MayeulC

I agree with most of what you said, but just want to point out that "riots" is much dramatized and overblown by the media. for what was mostly peaceful demonstrations, with a few nefarious actors (thieves, vandals, etc) using the chaos as cover, and in practice not much happened.

Now, in my opinon, the major issue with opening a business (regardless of the country) is that it is costly to do so, even if there are tax and state incentives. It's much easier to create a startup when seed funding is available.

I mean, the law is stupid, but as one of the commenters there says, there is a 2% revenue (profit?) cap, and I don't think that VideoLAN makes a lot of revenue.

Uh, I didn't see that coming. 8Gbps is deployed on most of the territory, that's potentially 5.5 cents/s, or about 200€/hour. Probably 1 or 2 orders of magnitude too high to be painless.

Given that the content providers (CDNs) are targeted, that would incentivize p2p. P2P networks are already employed by Microsoft to distribute updates, for instance.

What about CDNs? The article mention Valve, but their traffic comes from Akamai.

In any case, I would be surprised if this law ends up being adopted.

I lost the habit of using them when they got removed from Firefox (I used to be a heavy user), but I am coming back to them.

Here, I would tend to make a tab group per activity (so all products contained inside), so it is not too useful, but at least that allows me to collapse children tabs. It is also nice to be able to close the group and reopen it later.

Overall, I am fairly satisfied by TreeStyleTabs, but the extension must have a memory leak as it tends to use gigabytes of RAM.

Tab groups are available on Firefox mobile, but they have a major deficiency: new tabs are not opened in the same group, so I cannot have, say, a HN group with comments and websites automatically added to it, so the overhead of adding them manually is too high.

I use them as well, though a bit less, especially on a tiling WM (well, I could also tab my windows...). However, I'll often put a separate browser window on a virtual workspace when working on a dedicated topic.

Ah, sorry, my Android keyboard isn't that reliable, I tend to make a lot of typos when I type on my phone.

This is something I have been wondering for a while: do custom laser printer firmware projects exist? I'm mostly interested in that to remove toner manufacturer checks, and fix a few annoyances in CUPS and the printers' postscript drivers.

I imagine that is shouldn't be too complicated to support more HW in the agentic age.

You can also have infernal multiplexers to route the antenna signal to another pin, like the esp32-S3 has: https://docs.espressif.com/projects/esp-idf/en/latest/esp32s...

The output trace needs to be impedance-controlled, but it could be a regular GPIO, or an external antenna connector/footprint.

Thanks for the links, that looks useful.

I am a tab hoarder. It's usually because a task requires me to open a dozen tab, and I may have multiple tasks, including ones I intend to come back to later, if I ever have the time.

And no, I do not really have more cognitive ability. Or at least certainly not to the point where I can remember each tab.

I wish I had another way to visualize them. A mind map would be ideal, this is also why I like tab groups and tree-style tabs.

Let's pick an example: I am looking for a product (could be a car, screen, mouse, etc). I explore the product space, perform a wide search first to identify candidates. ~10 tabs. Then I make a short list. 5-15 tabs. For each, I will look at datasheets, reviews, manuals, search for common issues, availability of spare parts, etc. Dozens of tabs per product, but I will close most of these, and leave the most salient ones open. Prune the whole product from the tree if I can eliminate it. And so on (availability at retailers, second hand stores, etc), until the task is done.

Obviously the depth will vary from one task to another, but this is how I usually work on any topic, including literature review. So I only have "root nodes" with the next 20-ish tabs related to it. I don't mind archiving and restoring a tree later, but "out of sight, out of mind", and having them visible reminds me of the tasks in flight.

There are probably more effective ways to work, including spreadsheets (I use them when there is a lot of data to track), but browsers aren't bad at handling that many tabs, usually. Though I also tend to leave a huge quantity of tabs open on my phone: half-read articles that I intend to come back to. Wallabag may help there?

I would conclude by saying it's probably (bad?) habits, a locally optimum workflow, I should expend some energy to try a better approach.

Anyway, your two links are now sitting in my open tabs, hopefully I will come back to them at some point :)

I wouldn't call that laughable, that's 50% higher than the max altitude the F.35 can retain control at. It's number 13 on https://www.militaryfactory.com/aircraft/aircraft-ranked-by-..., with the likes of the space shuttle, the X15, sr71 (30 ft higher), the space shuttle and experimental planes: few production "planes" can fly that high.

