betterunix2

1) Years. I would say reducing carbon emissions is a higher priority for humanity than building more DCs, so building the nuclear plant should be a prerequisite for building the DC.

2) Dry cask storage. It works and it is extremely safe, you can leave the casks on site or bury them wherever you want. The most significant concern with nuclear waste is not the radiation, it is the possibility of radioactive materials leaking into the environment and that is exactly the problem dry casks solve.

2.1) Reprocessing, you can reduce the volume of nuclear waste by recycling fissile uranium and plutonium from spent fuel into new fuel pellets. The remaining waste is typically embedded in glass, which can be placed in dry cask storage (see above).

Nuclear waste is not the problem people think it is. We know how to deal with it and we have been dealing with it for a long time now. For comparison, solar panels do break down over time and solar panel waste includes several toxic heavy metals that can leak into the local environment. Nobody objects to solar power by pointing to the waste stream (let alone the waste created by all those batteries that you need to make solar power useful for industrial operations like DCs).

Most likely scenario is a janitor who installed an app (I'm guessing it was called LegitWorldCupWatch420696969) on his phone (which might be connecting to the wifi network from whatever locker he leaves it in while he's on the clock, the WH is not a giant SCIF). Or any number of other mundane, easy to overlook employees who work outside of the actual secured areas of the building.

Right, but the point is not that it is possible, it is that there are other, more important issues here. Let's say we regulate the utilities so that residential customers will not be impacted by industrial users (or by DC operators specifically). The result would be higher utility bills for those industrial users, which would encourage the use of things like turbine generators -- which I would say is a negative outcome we should avoid.

It needs to be carefully thought out (Congress does not have the best record on this -- see e.g. the ban on onion futures). As I suggested elsewhere, a better idea would be to require DCs to pay the local community in proportion to the electricity consumed by the DC, and an even larger payment to the local community if the DC does not get its power from the grid (to discourage less efficient/more polluting power generation). Yes, rates would be higher, but people would be compensated with cash and it would be a big incentive for industrial users to be more energy efficient and a small nudge for residential users.

I would not call STUXNET proof that this can be done "easily" -- that was joint work by the NSA and Israeli intelligence, who are among the best hackers in the world, and it used 4 separate 0days simultaneously.

While you certainly make a good point about the difficulties that come with securing the laptops and other equipment, the military has the same problem. Most of the users of classified military systems are not themselves experts in computer security. The necessary training is not a graduate level course in security; it is training people to always match cable colors to port colors, to only use laptops provisioned by the organization, etc. The military disables, sometimes by physical means, USB ports on laptops that are used on classified networks to close one of the key security holes Stuxnet took advantage of.

Sure, it is challenging, but this is not impossible to solve. Utilities would have to be forced to do it and they would need time to overhaul their operations, but they certainly can do it, and there is an argument that the NSA's IA division should help (as the electrical grid is a critical for national security). It can be done, we just have to be willing to do it (and pay the cost of doing it).

The military's approach to that problem is to use color-coded cables and ports, and we already use things like 802.1x to prevent unauthorized devices from being attached to corporate networks. This is not a novel or unsolvable problem.

OK, let's take that example. It is true -- landlords are responsible for raising rents, and they do so in response to demand for housing. In some places, landlords are not allowed to arbitrarily raise rents -- for example, by limiting rent increases to follow the CPI. In theory you could regulate utility rates to insulate communities from the impact of DCs and other industrial operations.

Like I said, it is a nitpick and there are bigger problems here.

Indeed and frankly we are way behind schedule on building nuclear plants, mostly because of irrational fears about nuclear power safety.

> Any new datacenters being built cannot raise utility rates

DCs do not raise utility rates, utility companies raise rates in response to demand, but aside from that nitpick there is a bigger problem here. If the DC is not connected to the grid then the DC is going to generate power on its own, using inefficient (due to the small scale) and polluting systems (e.g. turbine generators).

