I had the exact same experience and it was why I canceled my Claude subscription and switched to open models exclusively. Like, I have a test account, it has completely locked down permissions, and I can't get Claude to just use it to run simple integration tests. Instead it demands I set up some incredibly complex and fragile system for manually pulling tokens out of the browser before running automated tests, which is insane and utterly unworkable. I swear no one at anthropic actually uses their products, because if they did they would notice how awful the UX is in so many easily fixed cases. For example, in their harness, you can't report feedback about the UI without starting a chat session. Despite the fact that most of the harness exists outside of and independently to chat sessions. Apparently no one considered the possibility someone might have feedback about things other than chats. Or when you cancel your subscription, you are then forever locked out of your prior work chats. I already paid for those chats, I should be able to reference them going forward, what kind of scumbag thinks that's an appropriate design? Probably no one, they just don't test their own software because they are remarkably bad software developers.
You're talking past the OP commenter. Not all companies whose owners enjoy the benefits of limited liability are Limited Liability Companies (LLCs). The other commenter was discussing the general concept of a company whose owners enjoy limited liability, meaning they cannot be held responsible for the company's debts or other financial obligations beyond the value of their ownership stake, i.e. their _liability_ for the company's obligations is _limited_. There are many corporate forms that provide this benefit to their owners, including common US corporate structures like LLCs, C Corps, S Corps, and Limited Liability Partnerships (LLPs). It's important to note that the terminology varies across jurisdictions, so the specific structure we call an LLC might be called something else entirely or might not exist at all.
Which is precisely why the OP commenter framed their argument in terms of the general concept of limited liability. The argument applies to any company structured with joint ownership and limited liability, whether we happen to label that corporate structure C Corp, Private Limited Company (Ltd.), Gesellschaft mit beschränkter Haftung (GmbH), or even Limited Liability Company (LLC).
No I'm not. The OP clearly doesn't understand the basic concepts. "Because companies going bankrupt have no liability for debts" is misunderstanding what "limited liability" means.
>Which is precisely why the OP commenter framed their argument in terms of the general concept of limited liability.
That's not the general concept of limited liability. They are simply mistaken. They just parroted Reddit and BlueSky takes.
There is no model of the market that can remain stably accurate because the market will inevitably incorporate the insights of any model that is accurate until those insights are no longer accurate
Also known as the trading model's shelf-life. Based on what I have heard and learned, the average active trading model has a useful life of around 18 months. After that the rest of the market has adjusted and its edge is gone.
If rumours are to be believed, several "slow" hedge funds have models that remain useful and profitable for 5 years or more. Then again, those models are not used to conduct exchange trades but rely more on aspects of fundamental analysis.
Rather curiously some of the largest banks and tech-heavy asset managers tend to sponsor meetups and events like PyData - every now and then doing a slot on some of their open-sourced stuff. You might see a talk on 5-year-old trading model internals with large chunks of it opened up, or they could present some of the internal UI or visualisation libraries. The latter tend to be things that they no longer actively develop (they're "ready") but maintain for their ongoing persistent needs.
Of course anything that actually brings them money and/or gives an edge is not even discussed.
Honestly: People think that "trading models" (whatever this should be) needs to be somewhat "supersophisticated" and "extreme driven by whatever complex math" - the brutal truth, esp. for smaller trading shops is much more simpler:
- Standard approaches like trend following or mean reversion are working very well since decades. (I can speak only for the last ~25 years)
Complexity in trading is not about "building that one specific niche-super-strategy", but more about putting all the ideas in a reproduceable approach/process to repeat it over and over again to grow the money.
in order for your model to accomplish that, you would get very rich.
there will also likely always be more. in the limit in order to get an edge your model would start to infer insider information. for example, it's common knowledge by now that satellite imagery is used to measure car numbers in parking lots, that's a proxy for insider information.
so it's not even so much the model as it is the data.
even being able to forecast weather better than publicly available methods can be leveraged to gain a significant edge.
