Posts by Crazy Operations Guy 2513 publicly visible posts • joined 29 Jun 2009
You can turn those functions off, or you can replace the bootloader. Even stock android doesn't require that functionality. Apparently some apps do, but I've never encountered them, even security-related apps don't seem to care that the bootloader isn't checking. From what I've heard is that the only apps that care are those that are exclusive to a manufacturer / carrier and are trying to pull private keys embedded into the device. The only affect I've noticed is that when the device boots, a screen flashes up saying my boot loader is unlocked and points me to a website that tells me that it could be a security risk, especially if I wasn't expecting it.
I run LineageOS on a Pixel 3 and AOSP on Galaxy S10 (a work phone). Doesn't get much more 'modern' than them (Well, other than the version that came out in the last few months).
A lot more. Lennart Poettering is an employee of Red Hat. Aside from that, there is a lot of his design philosophy floating about in the company. NetworkManager came from Red Hat... They also produce a bunch of other minor, but equally problematic utilities that try to do a bunch of stuff at many different layers of the OS simultaneously, but just make a mess of everything.
The root problem is that Red Hat is a for-profit organization that lives on selling licenses and support contracts. Which presents two huge temptations: One, to make something flashy and impressive sounding so that users don't go elsewhere, even if the big splashy thing they make is barely functional and badly designed. And, two, to make something that requires support to get it running and working smoothly. Not saying they are doing those things, just that there is a very strong temptation to do them.
I upgraded away from RHEL to Debian. Red Hat has become a flaming pile of trash, especially now that they've thoroughly drunk of Poettering's Kool-Aide. I can only imaging getting worse now that the Big Blue Whale has swallowed them up.
Although, really, for the most part, most of my systems aren't a distro at all and produced using Buildroot. On most of my systems, I have apt installed to grab packages from the Debian mirrors, but most of my machines are single-purpose (DNS, DHCP, httpd, etc) and I just go without a package manager on those systems. Hell, most of my systems aren't much more than a kernel + busybox...
But then, I am an old grey beard, I've switched to Linux when the lawsuits against 4.2 BSD started flying. I've used Windows periodically since then on work computers and "gaming" machines since then, but never as my primary system.
"many high-end photo printers aren't supported under Linux "
But most printers, and especially "high-end" ones will come with a USB that allows you to just plug in a thumb drive and print straight from it. Even the cheap Brother printer I got from a big box store can print from PDF, XPS, PS, and so on. Plus it gives me a backup of what I printed.
I was printing like that long before I moved away from Windows because 99% of the time, the printer driver will come with all sorts of random crap along with the piddly single-megabyte .sys file I actually care about. The last printer driver I installed was a 250 MB download that also installed a buggy update service, a 'photo manager', some trialware image editor, and constant pop-ups telling me that my supplies were low and should buy at an inflated price from the manufacturer's online store.
I feel fortunate that I have the skill set to replace the OS on my phones with something clean. But not everyone can do this, and it might not even be possible on some of the lower-end phones with their proprietary SoCs that require substantial proprietary knowledge to compile a kernel for.
Google could easily fix it if they just required that phones shipping with their OS on them must allow for the user to remove everything. Google has the power here and yet they act like the victims. Phone manufacturers can't exactly move to anything else, Windows Phone is dead, and Apple sure as hell isn't going to allow their OS on anything they didn't make themselves. Yes, there are other OSes, but there is no way they can compete against the Google/Apple duopoly, unless they can figure out some way to convince app developers to also develop for their platform (And it would have to be a lot of them).
If anything, tighter security increases the risk of attack. Terror organizations like to recruit people who are frustrated with a government, especially if that government is doing it for reasons beyond their control of the frustrated person. Much like if you were stuck in a traffic jam and someone came up to you and handed you a bomb, telling you to blow up the car in front of you. A lot of people would do it without a second thought.
Of course there is also more immediate risk at security checkpoints. You have a bunch of people clustering in a small area, if someone were so inclined, they could rack up a very high body count (And the fact that that hasn't happened is pretty telling)
Really, every bit of evidence I've seen of the crash leads me to believe that the pilot probably just set take-off power, and left it at that power level for a bit too long and caused the turbine to overheat and eventually fail.
Probably left it at take-off power longer than recommended, trying to get as far and away as possible to minimize the risk of a repeat of IranAir 655 or MH 17. 99% of the time, leaving the engines at full power for extended periods of time is just going to result in needing the engine to be inspected and some parts replaced, and maybe a reprimand. And the engines were relatively new, so likely just a repeat of Southwest 1380 or 3472 would've been far likely the worst case.
The ADS-B data from the aircraft showed the aircraft climbed much quicker than what would be expected in normal flight, but would be consistent with staying at take-off power beyond the initial climb-out.
