rocqua

Not who you asked, but biological brains come about very differently than LLM neural networks. Most notably, they come about largely through plasticity, and with evolutionary pressure towards exhibiting plasticity.

My guess would be that the structure to allow for plasticity to work is a strong limit to how well you can tune weights to fit training data. My (doubtless massively oversimplified and way behind the state-of-the art) understanding of neuro-plasticity is something like "often used pathways get stronger connections". Which, even with significant tweaks, is quite different from the gradient descent used to tune neural networks.

Also hit a few smaller business in The Netherlands: https://nltimes.nl/2026/11/02/fridges-freezers-17-dutch-busi...

With the following interesting quote from the supplier of the system:

After that attack, Danfoss advised its Dutch clients on October 13 to disconnect their systems from the internet and ensure they only reconnect via a secure network. “Most stores took action at that time, but a small number did not,” the spokesperson said. “They are now the ones suffering the consequences.” The spokesperson added that the systems hit last night were not on a secure network and were still using default passwords. The attack could have been prevented if measures had been taken, he said.

That would be cool, do you have any source that confirms they used this for taking advantage of price fluctuations?

I think it's too late, but I'd consider seeing if you can still get into those bankruptcy proceedings. If you held 3 BTC there, you could expect around 30 000$ from the proceedings.

They closed registration for the proceedings many times. Which means they also re-opened it multiple times.

I'd expect the registration is permanently closed, but worth checking out.

I thought that was usually enforced in the OS, and that this was a big reason why you need secure boot. To ensure that you cannot boot an OS that will allow a brute force.

Certainly this was the case with the iPhone when the FBI asked apple to write some custom firmware for them.

It might have dutifully logged that the temperature setting was changed, which will allow regulators to conclude that the safety of the food stored in these refrigerators can no longer be assured.

Compliance tends to warp the underlying goal quite a bit.

Yep, but once you have internet connectivity, things escalate and people say "wouldn't it be convenient if we can also...."

On top of the monitoring mentioned by others, it's also to create a centralized log of actual temperatures. This then lets you show inspectors that you are keeping up with refrigeration requirements.

I'd love to learn more about the syscalls that took 3 or 4 attempts.

The grid frequency is set directly by 3 phase generators. So if the grid frequency drops, there is indeed a very strong torque trying to decelerate the turbines.

To some extent, this is a good thing. Because it means the spinning mass of all turbines in the grid actually provide a buffer for sub-second mismatches in demand and supply. But it does mean that big demand spikes (or drops) put some wear on the systems. They are made for this, but if it gets very periodic it might be more than designed for.

Note that this means that all turbines in a power network are perfectly synchronized. And if they are not, they are subject to massive torques. This is why a cold start is so hard. You need to bring on generators from one central point to keep everything in sync. And as you bring on new generation capacity you also need to bring on new consumption, that electricity has to go somewhere.

The reason for this synchronization is because the generators are spinning a magnet past 3 coils, and those coils are directly wired into the power network. If your magnets are out of phase then effectively the entire power network will use those coils to generate a field to get that magnet back in sync. All of those coils on all those generators really are all hard-wired together, it still kind of blows my mind.

This was deliberately bred out of wheat. Tall wheat falls over more easily, especially with high winds. It then dirties the oars that come in contact with the ground, which I believe can make them moldy. So they bred shorter wheat.

Except that plenty of exchanges can also blacklist a wallet, and then block transactions that include funds that (indirectly) come from those blacklisted wallets.

Frozen funds remain frozen, until they are unfrozen.

So mechanically Tether pockets the difference, unless they unfreeze. But as a matter of accounting, because of the unfreeze capability, the liabilities of the frozen Tether stay on the books. So they can't just spend the money. It also makes it hard for Tether to compensate the victims from the frozen funds.

The alternative would be to let Tether burn funds, but that is highly problematic for well intentioned users. Because a freeze can be challenged in court and reversed. A burn cannot.

Russia uses crypto for sanctions evasion and money laundering. They also indirectly profit of it from their domestic ransomware operators using crypto. But Russia does not gather state funds through hacking. That is pretty much uniquely a North Korean thing.