Just having a model that's predictably biased the wrong way when other models have already priced in the same things is pretty good. I mean, you have to look at where your model has no edge and figure out if it does have edges that the other ones are missing.
Coming at this from baseball, where the art is really knowing when not to take a game based on the line; usually if what you predict lines up too closely with Vegas, your edge is gone. (Hand-written A-life/evolutionary algorithm model of competing baseball equations that I've been refining for years, picks lately filtered through AI to isolate which ones fall into bands that are worth risking money on).
Damn, imagine you're the tribe's only successful hunter. You're the only person that's ever brought a deer back to the fire. You're even really good at it.
But the deer...sometimes they just don't exist. And you checked very carefully. No prints, no rubbings, no hair, no trail. Nothing whatsoever. You even left a bit of hair under a rock one time and the next time you checked it was gone. The only evidence of the existence of deer is the pelt you wear and the memory of last night's dinner. Something happens to the deer in the woods.
Some weeks, you bring back a deer a day. The tribe sings song in your honor. But every time this happens, there are those who grow quite suspicious when you all of a sudden "haven't seen a deer in weeks" because what do you mean you haven't seen a deer you said there were hundreds of them yesterday.
You can try to bring them with you to see, but most aren't interested. The few you bring happen to come when the deer are plenty. You can't convince them to come when the deer have gone because they just went yesterday and saw the foot-print highways and scarred trees; they know trees don't heal overnight.
We also have visual navigation systems that can use satellite imagery to produce guidance packages. They will become increasingly less useful the longer the imagery goes without updates after a Kessler syndrome type event, though. Another strategy, which I believe the Ukrainians have been experimenting with, is scattering small radio beacons throughout the battlespace to provide a network of navigation aids.
In a true Kessler scenario where LEO is denied for an extended period of time, I expect a variation of that strategy is how we would replace GPS for most civilian use cases. Likely by using cell towers to triangulate, possibly enhanced with additional transmitters to improve accuracy and coverage. Pure cell tower triangulation would have much worse coverage than GPS, obviously especially in rural areas where cell service is often already poor, but also anywhere that getting a signal from multiple towers isn't easy or possible. Cell towers are never distant enough to use the relativity-based timing tricks GPS relies on, so multiple signals would be required to triangulate. We'd likely end up with substantially less precise positioning, maybe something more on the order of the 10-20m of error we used to see with GPS before the US stopped limiting civilian access to the deaccuratized version. Cell tower triangulation also obviously doesn't work for aviation or oceanic shipping, where trident-like inertial+stellar nav would probably be the only workable option.
I assume in that situation they would be targeted specifically with ASATs at the beginning of the conflict, with the denial of LEO then preventing us from launching replacement sats.
This happened after I stopped working for the DOD, but I am 99% sure the program in question during this incident was something I worked on. It was effectively an anomaly detection system for identifying any weird behaviors detectable in WAMI (wide area motion imagery, which is high resolution, low framerate drone video data that covers a large geographic area). As an example from the unclassified sample project they asked us to do as part of the bidding process, our code flagged a car doing donuts in a parking lot (the sample data was just from a random US city on a random day because this was an unclassified open bidding process). The system was not and never was intended to be a "terrorist" detection system. The idea was always to flag potentially interesting events for a human analyst to follow up on. The disconnect it took for someone to blindly trust the program as a target acquisition system boggles my mind. Someone doing donuts in a parking lot should not be automatically targeted with hellfire missiles, yet in practice that is how they were using the tool. I still lose sleep over it.
> Your work will be used against innocent civilians in far away lands.
For now. I think this administration will absolutely have no compunction about using this against innocent civilians in THIS land as soon as they feel the need to to retain power.
Not sure about the downvotes, this is correct even if not pleasant to read (people need to be heroes of their little bullshit life story and all that).
It may be an afterthought for US citizens living in US, but its a deadly weapon pointing to 95% of the mankind and people wielding it are often shit to be polite. We the rest do care.
