detail). That's 120,000,000 kg, so a little bit less than I guessed. Now looking at 10km/h or 6.2mi/h or 2.78m/s, that's enough to come up with a big E for E = 1/2 * m * v². The momentum p = m * v is also rather unpleasant. If the momentum was suddenly transferred into my 2,000 kg car (e.g. if the ship hit my car), I could do my work trip in about 1s, though that's only a few times times faster than the ISS (which travels at 7660 m/s).

[I'm very busy, so hopefully I didn't forget/add a few 0s by accident. Anyone feel free to correct me if I did.]

You could try to avoid collisions with barriers, or make the bridge fail differently (only a few spans), or make it tolerant to a single support failing. But you need a pretty beefy steel beam to survive a direct hit by... Uh... 200,000t(??) traveling even at a slow jogging pace. (And then it transfers the energy into the rest of the bridge, which will very much not like it).

I guess we can be happy failure like this is rare, and that the bridge was not busy: reports indicate 13 cars, and about 7 persons still missing. This could be so much worse.

The best way to think about measures is after we know the full chain of events that have lead to this.

Take a look at "evcc". Careful: It's MIT licensed open source, but a a set of interfaces (to inverters, wall boxes, car APIs or smart meters) need to be unlocked with a sponsorship token; or patching the license check. -> https://evcc.io/en/

It's a stand alone server with a simple Web UI and takes care of charging your car with solar (or grid, if there is not enough solar). IIRC it offers three modes: 1. charge solar only, 2. charge with at least 6A, boost if there is more solar, 3. charge independent of solar. You can also configure charging targets, e.g. stop at 80% SOC.

The PR does not mention plaster work; I suppose the company is focused on the... errr... "shell construction"? Once plaster is added, I'd wager it's much the same as for any other house.

I found this video: https://www.youtube.com/watch?v=dyAoAMXAVEk

They seem to have issues with rain (no real surprise), but the process seems relatively straight forward.

I wonder if, for smaller scales, a cable-driven, flying delta could be used? For garage-sized buildings or smaller houses this could work (though I believe pre-fab to be cheaper).

I was also surprised not to see a flying probe system - I would expect this to be viable with modern 3D printer motion & control systems, but obviously this is highly non-trivial and has lots of mean details in the mechanical, electronics and software domains to solve.

I did not think of a die-bond machine (I suppose it bonds a wire to each pad instead of you doing it by hand?), but of course that also makes sense. And at least the motion system is much simpler.

A first step/experiment could be to automate creation of the gnd net. For that you only need a single tool head, meaning you can repurpose mostly any 3D printer motion system; for small increments, this could (later) happen during the die-bond process or become a precursor to a flying probe tool head. Of course I can not judge if that's a worthy investment of your time, or if you would enjoy building something like this ;)

Anyway, the effort, skill and dexterity are amazing! Spending 3 weeks soldering 1917 tiny leads seems to be just the icing on the cake :)

My UPS is my hybrid inverter. It offers at least 1kWh of usable capacity (after losses), more of the sun is up or the battery is charged above 20%.

Of course there is no native signaling on grid loss, but in practice that's not an issue; for planned maintenance I can force charge the battery and keep the servers running for a day or more, and actual power loss incidents are very brief in my corner of the woods.

While liquidity is a reasonable argument, I only know this from "rich people doing things/investments"; e.g. not selling shares in your company when building a new house, or a company building a warehouse.

I am not sure it actually applies to the cost of a phone. Chances are, if those matter, you'd be better off with a cheaper phone to begin with. (Substitute phone with other luxury articles).

It's a matter of the ABI, and most importantly the calling conventions. Plus, depending on the architecture, you'll not be able to use some of the builtin push/pop/call/return-like instructions anymore. This can be a minor caveat or a major performance penalty, depending on your program and target architecture. And I'm not even talking about compiler-builtins or libc stuff that will wreak havoc unless you rewrite all of that for your new ABI.

Also: The program still crashes at "random" spots instead of behaving unpredictable. But for that you could add some assembly at the beginning of each function, asserting that the sp is in the pre-defined stack range. If your hardware supports it, you could also mark the "overflow area" behind the stack as read-only and let it crash that way.

Due to a past job I'm biased to static program analysis for this task, something the post author completly neglects (doesn't know about?). At least if sound boundaries matter (e.g. safety critical stuff).

Absolutely. Try retrofitting this into a multi unit home and "just run a digital data cable up there". With some bad luck, this can easily cost as much as the whole unit.

I don't like getting everything on WiFi, too. OTOH I have ~30 devices on an VLAN isolated IoT WiFi network. This easily saved us a 4 digit expense and a lot of time(!) as opposed to installing new additional wire (redoing all the old ones was bad enough). Plus, I can do dumb stuff like control my office's window blinds individually, which is nice when only one of them has the sun blinding me while using the PC (temporary desk location due to ongoing renovations).

