This will really become the problem for selling commercial access to frontier models, if it proves to be true. (I tend to believe it will ).

If the models get thousand-fold cheaper to run, than the hardware needed to do it will be something anyone interested in more than very occasional use will be able to justify. Even if it ends up not looking exactly like consumer GPU/NPU offerings today, it will land in PC and likely even SBCs soon enough.

So now the pure AI companies will have big problem, how to charge enough to pay to build and train their next model while not pricing people out of their cloud offerings in favor of a $200 expansion card - or even a $2000 expansion card - and some maybe not as good but very good free-as-in-beer models, which both academia, non-profits, and hobby groups probably can produce.

Which is why I don't companies like OpenAI and Anthropic being able to continue with an inferences as the product business-model. They are going to have to be acquired by the Alphabets and Microsoft's of the world who can eat the costs of leading edge model development and fund them with margin from other lines of business, and want to do so because they offer "better" inference as a feature in their other proprietary software tools and platform offerings.

Setting VC money on fire has never been a sustainable business-model, eventually the activity has pay for itself or it has be vertically integrated into something that does.

It is very hard to prove a negative. However when it comes to stuff like this, proving it does not happen under likely conditions.

Proving you can use infra-sound to make people more prone to certain kinds of imagination under very controlled conditions is interesting but does not explain why people often think old buildings are haunted, even when some infra-sound is present.

Once you get to 'and all the stars are aligned' territory what have shown is maybe that one guy that had some sudden psychic break onetime could be explained this way, but that does not make "people think old buildings are haunted because radiator pipes vibrate" - "Plausible" in mythbusters parlance.

Which is exactly why we need to create some kind of 'concerned citizen' or 'public interest speculative' type of standing.

Obviously it can't and should not be just anyone can sue the government or challenge a law for any reason. That would making legislating anything successfully extremely difficult because anyone with with very different real reasons can come a long an challenge every law with some corner case, and it would also burden the courts excessively deliberating over issues that are not really issues.

On the flip side the can't sue anyone over warrant-less wiretapping is ridiculous. Nobody can sue because they can't show evidence they were personally spied on without due process, even when there is abundant public evidence unauthorized spying did occur, is travesty of justice and startlingly unnecessary barrier to accountability.

I don't have a great answer or real policy here, just the vague idea that we need to lower the barrier that is legal standing to something were you don't need to show a personal harm, but also not throw the doors to the court room open entirely.

Not at all. We are just smart enough to recognize that contemporary Democrats have more in common with the old Soviet Union than modern Russia.

It is still better dead than red, except red happens to be blue now, and we have to adapt to our so called neighbors representing a more clear and present danger to our nationals welfare than an evil empire on the other side of the world ever was...

The real problem is 401ks and more specifically 401ks using mutual funds as their primary assets.

The reality the public should have a lot of influence in terms of share holder action, given how much of their capital represents the ownership. However thanks Wall Street engineering a system where a large portion of the investment comes from people who will not own the shares directly, they escape public accountability.

Instead the shares and the votes are controlled by Wall Street insiders who can be relied upon to either not vote or vote as requested by the board members. To an extent that is how it was working.

Lately an even more frightening pattern has developed, where radical leftist elements in the banking cartel have captured big fund management firms and are leveraging the voting rights that come with holding the shares against the interests of the actual investors! Literally using your own retirement savings to ESG you right of a job!

There really isn't a good fix, but at the very least fund managers need to be prohibited from exercising voting rights attached shares held in their funds. We should also explore ways to encourage broader ownership by the public of publicly traded companies over institutional ownership. Perhaps there should be tax incentives, ie a lower capital gains rate on appreciation for voting stock held in an organization that reports less than 20% of its book value as shares in other organizations at both the time of purchase and time of sale. - yes there would need to be additional rules to prevent gaming of that.

what if it isn't a scam.

What if the scam has in fact been as many suggest decades of de-industrialization of American through policy bought and paid for by small 'international-class' that lined their pockets at the expense of opportunity and security of their fellow citizens.

What if these actions are being taking by those who correctly recognize the problem, but either don't appreciate the complexity or cost of the required solutions?

What if they do appreciate the cost and complexity but this is the best they can muster faced with massive opposition by a mixture of actual traitors, and an army of useful idiots they raised?

What if you're a member of that army, zephvark?

So Anthropic builds this super power model that is to dangerous to let just anyone near, but does not bother putting any real fine grained authorization controls on it. They just set access to to be control by some kinda tech-preview group membership they use for other things as well, and hope none of those people go sniffing around and find it?

  WTF how much effort could have been in the context of a project like Mythos development to create a few more IAM objects? These guys want the world to take them seriously about their advice on risk, controls, information security etc but don't have basic internal process control around discretionary access?

I would have to counter that argument that using a FOSS project without contributing for a for profit activity isn't great but the people who are behind project always knew that was a possibility, depending on what license they chose.

However just being a user, especially as a corporate entity, means more exposure from the project. Even if you don't publish the fact you use it, you end up with employees who know that might recommend it to others, move on use it elsewhere, contribute themselves, provide useful bug reports and test data etc..

This does none of those things, this is purely parasitic it borrows all the ideas and robs the original project of mind share. Now you *could* argue this is true of FOSS clones, of commercial applications... and I think it would apply to FOSS that isn't also free-as-in-beer case; though I struggle to come up with an example of FOSS that is both clone-ware and inst FAIB.

Going back to the invasion of Georgia the US government had, overtly (I don't think the IC state department people every really pivoted), to a mostly adversarial relationship with the Russian state.