I mean, this site is full of people that develop the weapons used to commit atrocities and genocides (both physical and digital ones) so I am not surprised about the downvotes..
that's why we outsource to other countries our software development, so you don't get blood on your hands violating the... essence of the constitution. i'm missing a word here.
hiring foreign data scientists and web developers.
The answer is not to avoid engaging with a corrupt institution. US is a democracy: vote, be active politically, go into organizations that need change.
So you are suggesting that you can change the US military industrial complex by.. working for the NSA or Lockheed Martin ?
Also it is quite questionable how much of a democracy the US is nowadays.
Gerrymandering aside, the populace does not get to choose the presidential candidates. And the whole two party thing ? Or the fact that to have a chance you need a super PAC with hundreds of millions ?
Yes, that's what I'm suggesting. If you feel that the problem is higher up or lower down in the institutional hierarchy, then go higher up or lower down. That's the civil approach, as I understand it.
Apart from the fact the if you stay true to your morals you will be incapable of rising to a position of power in such corrupt institutions, the degree of institutionalisation necessary to do so will definitely erode your ethics, turning you into another cog.
In my opinion, you underestimate the power of an individual. I'm not saying you have to be Gandhi, but he was one person. A problem with your assessment, is that if you're right, or, if everyone believes you're right, you'll have to wait for some external force to come in and change things, like a calamity or desperate revolt. By the time USSR broke apart, bread riots were happening, and, tragically, 30 years later we still have the same authoritarian, corrupt regime (in all but name) wrecking havoc at home and abroad. That's what a people get for believing in the story of disempowerment. That's my take, at least.
I agree in principle, but also agree with gp that in practice the things one has to do to rise to a position of power in corrupt institutions make it virtually impossible to be a credible agent of change, especially private/secretive institutions like military, intelligence, military-industrial corps. I guess politicians are the one role that is supposed to be able to operate without those institutional constraints but of course, (in the US) given corporate capture of political speech, political hyper-polarization, etc, in most cases they are just as or more constrained. It is a hard problem.
The Brits were bankrupting themselves with war spending, which culminated in WW2. After its conclusion, they started liberating every territory that wasn't directly profitable, including some that did not even want to be liberated. WW2 ended in 1945, India was liberated in 1947 alongside numerous other territories. Without WW2, India would likely have remained a British territory well past 1947. Beyond that, Gandhi wasn't one man - but the leader of a movement of millions.
As for Russia, I think your perception there is also driven by 'messaging'. There's a million videos of people streaming walking tours and the like in Russia. Here [1] is one done 4 days ago. That's a country that in the 90s was on the verge of collapse and was extremely dangerous. You can find similar stuff on e.g. Twitch because there's this somewhat odd subculture of delivery drivers, taxi drivers, and others streaming their days. Perhaps in part out of bemusement from this Western perspective paired against reality.
Snowden and other whistleblowers, of whom there have been many (many of which paid arguably a dearer price than Snowden did), made it plain obvious to everyone that US intelligence services are doing illegal things abroad and at home. Before that, only tinfoil-hat-types thought so: now everyone knows it. They moved the Overton window a tremendous amount.
I don't think actual change comes from people lobbying for some bill or whatever. It comes from a far more imperceptible shift in the zeitgeist of society, that is then eventually formally reflected in some way or another. For instance the Civil Rights Movement only succeeded because society had already largely shifted, before it had even begun, in a way that made it viable. A Civil Rights Movement in Saudi Arabia isn't going to succeed, because it needs the widespread desire for such to already exist.
And in this regard I think Snowden had a bigger influence on society than anybody or anything in my living memory. Prior to him there was a belief by many that government conspiracies couldn't happen because somehow if there were enough people involved, it'd show up on News at 9 or whatever. He made it clear that such conspiracies, far beyond the bounds of what even the absurd conspiracy theorist might propose, can and do exist. The NSA has friggin spies in World of Warcraft!