Admittedly, it isn't that useful anymore. Missiles can reach that altitude fairly easily, so it does not provide increased survivability anymore, and spy satellites, though less mobile, are more useful for day-to-day operations. But it is still a fairly unique capability.

I'm sure that wasn't the focus of the design, but I can also imagine that its balsa wood structure provides a lower radar cross section as well. (Edit: I made a mistake: looks like it's made of aluminum, not wood. Probably had another plane in mind)

An interesting anecdote was this interception of a U-2 over France in 1967 by a rocket-assisted Mirage. The pilot photographed the U.2 with a hand-operated civilian camera, leading to a diplomatic incident: https://theaviationgeekclub.com/that-time-a-rocket-booster-e...

The Steam Controller uses its left trackpad in exactly this way on the desktop, it's very satisfying to use with haptic feedback and a bit of inertia.

I am not sure where you got this dichotomy, but the appropriation process in the USA certainly has its own corruption issues, strong industrial lobbies are IMO a form of legal corruption (including campaign donations), and the current administration is in a class of its own wrt corruption and wasteful spending.

The USA certainly have the "best" (according to some metrics) economy in the world, but the reasons are complex, diverse, and mainly historical.

I will concede that some areas see fewer regulations, leading to a leaner permitting process (though it has consequences on road safety, environment pollution and gun safety, for instance).

However, it is hard to predict the future, and I am unsure if your last point will stand the test of time. For instance, if/when the AI bubble pop, the USA job market will probably suffer more than others. It is tempting to always see greener grass elsewhere, but Europe isn't in such a bad position compared to the USA.

CoT = Chain of Thought (I had to look it up)

Thanks for the insights. If you edit your post to include the definition, I'll delete mine.

It does a passable job of presenting the ideas be mind miniature inertial measurement units we find on our phones, but the details are completely wrong, so do not read too much into them.

For instance, decreasing distance increases capacitance (thankfully, imagine the capacitive parasitics Voyager 2 would induce into all of our electronics otherwise... I'm taking a ridiculous example, but the claim is ridiculous too). On the very same diagram, it shows two capacitor values changing, but those would be connected in parallel, so you would get a constant value at the end: the structure as presented is wrong, though it has the correct idea: inertia in a mass-spring system, readout (usually) via capacitance change.

So yeah it works, but if you don't know these devices already, you can't really tell what is accurate. The Wikipedia article is more accurate.

It takes time and effort to find vulnerabilities. GPUs are still a moving target, with frequent new models, and many different chips, with multiple companies producing different PCBs. Game consoles are a more static target by contrast.

You would need a juicy target, easy to procure for cheap for the effort to make sense.

It's probably doable, though I suspect that more recent models are probably better protected against that sort of attack. These are performance-oriented chips though, not security-oriented, so they probably make a few compromises. They may also use chiplets with a more secure "secure element", I have no idea (though you still need to communicate with it).

I don't think I've seen anyone attempt it, so it may be that they haven't specifically hardened their designs against these attacks.

Then comes the question of practicality: would you solder a "modchip" to your GPU? It may even be hard to reach the correct spot without interfering with cooling.

I would bet that this is doable with the same setup as the RP2350 hack, but that requires drilling into the specific chip that you are targeting, so great, but not very useful: there probably aren't many secrets to extract, you could decrypt the BIOS (I thought these were just signed with asymmetric keys?) and try to find an exploit in the code there, I guess. But this specific technique is not useful for putting your own firmware then using the chip.

Anyway, it's probably doable to glitch the power supply of some mid-to-late 2010s nvidia GPUs to have it accept custom BIOSes, though graphics card are specifically designed to keep the power supply stable, so you would have to desolder a lot of capacitors (and reconnect them after startup is finished?).

Except it's not really gyroscopes, but gyrometers. The literature is very sloppy with this, but gyroscopes give you an angle, while gyrometers give you an angular velocity.

Using a gyroscope, you can look at it an hour after starting it, and understand how your orientation changed. Using a gyrometers, you need to continuously integrate the result (including the measurement error) to get your orientation: it's far less precise and tends to drift. Both laser and mems gyrometers suffer from this.

Besides this, the video is full of slop and likely AI-narrated. Someone mentioned the bacteria bit, I can also mention the weird-looking PCB image, where traces do not connect correctly to the chip.