In the 20 years hence Russia has at various times sought to rehabilitate its image on the world stage thru negotiations like New-Start. The our government would have been able to leverage such events if convincing evidence existed. They did not, under Obama of all people they did not. Additionally Russia had no interests served by further antagonizing the US government thru most of that period, far better for them if we stayed out of their conflicts with the former Soviet bloc.

Conclusion - Russia was not then, though certainly could be now and with reason to be, responsible. However let's be real about that too. Why would Russia risk intelligence assets targeting research scientists. They are being rapidly depleted by our providing Ukraine with mostly end-stage cold war era weapons systems. Even if they (Russia) have more advanced weapons (hyper sonic missiles) at a battle field ready stage, they have no capacity to produce them in battlefield altering quantities. Try to delay us from developing tech we'd be unlike to field for a decade makes little sense, they'd be better served trying to compromise the production teams at our defense contractors and arranging some industrial accidents to shut down production than trying to target the research side of the house.

On the hand China.. who is very much considering a future war and has the resources to fight it; is looking at a situation where they could face next-gen weapons by the time they decide to pull the trigger. They have every reason to want to make sure they fighting the us with 21st century weapons while we are forced to rely on late 20th century relics, in the same way Russia is forced to rely on midcentury relics to fight Ukraine and our late 20th century stocks.

Russia for a foreign policy and security standpoint is a distraction! US policy makers need to internalize they are has beens who don't for the most part matter. They have firecrackers that make it impossible to ignore them entirely but by and large everything fling their direction is stuff we'd be better served to hang onto for an eventual Pacific conflict. Obviously with the exception of those arms with an expiration date.

The biggest problem with education in the US is the educators. There is no getting past that. Record spending per pupil, lots and lots of restorative this restorative that bias opportunities toward demographics that the left claims were historically disadvantage. The one consistent outcome is poorer outcomes no matter what color or economic stratum a kid belongs too.

Modern classroom theory and pedagogical theory are obviously broken. This is the only possible valid conclusion anyone rationally looking at what is happening could reach. However we have these unions which ensure that nobody does anything but fail upward in the public education sector. Basically the only real skill required to succeed (career) as educator is the ability to repeat the party line, and not molest the children (the latter being increasingly seen as optional in the eyes of union leaders as well).

there is no deep state and elections have consequences..

Yeah sure.

but you're not dealing with where those orders come from or setting the stage when you pose that question.

There is underlying assumption that if the order is received there is a point to it, that it could win war. If that person knew the enemy missiles were already in flight, would not be intercepted, and nobody they know personally or care about will see the sun come up tomorrow would they still turn the key and kill millions more people - just to get even..?

It really is a very different question. Now imagine you are not the guy in the bunker, you are POTUS. You know it is over. They entire lower 48 is going to be glass starting in about 8 secs over within the hour. Do you really give the order to return fire, or do think maybe just maybe you want to look toward the kingdom of heaven even if you have never held strong belief and hedge your bets and reject 'useless' killing? Because it really isnt about MAD at that point, it is back to Pascal's Wager.

Even if not specifically inventory shrink, it is still the same class for problem. It roll it up one level higher and either the delivery is successful or it isn't.

Be it first party Amazon vans or 3rd party shippers UPS/FEx/USPS, lost, damaged, and late packages occur in those lanes as well. It just an optimization problem. Either the the (presumed) increase of loss related to drone delivery is less then the offsetting cost savings, including reputation costs that translate to market share and recurring revenue, or they don't.

Either the drone delivers can be made reliable enough or they can't. Pontificating about it might be fun, getting upset about anything isnt worth it, because I don't really see much in the way of externalities here. Sure maybe something gets manufactured and never used because it is destroyed in transit. However if you deliver 10K somethings with electric drones rather than rolling petroleum powered vans, even the waste from an environmental standpoint might easily be offset for a net gain in efficiency.

Amazon's entire business is essentially logistics. They have not gotten to be the huge player they are by being wasteful. Rather they are where they are by ruthlessly optimizing everything they do for cost.

We all have cell phone cameras, tied to phones that mostly have the amazon app installed (if you are prime customer). You photo the package if the carton looks damaged, you photo the item the moment you open the carton lid if you see damage. You report the damaged item, 90% if the time time they agree to send you a new one right away and don't require you to return the damaged item.

So really other than as a customer, you item maybe being delayed a day while you wait for replacement, this is entirely a inventory shrink problem for Amazon to solve. Either increased loss of product (if there even is) costs more than traditional delivery methods or not, either they have incentive to invest is gentler product handling or they don't. Maybe it is just ML problem of learning which SKUs, and which type of SKUs should not go by drone and should instead go out on a truck, because of lower success rate..Maybe (probably) that training is happening already.

I don't believe this is accurate. It assumes that because I am dying I am going to kill you out of spite. If I am resigned to the fact that me and mine are doomed, I very well might decide that I would in fact rather humanity go on even if it is those where my enemies than to simple kill everyone.

I think actually there is quite a bit of evidence some degree of altruism is hardwired into humans, once our own genetic future isn't a consideration we are not actually given to wanton destruction of all possible competitors.

This fact actually makes 'the bomb' more usable in a tactical way. I think people who were fighting the cold war were actually keenly aware of that as well. It is preciously why they were able to find some agreement on things like anti proliferation. They understood that while these weapons remain *almost* to terrible to contemplate mostly they do not represent the 'end of the world' event sold to the public and perhaps most of the legislative bodies. Cooler heads realized nuke war would be horrible, morally unjustifiably most like but probably isnt MAD at all, probably ends up just being whoever shoots first wins...