QNAP seem to have a history of pushing out bad firmware updates.

After the Deadbolt ransomware, QNAP started enabling automatic updates by default. If you updated your firmware, QNAP enabled automatic updating regardless of whether you had it enabled or disabled prior. They didn't tell anyone. You just had to *know* that you needed to manually disable automatic updates each time you did an update. I do update my stuff, but I need to know an update is reliable before I go live with it. That is not much to ask.

About three years ago QNAP pushed a dodgy update that failed and corrupted my RAID array. I wasn't the only one affected. Luckily I was able to recover most of my data. I ditched the QNAP OS pretty quickly after that. Luckily I was able to get Unraid up and running on the hardware, although the CPU was a little too under-powered for much more than that. I wouldn't touch QNAP again, even if they paid me.

Sony has already got themselves into trouble for deleting purchased content. At the end of last year, Sony announced that it would be deleting Discovery TV content from users accounts due to licensing changes, despite the fact that users had specifically PURCHASED this content. They were just going to delete it with no recourse for users. It was only public outcry that forced them to backtrack.

Moving to a digital distribution method may have some advantages, but what is to stop Sony from doing this kind of thing again with purchased games? We're just buying a license to Playstation games, that relies on that license being honored forever. Sony is effectively just saying 'trust us', but previous behavior from them suggest that we cannot trust them.

The article talks about Dell moving to soldered RAM to solve weight and size issues. Dell created the CAMM standard in the first place to solve these very issues. They made the standard available.

But now, after the JEDEC have settled on the CAMM2 standard, Dell are moving to soldered RAM. Seems that there is clearly another reason for the move that Dell just don't want to admit publicly.

They make more money this way. Soldered RAM is cheaper to manufacture, and when people find that a model isn't upgradeable they typically spend more to go for the higher spec unit. Apple worked this out a long time ago, only they didn't stop with RAM, they did that with SSDs as well....

I do not believe that this is some economic tit-for-tat. I do feel there is something far more sinister at play.

China has learnt a lot from Russia's invasion of Ukraine, in particular how Russia was over-reliant on Western technology and how the international technology sanctions have impacted Russia's combat capability in a big way.

Xi knows that there will be a similar global backlash if they try to invade Taiwan.

I may be overly pessimistic, but I do wonder if this 'Delete America' policy is some sort of preparation for a time when China will become a global pariah (e.g. from invading Taiwan, and/or massively escalating the situation in the South China Sea).

Right now Taiwan is a key chip supplier for the rest of the world. If China invades, that chip production will stop. If China is successful, it is more than likely Western interests/Taiwanese military will destroy the Fabs to stop them falling into Chinese control. If China is unsuccessful, it is likely that the Fabs will be offline for months to years, and China would also be heavily sanctioned. Either way, strengthening domestic technology capabilities now makes perfect sense. To take it a step further, if China's domestic production were sufficiently advanced, they may make the Fabs the first target for destruction instead of attempting to capture them intact.

Unfortunately, I don't think it is a case of IF it will happen, but WHEN.

I have found a bit more information on some of the methods of 'poison': tools called 'Glaze' and 'Nightshade'

My guess is that, in addition to false image metadata, they use a form of steganography to embed a picture within a picture. This is done by adjusting the mathematical relationship between the bits of each pixel and their adjacent pixels. What may appear to be a solid block of color to the human eye could be an arrangement of pixels with subtely different values. While we see the image, or lack of, the AI training sees the mathematical relationship between pixels. So I can see how creating a strong mathematical relationship within pixel information could impact AI training. Steganography is not new, but it is interesting to see this being positioned as a method against AI.

However, I don't see how this could survive things that would destroy subtle relationships between pixel values: running a filter that adds low level random bit noise; reducing the bit depth of an image; or re-encoding with a significantly lossy codec (like JPEG).

I think some people will see it as a challenge to break this. Much like the constant battles seen with other DRM systems. However, a tool actively seeks to poison an AI dataset, is only going to serve to make people far more motivated to break it or workaround it. I don't think it would be long before someone creates a tool that can detect these images. Then it could either remove the image from the training dataset (which I can see the artists preferring) or altering either the image or training process so that the steganography is negated.

192GB RAM sounds like a lot, and it is a lot to most people, but when you're talking about mainipulating massive data sets as someone may want to do on a Professional Workstation, then 192GB RAM may not be enough by a significant margain. The previous Mac Pros could be expanded up to 768GB on the 8, 12 & 16 core CPUs, and up to 1.5TB on the 24 & 28 core CPUs.