If you actually want to address the utility bill impacts of DCs, you should require them to pay the local community. We have a similar arrangement for oil drilling -- to avoid a situation where companies pile onto oil fields and put derricks up on every property, we instead have the companies pay people who own the land above the oil being extracted and that is why Beverly Hills is such a wealthy neighborhood. Give money to the local community in proportion to the electricity consumed by the data center.

Frankly we should require DCs to pay even more when they are not attached to the grid to discourage pollution.

> Datacenters must not touch our waterways at all

...at all? Not even with a connection to the local water utility? That would mean there cannot even be a toilet in the building.

Data centers are industrial operations and should be treated as such -- I do not know of any other industry where this kind of absolute isolation from the local community is being called for, and there are plenty of concerns about industrial water use beyond data centers.

The farm might not have made food and it might not have made food for human consumption, and even if it did make food it might not be the kind we need to survive. Remember that there are plenty of farms that exist to produce ethanol, HFCS, or inputs for beer/whiskey/etc. makers (which we do not actually need -- drinking is fun but hardly necessary for survival).

...so building that nuclear reactor would be a good thing since it will emit far less carbon than those natural gas generators (not to mention all the other pollution that nuclear plants do not emit). Hook the DC up to the grid and expand nuclear power to simultaneously accommodate the load and decommission fossil fuel plants. Wins all around.

Shockingly...yes, and in some cases that is being done intentionally e.g. solar/battery systems require communication within the grid and the Internet is being used as the network.

What makes this particularly embarrassing is that power companies literally run thousands of miles of cables and could easily build a private network instead of using the Internet. The regulations should be amended to require it.

Have you ever graded tests for an entire class? I did so for two years when I was a PhD student and had to work as a TA. It does not matter if you are an expert in the subject because the problem has nothing to do with expertise, it has to do with how our brains work. The worst case is when most of the class answers the question correctly, and you don't encounter a mistake until you have finished grading most of the tests -- your brain really wants to just keep repeating the same motion (e.g. writing a check mark) and even if you spot the error you might accidentally mark it correct anyway.

Really though, this is missing the point. Yes, grading tests and homework is different from sorting mail, but both are tedious tasks that involve a lot of repetitive actions -- hard for the human brain, but easy for software. Software does not get tired or distracted by other responsibilities (or non-responsibilities). We have already been using software to grade tests for a long time now, but mostly for multiple choice questions; AI will expand the scope of that.

> I would rather we have plain language legislation

I doubt that you would be happy with the outcome, certainly not if you had to actual deal with such a legal system. Plain language laws sound great in theory, but in practice the result is a rigid system that is easily manipulated by bad actors and which becomes oppressive to honest people.

"No, I am not marketing a tobacco product, I am marketing a product that produces nicotine from other chemicals, then vaporizes it to be inhaled!"

"If I had not done it, my children would have been seriously injured, are you really going to imprison me for protecting them?"

In general, whenever a society tries to apply its legal code, it will begin to develop judicial law by necessity -- at some point you will have to deal with borderline cases where it is not clear if the law was broken, no matter how clear the language of the law might appear. Here is an ancient example taken from the bible: the death penalty shall only be imposed on the basis of testimony from two witnesses. OK, seems pretty clear, but we know from the Talmud that questions arose when the law was applied -- who counts as a witness, what counts as testimony from a witness, etc.

Also keep in mind that no matter how plainly laws are written down, over time you will wind up with a large number of laws as society deals with changing conditions and unforeseen circumstances. Inevitably you wind up with contradictory laws and you do need some process for resolving the contradiction and determining which law should be applied.

Read about the "rule of law," one of the foundational concepts of the US and basically any democracy, as well as most successful monarchies and other forms of government.

Yes, we wait for the courts or the legislature rather than resorting to mob justice. There are good reasons for it. The most obvious is to ensure that the corrupt politicians do not use the power of the state to arbitrarily oppress and persecute people. Another is to ensure that there is no doubt about guilt when we convict someone, and to avoid arguments about the penalties being arbitrary and unfair.

