Europe is quite conservative, in the sense that they would not invest billions into an unproven venture. It makes sense that it would excel at an industry that requires putting safety above everything.
The article says they did a lot of customer research and even lobbying, leading to fuel efficiency focus and reduced size, and sticking the finger up to various offended European countries (not taking delegates to US, eschewing RR engines). This seems like savvy being sustained over decades. It must be cultural.
Even if you ask every person to walk the earth what they want, that won't allow you to know future demand. The market shifted largely from hub-and-spoke to point-to-point during development. Without the benefit of hindsight, it must have looked like a solid bet.
A380 was also the result of "customer interviews", but after all the years needed to complete the project the customers have changed their mind, preferring direct flights over hub-and-spoke flights.
When A380 started, and even when it was delivered first, the answers to "what will be the preferred form of airline transport network organisation, in detail" was not yet fully answered.
And A380 simultaneously served as base (in many critical areas) for the quite quickly made A350 et al
IMHO Europe changed massively since the 80s and 90s in that regard, though.
Arianespace was pretty much SpaceX of the 80s and there were quite a few tech companies back then. Due to various reasons stagnation entirely took over Europe after the start of this millennium. Hard to say why. Certainly not putting all the blame on them (since Britain isn't doing that great either) but I don't think especially the Euro and the EU becoming much stronger helped.
The March 2025 blog post by Anthropic titled "Tracing the thoughts of a large language model"[1] is a great introduction to this research, showing how their language model activates features representing concepts that will eventually get connected at some later point as the output tokens are produced.
The associated paper[2] goes into a lot more detail, and includes interactive features that help illustrate how the model "thinks" ahead of time.
And, to pick an example from the research, being able to generate output that rhymes. In fact, it's hard to see how you would produce anything that would be considered coherent text without some degree of planning ahead at some level of abstraction. If it was truly one token at a time without any regard for what comes next it would constantly 'paint itself into a corner' and be forced to produce nonsense (which, it seems, does still happen sometimes, but without any planning it would occur constantly).
I don't think you're wrong but I don't think your logic holds up here. If you have a literal translation like:
I have a hot dog _____
The word in the blank is not necessarily determined when the sentenced is started. Several verbs fit at the end and the LLM doesn't need to know which it's going to pick when it starts. Each word narrows down the possibilities:
I - Trillions
Have - Billions
a - millions
hot - thousands
dog - dozens
_____ - Could be eaten, cooked, thrown, whatever.
If it chooses cooked at this point that doesn't necessarily mean that the LLM was going to do that when it chose "I" or "have"
That's why I hedged with "seems likely" and added "in context." If this is in the middle of a paragraph, then there are many fewer options to fit in the blank from the very start.
It's actually true on many levels, if you think about is needed for generating syntactically and grammatically correct sentences, coherent text and working code.
Just generating syntactically and grammatically correct sentences doesn't need much lookahead; prefixes to sentences that cannot be properly completed are going to be extremely unlikely to be generated.
that's actually the correct use of the phrase "the exception proves the rule"
the rule is that the parking is allowed; the exception is that it's not allowed on Wednesdays; they didn't bother spelling out "parking is allowed at all other times except"
You can also look at the price of opensource models on openrouter, which are a fraction of the cost of closed source models. This is a market that is heavily commoditized, so I would expect it reflect the true cost with a small margin.
If you make careful calculations and estimate the theoretical margins for inference only of most of the big open models on openrouter, the margins are typically crazy high if the openrouter providers served at scale (north of 800% for most of the large models). The high cost probably reflects salaries, investments, and amortization of other expenses like free serving or occasional partial serving occupancy. Sometimes it is hard to keep uniform high load due to other preferences of users that dont get covered at any price, eg maximal context length (which is costing output performance), latency, and time for first token, but also things like privacy guarantees, or simply switching to the next best model quickly. I have always thought that centralized inference is the real goldmine of AI because you get so much value at scale for hardly any cost.
Presumably the model is trained in post-training to produce a response to a prompt, but not to reproduce the prompt itself. So if you prompt it with an empty prompt it's going to be out of distribution.
The study seemed not very convincing to me, at least the way it was described in the article. To summarize: they asked crowdworkers to write a law who used legalese, but not when writing news stories about it or when explaining the law. From that the researchers concluded that people use legalese to convey authority.
But what if people just imitated the writing style of existing laws, but not with the intention to make it authoritative but because that is what they understood their task to be?
I agree. Building on 200 Prolific answers and inventing names for their "own hypothesis" called "magic spell"? Odd.
Lawyers have written like entire libraries on this subject, there are specialized journals examining the legal language used (e.g. in English: https://link.springer.com/journal/11196, https://www.languageandlaw.eu/jll, but there are probably separate journals for this in every language with 10M+ speakers, like https://joginyelv.hu/)
I understand this is not about the lawyers' approach to the problem, even if the author has a law degree, but a "cognitive sciences" department trying their hands on a problem that is new for them.
But it would have been helpful if they had at least attempted to provide a reference to some prior art in the legal field...