These features are nice and all, but I really wish that they'd do something about the memory/cache problems.
It'd also be nice if there was some way to keep videos from auto-playing or at least not just soaking up bandwidth buffering when I have given zero indication that I'd actually watch the video. This is particularly annoying in combination with the caching issues. A lot of my daily browsing is on documentation sites for the various products my company uses. One of them has a video up on every single page that is essentially a re-hash of the written text, rarely its useful, usually it means several megabytes of bandwidth wasted per page.
I'd also like the ability to just turn off most style sheets on some webpages, or tweak the style a bit. Like I'd very much like to remove a lot of background colors from chunks of text, or at least be able to shift it to black-on-white instead of dark-grey on somewhat-light-grey. I'd also live to be able to expand the width of content frames. I have a 4K monitor, but most websites are configured to only be 1024 pixels wide, so nearly 3/4th of my monitor is wasted displaying background pixels.
Also possible with some trickery in kconfig. If the onboard audio uses a different driver than the card, then bake its driver into the kernel and the card's driver as a module. The built-in will be detected and setup shortly after the kernel starts booting and the card would need to wait until /lib is mounted to load its module.
The problem is the very nature of Microsoft. It is a massive publicly traded corporation that can only survive by prioritizing profit above all else, and that includes the health of the industry at large.
As a publicly traded company, their first and foremost goal is to enrich shareholders. And that takes priority over any other policy they may have. That goal will outlast any executive that ever gets appointed. If an executive does something that sacrifices shareholder profits, they get voted out at the next board meeting and a new executive is installed. Being hostile to OpenSource is the nature of the beast.
They see a profit in supporting the OpenSource community right now, they'll flip back to being a patent troll the second that betraying the community presents a greater opportunity for profit.
Not sure if you are being serious, but I have to point out that there is a a .ibm and .apple.... Also a .microsoft, .dell, .cisco, .azure, .google and so, so many other TLDs owned by companies (Who precede to do absolutely nothing with them). And its not just tech companies, pretty much every Fortune-500 has snagged theirs. Many of them have tiny DNS zones that are just a handful of cnames to point people to the .com equivalent of the company's web presence.
Also of note: Amazon already owns 54 other TLDs, second only to a domain-squatting company named "Binky Moon", which is owned by a former Amazon employee (owns 197 TLDs).
If you want to see the horror that is the current DNS infrastructure see:
https://www.iana.org/domains/root/db
Yeah, I'm bracing for Blowhard Bob, or one of his identical clones, to burst in with his normal rhetoric about an "SJW Conspiracy" and how he is being persecuted. I can see it now, all full of randomly capitalized words and full of complaints about how everyone gets so offended these days (And completely lacking in self-awareness of how easily he gets offended).
One of my engineers came from a company that didn't even have any infrastructure of their own, rather they just used cloud-hosted and 3rd party stuff for operations. The company's documents, including the passwords to pretty much everything, were stored in a Dropbox instance that everyone had access. Their reasoning was "We don't believe in job roles, if something needs to be done and someone has the skill to do it, they should be able to!", a philosophy that they snagged from another start-up. They were trying to claim that "This is how they make Linux!", completely ignoring how wrong that is. They reasoned that if someone was malicious or incompetent, they could just undo their changes and push the application back out to AWS Lambda.
And yes, this company is in Silicon Valley (Well ostensibly, they don't have an actual office and instead employees work from home and/or WeWork type spaces).
I figured that GitLab might be doing something equally weird.
Well, yeah, that is an acceptable exception to anti-discrimination laws. You are refusing to consider them for the job because they can not meet the requirements of the job and not because of an immutable characteristic of the candidate (EG, race, religion, etc).
But, also, I am talking about the employment laws that are implemented at the federal level for all employment in the US, not the laws surrounding federal-level employment. There are exceptions in the anti-discrimination laws for employment in roles that are safety critical or are national-security sensitive.
Having used Subversion, you really don't have to do much to keep people away from it, just let people use that nightmare for a day or two and they'll run screaming.
(Yes, I know you are talking about the act of subterfuge and not the code version system of the same name)
Even though this wouldn't be *explicitly* racist, a lot of anti-discrimination laws, including the federal-level one for the US, forbid discrimination on the basis of "National Origin", in addition to race, religion, etc. So this decision would run afoul of those laws, just not for race.
Ummm, they should have many different classes of employees each with a different level of access to systems. Its called role-based-authentication and least-privilege and pretty much every company with IT infrastructure has heard of it and is doing it.
Google isn't loyal to any party, they'll support whoever they think will lead to the greatest profit increases in the future. They supported Obama and the democrats in 2008 because they were pushing bills to encourage adopting new technologies, like cloud-based email. Google now supports Trump and the republicans because they are pushing for deregulation of private enterprise and reducing corporate taxes.