I'd much rather see AMD expand ROCm to support more of their existing chipsets, and work on things like improving integrations with frameworks like PyTorch. It is one thing to build an expensive Tensor GPU, but not everyone has $40,000 per GPU (eg. Nvidia H100) to throw at AI. Yes, AMD are well behind the curve, but I think there is an arguement to be had that the lower end is also valuable too, and an area that AMD could really capitalize on, only their software support leaves a lot to be desired. They are already touting that they have a VRAM advantage. As anyone that has played with Stable Diffusion or LLaMA text generation may tell you, VRAM is really important to handle larger and larger models. Their software support lets them down. I tried to get an RX6600 working with Stable Diffusion... evenually I did, but it was a massive pain as my card was not 'officially' supported by ROCm. It did work, and was far better than CPU alone.... But, in the end I ended up switching to an Nvidia RTX3060, mainly because it had 12GB VRAM (the most I could afford at the time). Getting Nvidia working under PyTorch was super simple compared to the hoops I had to jump through with AMD.

Instead this is just a verification of an accepted payment method.

More worrying is that those who pay this subscription get priority replies and mentions. So what Elon is really saying is that the speech of those that give him money is more valuable than anyone elses. This goes directly against his PR spin for Twitter to be a platform of free speech. Those that pay the money will control the narrative. It isn't hard to see where that is going to go.

Time for the rest of us to move on to something else.

JPEG XL is effectively competition to the WebP and AVIF image file formats. Google controls the WebP standard, and is a member of the Alliance for Open Media which contril the AVIF standard, but Google have the same control over JPEG XL. I believe it really is that simple. Google have been quietly pushing WebP on users for a while now.

Absolutely not a hunch. DART was launched 21 November 2021 and impacted 26th September 2022 (ie. it took 309 days). The vast majority of close calls (asteroids passing Earth at less than 1 Lunar Distance) have either been detected with less than one week's notice or (more alarmingly) have not been detected at all until after the fact.

Being able to shift the orbit of an object in space is one thing, but for this to really work as protection we need to be able to detect objects early enough that an impactor can be launched and intercept an asteroid with enough time for an orbit alteration to prevent collision.

The problem is that, with our current observation capability, we wouldn't know until it is too late to actually do anything.

My first machine was an Amiga 500... I spent a lot of hours on that thing. I still have it. Last I checked, the caps were still good and it started okay. The same can't be said for many of the floppy disks I had for it. Sadly, I don't have the ability to attempt to recover the data (thanks ecpecially to the different standards between Amiga and PC floppy drives). I should probably make some time to sort that out.

I still have my first PC too. A 100MHz Pentium with 16MB RAM and a 1.27GB HDD. It came with Windows 95, but then I installed RedHat 5.0 on it.

The machine I first learned to program on was a BBC Micro with 32K or RAM. Fun times with BASIC and LOGO. I don't have one of those, but I do have a MiSTer FPGA that can run a BBC Micro core (not to mention Amiga amongst many others) so I can still get a nostalgia fix.

Deadbolt is not new. They have known about it since early this year.

Back in February QNAP pushed an update to QNAP devices, completely unannounced. Many users reported failed updates lost data and/or bricked some units as a result.

A lot of users, myself included, had specifically disabled automatic updates but QNAP was somehow able to push out the update anyway. This, more than DeadBolt itself, annoyed me the most. If QNAP could bypass users to install updates (regardless of whether automatic updates had been sneakily re-enabled in a prior update, or of QNAP had some method) then I could no longer trust them on my network or with my data. I pulled the NAS, built a new server, and have repurposed the NAS hardware for something else with a new OS (the hardware itself is still pretty good).

The fact that DeadBolt is still an issue for QNAP after seven months just tells me I was completely justified to ditch them when I did.

Marmageddon was a thing in New Zealand. After the 2011 Christchurch earthquakes damaged the manufacturing plant some production kept going, but production was fully shut down in late 2011 when a critical cooling tower was declared unsafe. Further damage uncovered during the repairs kept the plant closed far longer than expected. The resulting marmite shortage lasted from 2012 to 2013.

British Marmite has a different fomula, and tastes completely different, as does Australian Vegemite. Some people switched, but most held out with remaining jars reaching astronomical prices on TradeMe (the local Ebay equivalent).

Apple will just write it off as the cost of doing business in Europe. For a company the size of Apple, $5.7 Million per week is nothing to them. This fine is not a deterrent in any way. When Apple's revenue is over a Billion dollars a day (annual revenue $365.8 Billion in 2021), then $296.4 Million ($5.7 million * 52) of fines per year is nothing.

Apple know what they are doing is wrong. They don't care. They know that they will have to change eventually, but all they need to do is throw some tens of Millions of Dollars/Euros at lawyers and lobbyists to slow the process down enough so that they can maximise their profits while finding the next way to lock their customers in.

If the fines were instead exponential, doubling every week of non-compliance, then Apple would have to quickly make changes. Sadly, I don't think European law allows for this.