They allowed people to run their own homebrew, and allowed more third party apps like Kodi (formerly known as XBMC -- XBox Media Center, that grew out of XBMP) on their store.

Homebrew software had been a driving force behind consoles being hacked in the past. The console is obviously more secure, but they also removed some incentives.

This blogpost makes the point that the advent of the Raspberry Pi microcontroller made it easier to interface with a variety of hardware.

And it is true, but the capability was already quite cheap in 2017 and before: fx2 chips can do quite a lot, for instance.

I hope Bunnie's Baochip BIO approach will take off in microcontrollers, making that sort of feature more ubiquitous. You can do a lot of things with an ESP32 for instance, but the RP2040 still has an edge for I/O.

Also missed the opportunity to mention that the RP2350 itself had its protection defeated by fault injection.

Lastly, side channels and fault injection attacks are highly viable for mid-2010 and earlier hardware, but chip designs are hardening against them, for instance forcing a reboot when a glitch is detected. This makes current and next-generation chips harder to attack (see the laser fault injection on RP2350, though this is also being hardened against). It's probably a good thing to some extent (make it harder to duplicate my credit card), but I mourn the loss of control over my own HW.

Incidentally, tinkering with the first gen Xbox (which is the security equivalent of a swiss cheese) taught me a lot about software, networking and electronics when I was younger. I hope that future generations will still get to tinker like that.

Were the comments accessible in the bit history? If so, I wonder if the agent accessed it.

Yes, putting mesh radios on drones. As a sibling commenter said, Russia already does that: they do not really have access to starlink anymore (they still use it on drones sometimes, but there is a registration scheme in Ukraine that really curbed their use).

Both sides have used drone-mounted signal repeaters for FPV drones. FPVs use cheap radios, and signal extending drones can be reusable and carry a more expensive and jamming-resistant radio.

Since it is rarely a single drone attacking, but more often an entire swarm (to saturate air defences), having a mesh radio between drones can be useful, reduce the requirement that every drone must be equipped for long range communication, and enhance the connectivity back home.

Of course, both sides are mostly supplied by the same Chinese factories when it comes to radios, so China also gets to develop their drone warfare know-how. I assume that Ukraine also uses mesh radios, but probably to a lower extent than Russia does, as they can rely on Starlink.

To address your last point, I would also assume that it is possible to air drop signal repeaters from the swarm, especially for one-way trips, they do not need to last more than a few hours. Range will be degraded on the ground though.

Obviously, I would assume that most strike drones are fully autonomous, but having a connection back home means it is easier to adjust plans (air defense location, for instance) and provide intelligence (if target was hit, etc).

I am also wondering if the recently discovered SDR APIs in Wi-Fi chipsets will be leveraged for warfare. If I had to bet, I'd say: yes, pretty soon. These are a good fit for mesh radio, jam-resistant frequency hopping and wideband transmissions.

You can also prop up the "advanced search" in TB (ctrl-shift-f).

The timing is interesting, given that it comes days after Musk said Ukraine had gone "too far" striking Yandex data centers. Putting other considerations aside, these strikes could be interpreted as direct retaliation against strikes on Ukrainian data centers days before.

That said, depending on a single piece of tech is not a good idea (as evidenced by the Russians when they lost starlink access). Russians have resorted to long-range mesh radios, Ukrainians should probably integrate that as well (though it does increase BOM costs).

The other dependency is Chinese suppliers, who supply both sides but may prefer to favor Russia. It is a huge liability for Ukraine as it supplies 80% of the parts that perform 70% of battlefield duties, and could stop anytime (besides known design leaks to the adversary, etc). Russia is even relocating whole production lines to evade sanctions.

Ukraine has been fighting with one arm behind its back, this is the latest development in the series of seemingly-arbitrary limitations imposed on it.

It's fine if temperatures get lower during the night, but 25°C at night and 35°C during the day leaves little for cooling the infrastructure (roads, buildings), and constant warm temperatures are unbearable if you cannot even get in the shade when you want to cool off. Then, 35°C is also very hard to manage with high humidity.

I did really like PA, and still play it more often than SC/SC2/TA, but the UI was completely broken under Linux for a while (after a few years, it started running with software rendering, now the community more or less fixed it). To their credit though, they did have a Linux build.

This is supremely exciting, thank you for working on this. I also immediately thought of Iroh for networking (nit: though I do think you could include the port in the copy-pastable player id).