I've rebuild part of my Delta the last few evenings and today it had it first print since ~1.5y: The humble pressure advance calibration cube for Klipper; but with PETG [which I couldn't print before, but was all I had left] on a new bed, with new glass surface, on self-designed & resin printed mounts, with higher speed than I previously used (100mm/s)... - with other words, I did expect it to just fail, but it worked reasonably well! Dimensional calibration is next, and I'm waiting for the accelerometer to add resonance compensation.

This was a great decision! Created a USB stick for the RPi4 with Mainsail OS. No display necessary, just connected Ethernet (WiFi did not connect, didn't debug though). The printer setup was really easy, though I had to adjust the template configuration for my hardware modifications (mostly tmc2130). Most actual effort was in reworking parts of the hardware (remove case, solder new build plate, design and test the mounts, fix bad soldering that only caused half the heat plate to heat up [good bless thermal cameras!]). Overall, the UX has increased tremendously since I first heard about Klipper (IIRC Kevin just added delta support back then). I think the Mainsail UI has a huge impact on UX. Especially manual calibration of the delta kinematics with the Mainsail web UI is a delight. I scrapped getting piezo probes or stuff like that (for now).

In the ethics department I enjoyed "The Cambridge Handbook of Information and Computer Ethics" (978-0521888981). We read part of it for a ethics lecture during my M.Sc., and like most ethics/philosophy piece I found a lot of things to think about.

IIRC chapters are contributed by different authors, so it's not a single point of view (even though of course there is some filtering by the editor).

Since it's from 2010 it is light on more recent developments like AI, however (unlike software) it's rare for ethic ideas to become outdated. So I would not dismiss it due to age. Also it's 5$ used on Amazon, so depending on your situation it's not a huge deal to get it and skim it - IMHO this kind of work doesn't have to be read front to back, instead it's great to get back to it when you feel like and think about a few pages and explore new ideas, re-evaluate your own concepts, etc.

Description from the publisher: "The Cambridge Handbook of Information and Computer Ethics, first published in 2010, provides an ambitious and authoritative introduction to the field, with discussions of a range of topics including privacy, ownership, freedom of speech, responsibility, technological determinism, the digital divide, and online pornography."

Editorial reviews stolen from Amazon: "...This five-part work examines difficulties in the field of information ethics and offers practical applications and criticisms... Recommended..." --B. G. Turner, Faulkner University, CHOICE

"...This is a rich and fascinating book, bringing to interpretative debates much that has been hitherto unknown. The chapters are long and complex, and the argument is multidimensional and far-reaching." --George Lăzăroiu, PhD, Institute of Interdisciplinary Studies in Humanities and Social Sciences, New York, Contemporary Readings in Law and Social Justice

Thanks :)

And oh well, ours would be pretty good then, but sadly I just got new data from her: 30% or 260km down to 160km, and the lowest temps for the last days were only 0 to -5 degC. Still good enough for her daily total of 60km. One big factor: The dealership didn't want to add the heat pump, and we didn't want to pay that extra in full for a car we would not keep anyway (plus we were lucky and could get the car from their stock).

4" are 101.6mm, though overall they're more like 120mm. The situation for me is quite similar (initially placed on/behind the PC desk) However my build is rather wide (20cm) due to the double MDF of the walls. Not sure they're necessary, though (would shave off 38mm).

The issue are usually the woofers, since these need displacement. 4" is already quite small. Maybe look for designs with passive radiators on the front and the woofers on the side? (I'm not deep in enough to make concrete recommendations).

This! I recently had to rig something up in an emergency: Did a `pacman -S rpi-imager` since that's what Google says the kids use these days (it's new to me!), selected the RPi4, selected the headless Armbian, gave it my ssh public key, entered WiFi credentials and hit go. Then realized I could not find that lovely 3A USB power supply, and couldn't get one since it was saturday night.

Next I messed around with an old OrangePi just to realize they never updated the Armbian (even though I got the device tree into the Kernel!). Downloaded the ancient image from their Google Drive, got BalenaEtcher running again (its not in my repos) and then figured the SBC must be fried anyway.

Luckily I remembered that I still had a RPi1 that was collecting dust and should work with any old USB charger. Checked `rpi-imager` and indeed, it supported that old thing! Created the SD card (config was still saved), grabbed an old USB WiFi adapter and powered the thing up. And it just worked :)

The emergency was: My 20m USB extension cord with the zigbee coordinator failed after 2y, and I could not turn on a few of our lights. Using `usbip` I now forward it to the automation server from that RPi1 (I initially didn't think about the RPi1 since I wanted to run zigbee2mqtt on the device, but eventually found out about usbip). Bonus 1: I still have the spare RPi4 that I wanted to attach to the 3D printer because I could re-use the old one :) Bonus 2: I now know that rpi-imager has a ready-made Klipper image and I don't need to set it all up by hand.

Next time I consider getting another SBC, it will be an RPi again.