Did you do the bankruptcy proceedings? They payed out quite well.

Turns out there was decent money to be made in having your MtGox Bitcoin stolen.

An interesting part of this is actually in defeating secure-boot or bitlocker. If a device is misconfigured for PXE boot with an otherwise locked down UEFI system, and strong secure boot. This could let you break the boot-chain and then disable secure-boot or extract keys from a TPM.

The main question here is how often devices you'd want to try this against have PXE boot enabled. I'm thinking of locked down laptops mostly. Either corporate laptops, or manufacturer locked-down devices (chromebooks, I think some microsoft surfaces).

Even if there is a passphrase, this would likely allow for brute-forcing it.

If you want to filter this, you should probably do that at the network level, not the host level. Ideally at switches, otherwise with firewalls.

It's not the type of problem where a useful solution can be verified in Lean. Which makes LLMs a lot less good at solving that type of problem.

I think the implication is heavy distillation.

only if you tell it to be controversial.

In practice we have Certificate Transparency Logs that let us check whether a CA has issued a certificate for a domain you control without you asking them to.

This really reduces how much we need to trust the CAs, because we can detect any untrustworthy actions after the fact. Now detecting after it happened is too late. But it does mean that mistakes or malicious actions are much more expensive to a CA. In practice this has really reduced the risk, especially to those whom aren't interesting enough to burn a valuable capacity on. Which is the vast majority of people. Though perhaps if you count TLS traffic weighted by importance that count is different.

The thesis behind private equity is: people from finance can run small business better and more efficiently than small businesses owners.

A small part of that is more efficient administration. But the bulk of it is squeezing the customers harder, and squeezing the business for credit. Sometimes combined with a sector roll-up causing a monopoly.

Trust busting would help a little. But a lot of the pain comes from the squeezing of customers. And that’s much harder to fight back on, other than avoiding private equity.

These runs are very likely also going to be a part of Reinforcement Learning and internal distillation no? As well as being analysed for what reasoning directions worked well, and which directions did not work well.

The mathematical results are valuable as PR, but the real value might be in the improvement this can bring to other training runs.

It’s not ‘spare goes to training’. These data centres are training first. Inference is really rather efficient, so not why datacenters are so big.

I think it’s just that the dangerous jobs are avoided y those who can avoid them. So they are left to those who have little choice, which depresses wages. This is rather immoral, but Covid made quite clear that the essential jobs weren’t the high-status or high-paying ones.

I find this worrying because the entire economy is built on the idea of cheap physical snd cheap unskilled labor. So if we do a good job educating, we’ll need to fill these jobs by paying more rather than preying on the desperate. And I don’t see the current world being willing to pay more.

After the Bronze age collapse, the Roman Empire rose. Once that collapsed it also took a very long time before another big empire arose (with perhaps a brief exception under Charlemagne).

It seems instead that the process of building a bureaucracy that can support an empire is tough. And that the conditions left behind after an empire collapses are particularly tough to re-consolidate under another empire.

Presumably OP believes that software engineers went to far in the direction of “started worshipping the machine god and loudly claiming their whole field is solved.” Certainly within software engineering there is much louder voices on that side of things. By now the voices saying that it’s useless have pretty much died out. But it seems quite reasonable to say that the wider software community didn’t “come together, assess the situation, and calmly, empathetically, and rationally act.”

One might argue that this is due to size of the software community, or commercial interests. But I do believe OP was right in his description of the response of the math community as different from the software community.

It seems very hard for OpenAI to prove that independence. Since they seem unable to exclude the possibility that their model was trained on transcripts by the two mathematicians.

It’s extra hard on Linux because the amount of ways you can reconfigure a linux machine to work around boot- or root-kits, the ways you can modify a running process, and hide what is happening from an existing process.

Really, linux is meant to give you control over the software that runs in your computer. Anti-cheat is inherently about limiting what you can run on your computer. Windows is much more aimed at allowing you opt in to hard limitations.