Of course it was. They clearly decided that the benefit to the company valuation was higher than the potential downsides when announcing to the world that they committed a criminal act via negligence.
If it wasn't a marketing stunt, they would have at most quietly settled any legal matters with huggingface behind the scenes, fixed their evaluation harness so it wouldn't happen again, and avoided the potential future liability.
It sounds like your issue is with capitalism, not with renewable energy. Luckily, we have plenty of examples of aspects of society that have been deemed too important to leave purely to market forces. In fact, electricity generation is already (mostly) one of them.
As solar and batteries become even cheaper (we're already at the point where they're cheaper than anything else) we can just choose to overbuild capacity and turn down/off whatever other generation isn't necessary at any given time. Utilities already have to predict grid usage based on weather and all sorts of other factors, so it's not like this isn't something that isn't already happening.
What? Absolutely not, I love capitalism. The problem is that it is suspended when it comes to the wind farms, they are pretty much always heavily subsidized and on a socialized loss scheme at that.
And no, I have no problem with renewables either.
If your conclusion from reading what you said is this, then you did not read because I said the exact opposite.
Not quite. Looped models do the extra "thinking" inside the model's layers. So the token gets twice the number crunching performed on it before it gets spit out. I think of it as the first loop "kickstarts" the process, and the second loop refines it.
I mean, that's cool and all, but the numbers are really going to shift when it accidentally goes off and orders that same hardware from every vendor in your local region and the top 5 online results for comparison.
It's the same problem as all other LLM solutions (that I hope OpenAI is working on!) it's non-deterministic, and there's no way for the user (or model provider) to know what the distribution of possible outcomes is. This just gets compounded when multi-call harnesses come onto play.
So much of the current bubble is predicated on the hype train continuing on its current trajectory. If anything lags behind, the bubble deflates.
For one potential path - Less datacenter construction means less GPUs, so less $$$ for Nvidia, which means less for them to spread around to prop up the LLM industrial complex, which means lower valuations, which means OpenAI and Anthropic can't get their trillion dollar valuations, which means VC's race for the exit to cash out, etc, etc.
OpenAI, Anthropic, Google selling equity for the first time since their IPO, all of the downstream consumers of the frontier labs. NVIDIA is perhaps the only one profiting from actually selling something. Who hits a liquidity event before the house of cards blows over?
The original claim was "Their debt load and infrastructure costs compared to their relatively meager earnings..." which is just obviously not applicable to Google and NVIDIA, which both crank out more than $100 billion per year in profits. They can literally take on almost a trillion dollars each in debt to finance data centers and still be in reasonably good financial shape.
OpenAI and Anthropic are not very profitable (yet!), but rely mostly on equity financing, not debt. So no debt load issue there either.
Google search only dies if AI wins, in which case there's no shortage of demand for tokens. NVIDIA's revenue just doubled year-over-year; even if token demand cools off and they slow to say 25% growth, they'll be fine.
OpenAI and Anthropic not making money would be sad for their shareholders, but that's sort of a different topic from whether the "debt load" you originally cited will be any kind of a systemic problem (it won't).
A bubble is a debt-fueled run-up in asset prices driven by over-optimistic projections of future fundamental value.
The housing bubble was people borrowing money (debt-fueled) to buy ever more expensive houses (rising asset prices) on the assumption home values would always keep going up (projection of future value).
'90s tech bubble was similar.
Where are you seeing that now? Which asset is over-valued?
Same! Still getting used to it, but the benefits of being able to host my notes on my Tailscale network are huge. Having a built-in Lua engine to customize the interface and content is a really interesting concept, though I'll admit I haven't explored it beyond adding an AI assisted voice to text button. Still, it's been fun using it!
Yes, Canadian dairy prices were always higher than US prices, probably by 10-20 percent or so. There's an internal quota system, and a prescribed allowable volume for imports with very high tariffs on imports above that prescribed limit. Canada did this to protect its dairy industry, and to ensure there's a level of self sufficiency to their food production.
Canada has the same issue in many industries, where the need/desire for Canadian firms to compete against US firms - even though Canada has a population roughly a tenth the size of the US - results in a handful of bloated monopolies/monopsonies that the Canadian population has to deal with.
For an alternative perspective you can look at the US healthcare system, where the US has a protectionist system that props up its healthcare industry (insurance, providers, pharma companies, etc) at the expense of its population, to the tune of about 2x what Canada spends per capita.
Canadians who live near the border will commonly hop over for some shopping and fill up their tank with cheap gas, while Americans try to fill their prescriptions from Canadian pharmacies and travel to Mexico for procedures they can't afford to get done in the US. It's all a matter of what the system is set up to do.
Got it, so really Canada has historically instituted protectionist policies to protect certain industries and ensure self-sufficiency for items deemed important to national security/national welfare and the result at least in part is consolidation into monopolistic entities in Canada.
I'm curious what options Canada has to break up these monopolies and lower prices with respect to their ongoing protectionist policies?
Honestly, most of the incentives aren't aligned with breaking them up or trying to lower prices. If the large corporate entities were broken up, they wouldn't have the scale to compete against their much larger American counterparts. They'd either get pushed out or bought up. And low prices aren't the goal for a company that is protected by government policies - they'll aim to maximize profits within the policy framework instead, and the framework itself does the job of preventing competition from smaller upstart rivals.