That sort of skepticism played a key role in fueling an anti-establishment mindset. And that mindset is what fueled the rise of everything from Trump to the DSA. I'm certainly not claiming those are good outcomes, but merely part of an increasingly visible shift away from the old establishment. And I think this shift will be, in the longrun, good for America. And I think in hindsight, it will be clear that a lot of this change started happening with Snowden.
I'm probably too focused on the Afghanistan war but I don't think that voting differently would have extended or shortened the war.
Both republicans and democrats wanted the war in Afghanistan back when 9/11 happened and they needed to get revenge on the scapegoat. Meanwhile the end of the war was initiated by Trump, then slowly concluded by Joe Biden.
So both the start and the end of the war had bipartisan support. Nothing against democracy but war doesn't follow the election cycle. It might start in one though.
Whenever you introduce (or grow) the presence of computer automation you need to consider the human / social interactions. You are responsible for the whole system, not just your software.
It is well understood that human diligence falls dramatically in the face of automation that is perceived to be highly reliable.
There is a tendency for software people to say “it has a 5% error rate, so make sure to carefully check every time”. Even though we know that humans simply won’t check carefully the results of a 95% accurate system.
> There is a tendency for software people to say “it has a 5% error rate, so make sure to carefully check every time”. Even though we know that humans simply won’t check carefully the results of a 95% accurate system.
This should really be framed differently during training. "Assume you have 100 neighbors in your area/house. We expect you to kill 5 innocent neighbors each time you use the system, and don't actually think for yourself. (After using it once every day for two weeks, you will have killed half of your neighbors.)"
Suddenly 5% doesn't sound that small/95% not as good to the laymen.
IMO you should increase the error rate synthetically, adding tests. So you get 10 targets to review, but 4 of them are known to be bad and the system is both checking the user, and also increasing the perceived error rate to encourage diligence.
Brilliant! This is the logic behind the birthday problem where you only need 45 people present to have greater than 90% chance that 2 people share a birthday.
Even rare events become very likely when you repeat it. A 5% wrong chance means half are dead after 10 tries
> Even rare events become very likely when you repeat it.
True, but in my opinion 5 innocent dead out of hundred is not remotely small. For any serious automated decision anything larger than per million is completely unacceptable. (For reference, among Germans the amount of Jews killed by Nationalsocialists, is "only" 2.3‰ [160k vs. 69M] I don't think anyone considers that to be small.)
> A 5% wrong chance means half are dead after 10 tries
Btw., after using it for a year every day, the US population would be 2.52 people. And I think the US military plans to use their expensive systems for much more than a year.
A computer can never be held accountable, therefore a computer must never make a management decision.
We have CEOs designing AI systems to replace them (or at least their workload). The DoD has done nothing to provide accountability for this war crime. It's a scapegoat to move on from, not a problem or justice.
I have generally been bullish on data centers independent of how AI demand/load evolves. People will find a way to use that compute, even if it isn't the precise use we expect. As scale increases and the price per computation comes down, we will be able to brute force solutions to problems that otherwise would be intractable or prohibitively expensive to solve. And there are effectively an infinite set of those problems.
This tracks the evolution of how we use cloud computing and GPUs over the last 20 years. Cloud computing was originally just about saving companies from needing to maintain their own server racks, but has unlocked previously unserviced demand by allowing people to build an app or service and scale it to meet rapidly rising use without needing to invest a ton upfront in hardware. Suddenly a hobbyist could spin something up in their spare time that previously required many thousands of dollars of investment. Or when someone wants to run a single large scale computation, they can do it without needing to waste capital on maintaining idling servers to meet an occasional demand spike, like when a company I used to work at moved from running atmospheric calculations on a server in a closet to the cloud and were able to achieve double digit accuracy increases with the increased scale, while spending less overall on batch computing jobs.