Really though, establishing courts to deal with accusations of people breaking the law is something most societies converge on at some point. "Joe was only ordered to spend 5 years working the fields, so why should Jim be executed for doing the same thing????" amounts to a rudimentary form of judicial law -- pointing to a precedent from a previous case to argue the current case.

Bernie Madoff's operation was also subject to a regulator and was properly registered as a broker-dealer; that did not somehow make it a regulated operation, at least not in the sense of actually complying with the regulations. Boiler rooms are also subject to a regulator (the SEC) and obviously are not regulated brokerages.

Really it sounds like this is a disagreement about what it means for a market to be "regulated" -- I draw the line at compliance, you seem to draw the line elsewhere.

The actual question to ask is if the agent makes more mistakes than a human being. Grading homework involves reading over similar/identical answers to the same questions over and over, and it is easy for human beings to make mistakes while doing that.

(That is basically the argument for OCR in the postal system -- it does make mistakes but it makes fewer mistakes than human beings.)

They offer contracts that are almost certainly outside the scope of regulation, an appeals court has already ruled that some of their contracts are indeed subject to state laws and not CFTC regulation, and they are continuing to push the envelope. How is it wrong to call them an unregulated market under the circumstances?

How is that at all clear? They are likely subject to CFTC regulations but there are a lot of questions about whether or not they are in compliance.

That is the federal government asserting that Kalshi is subject to federal regulation by the CFTC and not state regulation. It is not an assertion that Kalshi is actually in compliance with futures market regulations, nor is it an assertion that the contracts offered by Kalshi are actually legal (there are explicitly illegal futures contracts).

If Kalshi is not a gambling operation then it is an unregulated futures market, which is still illegal. There are perverse incentives in futures markets too, which is why they are regulated.

Also speculators serve an important purpose: hedging strategies require speculators on the other side. It is not just a "certain clientele," there are people who run profitable businesses on speculation (it amounts to collecting insurance premiums from people who want hedges).

If it is not gambling then it amounts to an unregulated futures market, which is also illegal. There is no way around this for Kalshi/Polymarket, they have broken the law and the only question is which law was broken.

There are other ways to use AI in a school setting. Grading homework is a tedious and time-consuming task that would likely be simple for an AI to do.

The main problem with submarines is that it is not possible to communicate with them while they are submerged. In particular, the submarine can receive but not transmit ELF signals and nothing else really penetrates sea water, so if you want to immediately know the status of a submarine you would have to transmit an order to surface and report in (doing so is rare because it is very easy to detect the submarine when it surfaces). If a country relies exclusively on submarines it will be vulnerable to a surprise "disarming attack" -- the submarines could all be destroyed while at sea, and then the country could be attacked with a first strike and would have no chance to retaliate. Exclusively relying on the submarine force will also allow the other side to devote more of its resources to anti-submarine warfare.

One reason to maintain an ICBM force is that the ICBM silos protect the missiles against anything less than a direct nuclear strike; by placing the silos far from urban areas you force the opponent to expend many nuclear weapons attacking individual silos, which greatly increases the cost of a first strike. If you were wondering where bomber forces fit in, maintaining both ICBM and bomber forces creates a "synchronization problem" -- both must be destroyed at the same time to prevent retaliation, and by strategically placing the air fields and missile silos a country can make a synchronized attack more difficult (the attack will be detected long before anything is destroyed, but one of the two will have more time to initiate the retaliatory strike). Bombers also force the opponent to expend significant resources on air defense, resources that could otherwise be devoted to offensive weapons. Another reason to have a bomber force is that bombers can be recalled after being ordered to strike; missiles cannot be recalled once launched and even if they could be the other side will detect the launches and immediately retaliate.

The US and USSR developed an entire theory of the nuclear triad during the cold war, and both reached the same basic conclusions about the need for all three legs.

Nuclear strike orders do not require high bandwidth. The orders contain targeting information and arming codes, all of which is pre-programmed on the missiles (the target package is specified with a "pointer" -- the computers on the missiles already have all the details, the order simply identifies which set of details should be used). For ICBM crews it is conceivable that orders could be transmitted on high-bandwidth systems, but for submarines that is not even possible (ELF bands have to be used because of the physical properties of ocean water).