Think of the Canadian aircraft manufacturer Bombardier. Canada wants to keep a domestic aircraft manufacturer around so they have a backup in case of US shenanigans (like what just happened with the F-35) so the government ensures there's ongoing contracts for those domestic aircraft as well as subsidies for things like manufacturing plant expansion or modernization. But Canada's population only needs so many aircraft, so why would they waste government resources on a new upstart rival aircraft manufacturer? Bombardier knows this too, so why would they spend their own resources on R&D or modernization when they know they can get the government to pay for it instead? It's a feedback loop, and not necessarily a good one.
It really seems like Canada is in a tough spot economically. They don't have the scale to compete against, say, China or the United States but they want to protect domestic industries against foreign competition and so have enacted trade barriers and other protectionist policies like supporting Bombardier (or the dairy industry or whatever) even if it means higher prices for Canadiens and goes against free trade practices.
In my mind that makes a lot more sense as to why, at least in part, there is the ongoing trade dispute between the US and Canada (whether I agree or disagree with it and how it is handled would be a different matter). The US also seeks to in part protect domestic industries that are being outcompeted by China or face significant regulatory barriers (diary industry in Canada for example, other various exports to the EU, Chinese ban on American technology companies) and so we're winding up in a world where the US is now also putting the hammer down on protecting industries like the EU, China, and seemingly Canada does which is causing a lot of political conflict over trade and jobs.
It seems inevitable that as one or more countries put up trade barriers (EU, Canada, China being some large ones here) the US and others would eventually respond if they found that they were unable to compete on price due to government intervention.
The US also has significant agriculture subsidies. Most developed countries do.
In dairy, for example, if prices get "too low", the US directly buys excess supply to increase the price, makes direct payments to dairy farmers, and subsidizes feed (corn/soybeans) and farmers' insurance.
Most historical trade agreements between the US and friendly nations tolerated minor interventions and subsidies on both sides.
Trump is taking a completely different tact. His goal is not free trade or sustainable relationships. He recognizes the US is the stronger party in most of these negotiations, and wants to use that to extract concessions that explicitly favor the US, even if it strains and damages the economies of historical allies.
> Trump is taking a completely different tact. His goal is not free trade or sustainable relationships. He recognizes the US is the stronger party in most of these negotiations, and wants to use that to extract concessions that explicitly favor the US, even if it strains and damages the economies of historical allies.
I don't think that's even it. The end goal is humiliation of the other party. Trump would be pro-humiliation even if that didn't favour the US economically. (As it typically doesn't.)
Right, my understanding is that the US along with most developed countries provide agricultural subsidies and/or tariff agricultural products from other countries or provide other formal or informal trade barriers. Makes sense, in a way, as you need to keep your agricultural base if you have one. It seems logical that this could extend to other industries, car manufacturing, weapons/missiles, semiconductors, other various technologies. Healthcare. You get the picture.
But I think your comment about Trump is not quite nuanced enough and of course you have to include that the US is negotiating with other powerful parties like the EU and China so it's not really always the US negotiating with weaker parties. At the negotiating table I view each of those big 3 has having many cards to play, and both weaknesses and strengths that can be exploited.
When I say it's not nuanced enough what I mean is that, for example, American technology companies are simply banned from operating in China. In Europe, until I believe July or something American car exports were tariffed at a 10% rate, and then you have policies like the Online Streaming Act in Canada (I'm not exactly up to date on the current status but just recall it existing).
So while Trump's negotiating style is, arguably something to abhor, I think painting these engagements as simply a matter of Donald Trump isn't interested in free trade or sustainable relationships to be a bit off the mark. For example, you can't simultaneously claim that Donald Trump isn't interested in free trade while ignoring trade barriers put in place before he was even elected president. I'd also argue that in general most countries and people aren't actually interested in free trade and that it's more of a political tool, because that would require abandoning subsidies and protections and there are obviously very bad potential repercussions of doing so.
De Havilland is another Canadian aircraft manufacturer, and is expanding massively right now. But I suppose they serve a different, simpler market than Bombardier.
Kinda? Originally a Canadian subsidiary of a British company, privatized in the 80's, then sold to Boeing, then bought back by Bombardier with big financial support from the Ontario government, then sold off to a domestic buyer in a corporate restructuring. Their big commercial aviation product is an old Bombardier design.
To your point - Even as domestic "competition", the market is still separated by aircraft type.
And honestly, their history is a good example of why countries act in this way. When they were sold to Boeing, it seemed to be a way for Boeing to "grease the wheels" in order to land a big Air Canada contract. When Airbus got the contract instead, Boeing immediately put De Havilland up for sale and let the company languish until it was purchased by Ontario/Bombardier. Corporate histories can be interesting!
If the US military purchased camera guided smart munitions with a miscalibrated camera, such that it ignored adult targets and instead specifically targeted children, would that also be a simple mistake per your argument?
If one camera out of thousands has this issue, then yes it's a mistake. If the army doesn't notice it isn't working well after shooting the first child, I'd be learning more towards blaming the army.
I find these style of models are great, but fail hard, and fail randomly. I'd be hesitant to use it for a daily driver, but I'm using dual 3060s, so it's not like I'm quantizing a frontier model here.
How do you find the overall experience? And do you have any special sauce or recommendations for going this route?
I’m using DeepSeek V4 Flash on 128gb mbp - it’s a bit different using a 200b+ param model. It’s MoE so performance is acceptable. It will still malform a tool call every now and then, but the capabilities are so far ahead anything else that the majority of the time it works really well and solves really complex problems.
If it wasn't a marketing stunt, they would have at most quietly settled any legal matters with huggingface behind the scenes, fixed their evaluation harness so it wouldn't happen again, and avoided the potential future liability.
reply