The problem I have with ECC discussions is that it’s always anecdotal. There is one Google study finding memory flips with some probability in their datacenters, but it was awhile ago. Everyone repeats the statement DDR5 on-die ECC isn’t a replacement for regular ECC, but nobody ever shows data on DDR5 bitflips. That’s not a knock on this article per se, but it’d be nice to have actual data here.
Good to see Codex as a datapoint: I haven't used it. I can confirm though that this is certainly what Claude does, with the addition that Claude's adherence to its own string-match based allowlist seems even less deterministic than OpenCode's.
The developers are used to Java, and as someone who's worked with both Java and C++, Java is simply better for the kind of iterative development minecraft seems to do. From what I've seen (which admittedly isn't much more than the public facing history and the minecraft live/minecon events), there isn't really a strict roadmap and they seem to value messing with features a lot. Also, the game was only rewritten in C++ to support multiplatform ports, and bedrock isn't even the first time they did it (though, the other C++ versions, also known as legacy console, weren't handled in house). The existence of a windows version is a strange curiosity stemming from the microsoft buyout and their ability to do so since that codebase is insanely portable.
Also, mojang (not microsoft) really appreciates modders, which java edition is strictly superior for
Thats fair. There are a number of safety checks before the original files are deleted, and this is also why the keys are deterministically re-creatable.
But, I also don't think there's any substitute for backups when using encryption software.
For those asserting that Kimi identifying as another model is evidence of distillation, is this evidence that Claude was distilled from Chinese models?
> Upload your file anywhere: You can upload your org file to any server you want. You can use Apache, Nginx, GitHub, GitLab, Cloudflare, Dropbox, or any other server or service you want.
to me sounds like "You can step into a wardrobe, close the door, look into the darkness and start talking".
Typically, no one will notice the files you upload to your server or GitHub etc. No one will hear you talking in the wardrobe
Which makes sense, but as long as the tool itself pretends to provide security features, the fact that these features don't work is a big problem, as it may provide a false sense of security to end users.
It'd be better if they had absolutely no permission enforcement and delegated it entirely to another program, as you say.
Something important not discussed here is that you need not only error correction (when possible) but also error reporting, so you can know you RAM is failing. Sadly, is not possible to get DDR5 on-die ECC error counts reported to the OS. Alternative solution, use IBECC with DDR5, where the CPU reserves a portion of the RAM for parity bits and does the check / reporting.
For the past few years, I have been learning different programming languages. In the process I realized every website starts with the exact same boilerplate: Auth, OAuth , stripe, s3 file uploads and these days AI chats and/or AI embeddings.
To make learning and building, I bundled these into templates across multiple programming languages. Inspired by Erlang's OTP, I also built an actor supervisor system around each of the stacks to handle the process reliably.
I have a few previews for everyone to play with and provide feedback.
Yeah someone thinking they screwed up to the tune of tens of mlllions of dollars (or billions, someone said 2^30 error); this is different to an auto accident or something. Most people can't even fathom millions or billions of dollars.
I'm stressed thinking about it, like from a theoretical standpoint.
This might work really well in practice idk, but I think it's not allowed by big O to pick a maximum supported size. Otherwise you could just make a lookup table. Your algorithm must be ready for anything.
> Why did Airbus go to "the cloud" in the first place?
The same reason that so many companies went to "the cloud", because the hyperscalers warpped the C-suite round their little finger.
There are two things any C-suite like doing that the hyperscalers could help them tick boxes on.
First, shifting capex to opex. They love doing that because of the accounting treatment. Opex goes into your P&L, so it reduces your taxable profit and therefore it reduces your tax bill. Meanwhile capex goes onto the balance sheet and is deprecated over the asset's life, so the "only" benefit you might get is to claim allowances (e.g. R&D) from your tax authority.
Second, it allows lots of things to be made "someone else's problem". Security ? Their problem. Staff ? Their problem. Tick-boxes to keep the compliance department happy ? Their problem.
So the hyperscalers can turn up and spin a ready-made story that the C-suite will lap up. Perhaps accompanied by a nice lunch and a round of golf. ;)
There is no doubt, companies such as Airbus have the technical competence and the financial means to do it all in-house. Sadly the C-suite calls the shots ....
They need to focus on archaic government processes which rely on document formats that require proprietary software to view properly, while good open source software for the same task exists. And where lacking, focus on improving open source support for a particular task.
A good start would be to cut use of this OOXML Transitional format entirely. Default to LibreOffice's default formats, and (perhaps) support OOXML Strict where it helps transition.
That way any government employee saving their work in OOXML Transitional or other proprietary format, would automatically stick out as a dumbass who failed to do their job. Pass documents that are published through some validators / allowed-format checkers, and the public only gets stuff that's viewable with software of their choosing.
None of this is hard. It's mostly about changing defaults & a bit of employee training.
Interesting article, and I agree that ECC needs to be more widely available. But I think this is a major overreach:
> We need ECC RAM to be more widely used. Ideally we would have some government action to force this given the ongoing cost to society in corrupted data and lost time due to RAM hardware errors. I think that at minimum we need sufficient taxes on non-ECC RAM (and EC4 RAM for DDR5) to make it more expensive when bought new than ECC RAM.
ECC RAM is not needed for every application and mandating it for every gaming rig and machine used to run Chrome will just increase prices unnecessarily. 99% of the time it's also not a "cost to society" but a cost to whoever owns/uses that machine. If they want to eat that cost, it should be up to them.
Not OP, but your comment had me looking into it: the fact that it's Firefox-based is reason enough to give it some consideration: the Chrome monopoly we are currently under is not compatible with a healthy and neutral web.
Rolling this stuff back doesn't instantly restart all of the science that was built up for decades. Nor does it get the smartest people back into the country after being kicked out or scared to stay.
Damage is done, we'll forever be behind, this isn't the place to go to research anymore.
Pointing out the context for an argument and the many articles and discussions that have taken place on that subject in the news is not the same thing as participating in or taking sides in the discussion. I take offense that you have made this personal.
For any US visa you do, and you can't take electronic items or bags with you. Costs a lot more too.
A reason I refused to go to Syria a couple of years ago. A colleague who did go said "oh I have a US (i) visa so I'm fine". I asked "how about in 20 years time when you've left a company which will pay for you to have a US visa. He hadn't considered that.
A friend of mine went to Libya on behalf of the US military about a decade ago. He is no longer eligible for a US esta and thus has to get a visa if he goes on holiday.
Last time I went to India I had to get a real visa which involved travelling to an office in London and waiting around for a couple of hours.
It's quite normal for people of "lesser powered" countries who don't have the power of passport that the US and Western European citizens do. Even before the Russian invasion of Ukraine, you would have to fly for hours to a major Russian city a couple of weeks in advance to apply for a visa to visit the UK.
When I went to Pakistan in 2010 it was even worse. Well until I gave up and then used a visa agency which just handled it all for me.
Last time I tried to go to India I went for an electronic visa, and after 7 days of "waiting" they updated it to say "declined". No explanation.
It has happened people I personally know. It wasn't just a truck passing, it was a truck passing with a good cross wind, which is why it isn't common. This wasn't a small tow vehicle, it was a F350 which is a large truck even by US standards, and it has a heavy duty frame. At least the truck was strong/heavy enough to say out of the ditch and could safely be moved to the side of the road unlike if it was a smaller tow vehicle.