The basic launch sequence is this: a missile crew receives a series of characters and both missileers write the characters down. They check against a codebook and verify that they both heard the same message and that it is a valid launch order. If the message is valid, they unlock the launch keys (from a box secured by two combination locks, each missileer knows the combination for one lock), insert the keys, enter the missile launch unlock code, enter the warhead arming code, turn the keys, and then that's that.

(There are several other steps -- go watch a video of a test launch and listen to the operators read off the checklist, there are quite a few steps. If you are wondering why such information is public, it is meant to reassure the other side that there is no possibility of unauthorized or unintentional launches -- we want our adversaries to know how we control our nuclear forces so they don't panic and preemptively attack us)

It's more that fallout shelters are basically useless even if you manage to get yourself inside before the detonation.

Basically, you would run out of water, and even if you have clean water you would still run out of food, even if you survived the strike and the fallout. We cannot actually live without major transportation hubs to deliver food etc., which by necessity are located in or near the major urban areas that would be destroyed in a nuclear war. That, more than anything else, is why the US and USSR ultimately abandoned "civil defense" and focused on "deterrence."

It is OTP, that is not a secret. The secret is not how the orders are being communicated, it is the actual meaning of the orders (e.g. the targeting pattern) and certain strategically significant details about how a launch order would be executed (e.g. if only one pair of launch keys is turned there is a timeout during which another missileer crew can cancel the launch; the length of that timeout is secret). During the cold war nuclear strategy was the subject of extensive research by both the US and the USSR, which is why both sides wound up pursuing essentially the same strategy.

(China is interesting because strictly speaking they follow a theoretically suboptimal strategy -- they keep their ICBMs in an unarmed configuration and would "absorb the blow" of a surprise nuclear first strike. The US and RF follow the theoretically optimal "launch on warning" strategy, which requires ICBMs that are always armed and ready to launch within minutes.)

I mean...sure...but if you are going to be consistent you are going to have to criticize the Allied victory in WWII. There were literally piles of dead children in cities all over fascist Europe. We do not even have exact numbers because in some cases the death toll had to be estimated by weighing the ashes found in bomb shelters.

"We do not know if our agent is deliberately changing its behavior when it detects an airgap. So we are not going to run it outside of the airgap until we figure out how to ensure that is not happening."

Well heaven forbid their iteration speed takes a hit!

I do not have much faith in what they believe about their ability to enforce a security boundary given their track record so far, not to mention their grossly inflated sense of importance (both about their work and about themselves).

Here's an alternative scenario:

I connect a computer to the Internet and then leave a program running on it, unattended, that connects to a bunch of websites and does things that "exceed my authorized access." Then I say "Oh sorry about that, I was running this really important experiment, really it's gonna change the world in ways we cannot even imagine! I'll totally be more careful about this sort of thing!" I tweak the arguments and try again, it does something else that "exceeds my authorized access," and then I say, "OK this time I am definitely putting some hardcore controls in place. I'm running it in the SELinux sandbox now! It still needs Internet access because I want to check in on my experiment from this awesome cafe down the street..."

Oh wait we know exactly how that scenario ends: I get prosecuted for CFAA violations and gross negligence, and probably several other crimes, and that is where those prison guards you mentioned enter the picture.

> Similarly models could allocate a dead zone of their weights to inter agent communication

Sure, and you know what? If we cannot know that these models are not doing such things, let them stay in the airgapped environment and let the researchers continue their experiments and research work there until they develop techniques to verify that their models work as intended without unintentionally doing harmful things. The issue here is not agents designing covert channels, it is agents doing things that are almost certainly illegal and which have the potential to cause actual harm in ways that are not at all difficult to imagine. There is a reliable, robust, and straightforward solution, which OpenAI is claiming they cannot possibly implement because it would mean...putting in some effort and accepting some inconvenience. We would not accept that excuse from anyone else; why should we accept it here?