Google isn't unique in that regard, every major corporation, especially publicly traded ones, are loyal to no one and nothing but the dollar.
Oh, I am very much aware that its showing up in the western world. I suspect China was inspired by us when they created their system. I also very much believe that the Orwell estate should sue Amazon and Google for plagiarizing 1984 so blatantly.
The Social Credit System.
Its not about building a positive score, but rather avoiding scrutiny. The rules of the SCS are extremely broad to the point where even a basic look into the life of the average person will result in them being denied access to a lot of necessities. The point of the system is that they don't have to create a blatantly draconian law to hang their detractors, instead they let people weave their own nooses one antisocial fiber at a time until the state decides to hoist their self-made noose. Even a lifetime of supporting the state won't add up to enough points to avoid the gallows.
China lacks the personnel and technology to actively monitor their entire population, but they don't need to. They punish the most noticeable on their radar, which convinces most people to avoid doing things that would cause them to be noticed. Things like being part of the minority not parroting state propaganda.
Its less a gear change like a transmission and more of a governor like you'd see in modern turboprops. I too may mistaken, I am going off of what engineers have explained to me by some GE's turbine engine R+D engineers that had worked on the LEAP (We are members of the same EAA chapter). I tend to trust them as they hold zero loyalty to their employer and instead equip their craft with engines from multiple different manufacturers even though a GE engine would suffice.
They also seem to have permission to grab whatever they want from the GE Boneyard, one of them has a Velocity V-Twin with a Rockwell-Collins ProFlight flight deck that came off a King Air).
The problem with MAX's is that Boeing was allowed to determine whether they could reuse the type certification from the -700/800/900 model series. This is normally allowed when the proposed modifications do not modify the aircraft's performance and handling characteristics significantly like when the 900ER was rolled out. However the MAX series, which used larger engines that were moved forward, produced significant changes to handling, which are corrected with the MCAS system (Which was approved as a pilot aid, rather than as a critical component to keeping the aircraft properly trimmed). Boeing never should have been allowed to make that determination.
The new type certificate would have required much more rigorous testing that would've likely caught the problem and redundancy would have been mandatory, or the FAA could've reached the conclusion that the aircraft is unsafe by design and denied Boeing approval until they redesigned the system so that the aircraft can perform properly without no matter what failure mode the MCAS system experienced. This would have also required that pilots operating the MAX series to have received a new round of training that would have included awareness of the MCAS system and troubleshooting issues with it.
In my opinion, any aircraft that requires electronic system making calculations to maintain stability is an aircraft that needs to stay on the ground until it it gets recycled. I'm okay with fly-by-wire systems since they are designed to fail-safe so that the controls directly modify the position of the control surfaces instead of being put through a computer that prevents moving the control surfaces to unsafe positions.
Assuming both FADECs are on different software versions, but I wouldn't be surprised if both have had the same modifications made to prevent a thrust mismatch between engines.
Given that the PW1000 engine series is a geared-turbofan design, I'm assuming the software change is to expand the range where the lower gear ratios are used, but accidentally deviating outside of the safe range for air stream densities entering the compressor stage and causing an over-pressure situation causing cavitation on the compressor blades. Or maybe a back-pressure situation causing the fan's vanes to cavitate.
But thats the problem with squeezing your margins for efficiency, sometimes you end up with your performance curves crossing to points outside of your new margins...
Engines are independently tested, twice even. Once in approval of the engine itself, then a second certificate indicating its suitability with a specific aircraft type as part of the aircraft's type approval. The manufacturer also receives certifications of their own indicating that their manufacturing processes are sound, they have appropriate testing facilities, they perform their due diligence in auditing suppliers of components, the employees building the engines have the appropriate certifications of their own, and so on. Parts of the supply chain also need to be certified (Although those certificates are handled by other agencies).
While this is a very positive move forward, it feels very much like a 'fox guarding the hen house' situation. Google is in the business of tracking and profiling us and I can't help but think that they are leaving some intentional gaps in their policies to allow them to do their tracking, but lock out their competitors. Until the company is split, I will forever assume that any move on their part is a cynical ploy to make more money and/or eliminate competition.
Yeah, but that pushes the legal problems to them and away from ImageNet. If Getty provides an illegal or unauthorized photo, its on Getty, from a legal perspective. At least it would provide ImageNet with ability to sue Getty for providing a faulty product if ImageNet uses a bad image.
There are dozens upon dozens of companies that are sitting on top of literally billions of photos that have already been cataloged, tagged, and the people pictured have given full consent for their image to be reused in diverse contexts.