Using Lua files for AI does not seem like a great idea if you have control over the game source, though you can likely run run one lua thread per AI player.

I wonder how much your project is an attempt at a faithful reverse-engineered version vs an improved one, though given that you improved on networking, I would guess the latter? There are many features I would love to see added in, such as multi-monitor (multi-window & multi view) support. My concern is: as you diverge more and more by adding features, you risk making it harder to compare to the original for accuracy.

Anyway, as I failed to install FAF in my Flatpaked Steam install a few years ago [1], I would really appreciate a native Linux build for this (though I guess I could try the .exe under Steam/Proton).

I also played a bit of Supreme Commander on my Steam Deck, and while it was playable, I would appreciate better controller integration (at least Steam Input to simplify UI overlay creation).

All of that, and more, would be doable with a decompiled port.

[1] It would have been possible, I needed to shoehorn a java runtime in there, but it didn't sound fun anymore and only had a few hours to play with friends

> The Schuko [...] outlet is not flush

The look can be improved by using receding covers. I am sure they exist for Schuko outlets, here is one for a type E socket (CEE 7/5, usually compatible with Schuko AKA 7/3, using 7/7 plugs): https://www.domomat.com/53812-prise-de-courant-affleurante-o...

As a side note, the type E is polarized, though on occasion you can find modified 7/7 plugs that accept both orientations (I see this most commonly on PLC adapters). IEC 60906-1 was designed to replace Schuko but was not adopted.

Well, I have a VPS, plus multiple physical servers in different locations (parents, etc).

The VPS is useful for:

- sending e-mails (hard to improve your reputation on residential IPs and get a reverse DNS record)

- Unifi controller, though I appreciate Unifi less and less.

- Uptime monitoring / status page

- I think I'll move my headscale node there, since it has the highest availability -- Internet was out for a week last month, which is exceptional, but rarer in a DC

- backup Internet connection (especially incoming packets, as it's hard to port forward on mobile Internet).

- netdata for monitoring

Now, I use my servers a lot more:

- HomeAssistant on all of them

- power monitoring to warn of power cuts when not at home (freezers!)

- netdata

- yggdrasil

And for the one beefier instance:

- esphome

- Wireguard links between it and all nodes/VPS, plus a few devices -- I like Wireguard, but Tailscale is much easier to manage for full mesh and just adding personal devices.

- jellfin, *arr and torrents

- calibre web (mostly humble bundle DRM-free books, actually)

- e-mails (receive, I should make it more redundant though)

- Matrix server (synapse). For years this was the most resource intensive service. It's much better now. I do not recommend due to spam waves.

- headscale

- Nextcloud, OnlyOffice, Kanboard: for productivity

- Prometheus, grafana & monitorix: monitoring, plus netdata aggregating data from other nodes

- Open Sondage, Framadate, I hate money: for friend groups

- Vaultwarden: password storage

- Wallabag (should use it more or may retire soon)

- Ntfy: for UnifiedPush

- LDAP server: for centralized auth

- Borg backup

- Some custom web apps and probably more I forgot

Services I plan to host soon: Garage (S3), split across nodes, Ente (currently using the hosted version)

Services I retired or only start on demand: Funkwhale, Peertube, ampache, overleaf, ethercalc & etherpad (both had too high CPU use on idle, not worth 5-10W). At some point I had game servers behind systemd socket activation to only turn them on on-demand.

Yunohost is pretty good for starting, but it's quite amateurish, I prefer declarative configs and containerization these days.

I also think pretty hard about ways to get my power usage down (around 30-40 W on my main, next upgrade should bring it to 25), so I'm monitoring the usage and correlating to power use (via a smart plug). I wish there were more turn-key solutions for putting nodes to sleep in kubernetes & others, I'll likely investigate this at some point. Tiered storage systems also tends to focus on perf, not power efficiency.

And yes, I'm considering moving more things behind tailscale to reduce the attack surface. Some containers listen to localhost and I use ssh -D to access them, but tailscale would be more convenient.

FYI your Samsung televisor may not have been actual "mini led" https://arstechnica.com/gadgets/2026/10/tcl-is-right-to-ques...

Speech to text I assume? Maybe it has to do with your a accent or pronunciation? You could contribute a bit to Mozilla's Common voice, if that's the case. I assume it is part of every STT training corpus.

This is not far from piracy, but I got them on G2A. Same advice if you want to get Unreal Tournament and the original Prey on Steam.