GPUs, as the name implies, were created for graphics, and primarily for gaming graphics, but then more or less accidentally ended up enabling the present AI boom, which depends on a scale of computation that would have been impossible with older CPU architectures. Maybe someone at some point predicted this, but I think for the vast majority of people, it was extremely surprising that a niche gaming product would enable an industrial revolution level technological leap forward.
LLMs are just one way that increased compute scale unlocks seemingly magical results, but they are far from the only example and I have no doubt that there are many unknown examples remaining to be discovered yet.
I think you're conflating the impact of a small bubble on a narrow market space with our current situation. The current bubble is larger than anything ever seen before, and is now encompassing nearly the entire economy.
New datacenter builds are so far along the curve of diminishing returns it's absurd. No one is going to want to pay to run a datacenter that costs 10x as much to run for the same compute.
And the problem is with all these "freed" resources, the entire pipeline will be affected. No one will want to buy any new silicon if they can buy a B200 for $1000. We could potentially see a decade or more of stagnation in the chip sector, or even significant regressions in capabilities as foundries are shut down due to lack of demand.
The impact of what is coming scares me to my core. I don't think we're going to bounce back from this any time soon.
I think actually it's quite the opposite and you're the one conflating the impacts on a narrow market space with the larger economy. Right now, AI dominates growth in the stock market and demand for chips, but it is by no means encompassing the whole economy. Neither the stock market nor the chip sector are the entire economy. My grocery store isn't going to go out of business if Anthropic does.
A lot of investors may lose money as a result of the bubble bursting, but that does not mean the underlying asset will be forever worthless, just that it didn't provide sufficient returns sufficiently quickly to justify the upfront investment for the investors that funded it, at the time they made that investment. A different investor who could afford to ride out a period of reduced demand might have an entirely different experience.
Imagine you take out a five year loan to buy a truck to deliver packages, but then for the first two years you operate it, gas prices are elevated so you struggle to make the payments on your loan and end up not making as much money as you had hoped to originally, perhaps even to the point you need to declare bankruptcy and sell of the vehicle. But if not and then gas prices drop back down and demand shoots up for the last three years of the loan, you could then make up the difference. If the first two years drive you into bankruptcy, that is difficult for you, but amortized over the entire five year loan period, the truck may actually have been a profitable investment for someone who could have afforded to ride out the first two years. And just because you go bankrupt doesn't mean the truck stops being a valuable asset, it's just that you don't end up benefiting personally from that value because your timing was bad. From a macro perspective, the overall economy doesn't suffer, except to the extent that it might have been more efficient to invest the capital that went into procuring the truck elsewhere during those first two years. But only possibly and only on the margins, because the truck remains a profitable investment over the course of its entire lifetime.
Don't confuse the success or failure of individual investors or businesses with the success or failure of the overall economy. Current data center build outs premised on fanciful projections of demand for LLMs may end up not being profitable in the short term while still being profitable over the entire productive lifetime of the asset, if sufficient demand is found elsewhere or if demand for LLMs picks up later. Similarly, I anticipate at worst we will see chip prices plateau for a while if there is a pullback in LLM demand, but we won't see them fall and they will continue to rise over the longer term as more demand is generated elsewhere.
As one small example: we have barely begun to scratch the surface of what we can achieve with robotics. Think about the demand for video processing if you have tens of thousands of robots stocking shelves in supermarkets generating video all day long. On board processing will of course be the obvious primary demand for chips, which doesn't benefit data centers, but central processing of video to extract useful data from the entire fleet will generate demand for data centers. As will large scale training jobs. Now multiply that thinking across the entire scope of industries where robotics may be useful for replacing human labor, and you're talking about an extremely significant amount of valuable computational work.
Yeah but general purpose data centers that allow hobbyists to spin up an app or a service easily is exactly that: general purpose. AI data centers are built for the very specific kind of math that LLM's require. Even the GPU's can't even be used for something like cloud gaming, which isn't very popular anyways.