1. Fullrange speakers are great, don't underestimate them! The ones I looked at need some help at the low end, but you get away with a much lower crossover frequency compared to a tweeter. Since they can have a relatively big diameter (compared to tweeters) they have a stronger "beam"/focus; annoying when moving or having multiple listeners (off-axis sound is horrible), but for solo listening the stage/imaging gains a lot from not being disturbed by those pesky reflections.

2. Don't add a crossover if you don't have to. They mess with the phase at the crossover point and are hard to get right anyway. Just do it all in a DSP and have an amp for each speaker. You can actually combine room correction (often via parametric equalizer) and the crossover itself into one big FIR, if your DSP is big enough. You can also play with crossover frequencies, slope, sacrificing headroom for a deeper frequency response etc. - in the analog domain, the costs for the passive components quickly add up. That's why MiniDSP + array of amps can have a huge impact.

3. While I'd go with MDF any day, most designs will sound terrible without padding (or too much!). For your 3D printed design I'd expect proper padding to be the lower hanging fruit.

Anecdata: My 4" full range speakers are fully active stereo speakers: 4" FR + 4" woofer per cabinet (based on "Double Four", but with a different FR), 4 amps and an ADAU145x DSP for crossover & room correction (Beocreate 4). They play down to IIRC 40Hz (-3dB, or was it 0dB?) with ultra precise voice/instrument locations. (Not all that glitters is gold: IIRC they're limited to 92 or 95dB and I still need to redo the DRC since we moved).

Edit: For prototyping, there is the free Windows "Equalizer APO" I used to output the four channels using the onboard sound card and fed them into an older 5.1 receiver (with 5 integrated amps) connected to the speakers. That way I could rapidly swap the FIR filters and compare them to insanely large filters. This is nice for experimenting, and with a decent sound card can eliminate the need for a DSP when used to build dedicated PC speakers.

May I ask for more precise anecdata: What EV is that? And which EV are worse for the colds snaps?

Around here it's -5 to -7, so much less severe and if I understood my wife correctly, the Ora has a ~20% drop in range. Still good enough for our use case, but when I'll replace my PHEV with a proper BEV this might be more relevant.

Living in Europe the GDPR applies, and with Bosch there is a better chance that they stick to it compared to other brands. Of course I don't know for sure.

The app transferred 1.56MB of data in 3.5 months. While the could include lots of data that I don't want them to have, at least it's not a massive amount indicating non stop tracking.

Ultimately you're right, and I would love to have the two appliances fully local (the other smart things don't even have Internet access, thanks to awesome projects such as Tasmota or Valetudo), including fancy program selection.

Our washing machine and dryer are networked (Bosch). Current benefits are secondary reasons and I could live without them:

- Notifications when done: Nice, but a beeper would work, too.

- Select program via App, based on what I put into the machine: Occasionally useful when I'm not sure. Also, I can put in my wife's cloths and let her pick the correct program (I did that wrong once and we both don't want that to happen anymore).

- The dryer can load the correct settings based on the washing machine's last cycle. Very convenient.

What's planned (primary reason): Our photovoltaic is also networked, so the plan is to eventually start the cycle when we get energy for essentially free. This saves us money and allows us to use more clean energy instead of the grid mix. I guess the non-smart variant also supports this by setting the "finished in x hours".

What I don't like: Our smart home is fully offline, except for these two appliances. I thought there was a local API, but last time I checked I could not find the info anymore. Also, they sometimes lose connectivity to their cloud (not sure why, maybe hick-ups after that archaic 24h disconnect) and need to be restarted to reconnect.

Edit: Ah, mind we did first replace the old, power-wasting dryer with the new one (amortized after ~2y). Later when the old washing machine started failing, we got this one (which is also better suited for our needs, regarding size; and generally uses less water IIRC). I wouldn't have replaced good units for the sake of smartness. Obviously we could have saved some money by getting more basic appliances, but we had good timing on offers.

Sometimes the indicator LED acts as a voltage reference for some circuit. In that case, destroying it is a bad idea.

Also, I think most hotels would object ;-)

Dennis Giese - the guy who did & published many of the actual vacuum roots - actually did a talk at 37c3, just yesterday! Together with Braelynn he published rooting methods for Ecovacs. Expect easy & comfortable root using the UART debug connector.

Valetudo support seems to be coming, however "it's done when it's done"; the necessary work is more dramatic due to the different protocol Ecovacs uses for cloud communication.

Until then: I would make sure not to update the device firmware. Just be prepared to wait and keep an eye on the release notes and/or Valetudo webpage (unlike with other projects, you can safely assume it to be up to date).

I think I've seen those pumps here in Germany as well, and I'm pretty sure you can get AdBlue in canisters (I have one in the basement that I forgot putting back into a rental).

Yes, this (plus that the whole thing is more convenient than filling up the washer fluid) makes the whole "we don't want the customer to fill it up" even more absurd.