It would have been fairly cheap for Image-net to just get in touch with the likes of Getty Images or ShutterStock and negotiate an "Academic redistribution" license. Then feed those images to Mechanical Turk to produce the boundary boxes and normalize the tags.
As for the CSA images, those should be split into another dataset that is strictly controlled. I figure such a library could be controlled by a law enforcement agency like the FBI that acquired the images by having victims / parents of victims sign a contract to allow the images to be used in such research like they do for getting permission to reuse confiscated material in sting operations. Maybe pair it with a library of equivalent images reproduced entirely with adults so as to remove as many possible variables when learning to differentiate between legal and illegal images Like avoiding that 'skin cancer' AI that was making decisions based on whether there was a ruler in the image and not the appearance of the skin blemish).
The performance issues seem to lie in Firefox's rendering / add-on pipeline, so cleaning up pages before they even make it into Firefox's pipeline will make it perform much better. I've been running Privoxy to do exactly that on the network level. ABP, uBlock, and NoScript will run much better as well since they won't be bogged down with crap either. Other proxies will work, I just prefer Privoxy for its simplicity and ability to import ABP's block lists without too much difficulty.
I also prefer the network proxy solution since I can provide ad-blocking on all my devices without the devices needing to support add-ons.
Patent trolls do sell something: Licenses of their "Intellectual Property" to people that they have sued for patent infringement. The basic procedure is sue a company for something that technically infringes the troll's patent and the victim cannot easily replace, and then as part of the settlement, the troll offers 'reasonable terms' to license the infringing patent.
I support them doing this because each dollar going to fighting patent trolls is one less dollar towards infecting their desktop environment with more systemd effluence, or spreading the infection to other software.
I support going after patent trolls anyway, this other bit is just icing on the cake.
Storage dependent on a high-level protocol to function? That's just asking for trouble... Well, so is relying on DNS TTL-abuse to do the work of redirecting clients when a storage endpoint moves rather than doing the work of writing the protocol to handle that sort of thing properly.
By properly, I mean something like the client connects to a 'storage Director" service, sets up a persistent connection and requests the location of a specific storage bucket along with a flag of "Inform me if the location of this bucket changes". That way the client isn't hitting the server every few second to re-resolve the bucket's location. At the very least, a DNS outage would only cause failure of any new client connecting to the director system, but already connected clients wouldn't have a problem.
Why not just move firmware to be read-only by default except when inside of the BIOS configuration utility.
Or set it up where all devices have some kind of ROM that stores a known-good copy of their code. When the system boots, the BIOS enumerates devices, then uploads any updated device code to a chunk of RAM running on the devices, then send an 'initialize with the software in RAM, ignore your ROM' (Or if the BIOS lacks any updated code, tell the device to initialize normally). When a user goes into BIOS, there is a big list of the firmware files that the BIOS has and the user has the ability to load updated firmware from external media, or disable/delete the existing firmware files. The OS would only have read-only access to the BIOS. All the firmware on the machine, including the BIOS itself, can be reset by opening the machine and shorting two pins, just like what you'd do to wipe the BIOS's configuration data.
A piece of electronics that can be forever tainted by having been in a system that had at one point run malicious code is a terrible model, and trying to fix it by doing anything other than just nuking it all and starting over properly is just foolishness itself. Especially when the proposed solution is to further reduce the owner's control over the machine.
Solving a Rubick's isn't all that impressive, especially with the seer volume of computational power that was put into it. I've seen Rubick's Cube solving devices built out of a couple servos and a color sensor wired up to a Arduino. Sure, the system didn't "Learn" to perform the task, but it was programmed by a hobbyist working in their spare time. And I am willing to bet that that programmer didn't expend nearly the level of resources 10,000 years of CPU time would burn.
A human produces approximately 15 tons of CO2 per year to support. Assuming that the AI burned 50 Watt-hours per hour of compute time, you are still talking about 4383 tons of carbon dioxide, or 300x the amount a human would expend. These are extremely rough calculations based on averages and assuming that that human expended 8766 labor hours to producing the code. This is also based on per capita emissions, so includes industrial emissions in that calculation (Although some of those industrial emissions would be from producing necessities for that human. Really, there are just far too many variables to deal with here). And this is even being generous in assuming that the AI can be considered as complete rather than only 20% accurate, so we are ignoring the future computational hours needed to solve the problem (Which I think fairly balances out the somewhat pessimistic wattage calculations of the machines in the AI cluster)
"possibly many years in prison in the US."
I would think that he'd only be imprisoned for a few days ending in him leaving with his nose and toes at the same altitude. Assuming he makes it to sentencing and it doesn't just happen in pretrial holding ala Epstein.
Doesn't even have to be assassination, Assange clearly requires mental health care, and I would't put it past the US to press him beyond the breaking point (Even just being in a US prison might be enough)
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