That math is very much general purpose. It's just linear algebra under the hood, and linear algebra is a wildly useful toolkit with an unbounded set of potential applications, including many applications other than LLMs. That's how something developed for gaming evolved into something used for text processing and generation. For example, we have barely begun to scratch the surface of what we can achieve with robotics, and I have no doubt that further advances in robotics will generate an enormous amount of computer vision work that GPUs are extremely well designed to handle.
Bullshit. Even if you don't think LLMs alone are enabling that level of advancement, GPUs in general very much are. They are the critical enabling technology underlying self-driving cars, autonomous drones, camera-based robotics controllers, faster/better drug discovery, faster/better modeling protein structure and design, ML weather forecasting, materials discovery, and many other applications.
Fair, but my point is that a data center is not pets.com. They are a much more flexible asset, and are even more flexible than AI itself. Pets.com failed because the physical delivery infrastructure and consumer buying habits that support Chewy today did not yet exist and took years to create. Data centers can be repurposed from supporting the current generation of LLM-based AI models to an infinite variety of use cases. Pets.com could only ever hope to deliver pet food.
And Amazon.com could only ever hope to deliver physical books?
Anyway, it's clear an AI data center has utility for crunching AI inference, and that there is and will continue to be demand for AI inference. The trillion dollar question is whether you can make money doing this, and so far the answer is no.
You are not disproving but rather reinforcing my point. Amazon survived the dotcom bubble precisely because they could hope to do more than deliver physical books. It was a more flexible, generally applicable business model than pets.com, and they started out targeting a market more susceptible to ecommerce conversion than pet food delivery proved to be. Similarly, a company selling a product that can only be used for the limited use cases we have found for chatbot-style text generating LLMs will be more likely to fail than a company selling something that is more generally useful. Data centers are more generally useful than LLMs, and the companies building them are more likely to survive than those whose fortunes are pinned entirely to unrealistically high expectations for replacing white collar workers with LLMs.
Data centers aren't just physical buildings though - the vast bulk of data centre costs is the racks of servers in them, and those are being bought and installed on credit at very high costs.
Its not about "is a computer thoeretically useful" its about if it's going to deliver value to cover the cost of installing it in the first place over it's expected lifetime.
You aren't substituting generic compute: you're substituting highly optimized GPU dense server space. The application is more limited then it looks.
Yes that’s what I meant thanks for clarifying. I think local models are the future but I don’t see anyone serious arguing it. If it were true these city sized data centers will be graveyards in 5-10years
I'm shocked Netflix hasn't just settled to keep this quiet, because the entire situation makes them look not only like an abusive, discriminatory employer but even worse, horrifically incompetent at being abusive and discriminatory. Plenty of companies treat their employees like shit, that's not news. But any competently managed company knows how to do so without incurring massive legal liability like this, and it speaks to a remarkable lack of professionalism and basic competence at the leadership level and within the legal and HR teams.
If the accusations are true as reported, their in-house counsel fucked up terribly by assuming all ketamine use is illegal and recreational without bothering to conduct the very cursory investigation required to determine that it is in fact a perfectly legitimate option for treating treatment-resistant depression. That single error drove them to fire an otherwise exemplary employee for an obviously illegal reason in a manner that screwed him out of significant compensation including severance.
This was an obvious, unforced error by their legal team. Competent lawyers and competent corporate leadership would recognize that the relatively small amount of money they may save by litigating this issue is not worth the damage it is very predictably inflicting on their corporate culture and public brand. The lawyer who publicly admitted that ketamine factored into their decision should be immediately fired because he is demonstrably a terrible fucking lawyer, and he seems to be representative of the quality of legal advice Netflix is receiving, because any attorney with half a brain would have immediately recognized both the underlying legal error that incurred liability for the company and the business context that makes immediate heavily NDA'd settlement the optimal strategy for defending, even were the original legal error not so egregious.
My main takeaway from this situation is that Netflix is managed by remarkably incompetent leaders who are advised by even more remarkably incompetent lawyers.