I would argue, that the null hypothesis is that it is not, and that anyone claiming that there is a mote of consciousness are the ones with the burden of proof.
Presumably the intended outcome would be that sites would be pressured into not being malignant actors with regards to user privacy, rather than everyone just ignoring the security notice.
From first principles, 27b dense vs 13b moe, means you spend ~2x more memory bandwidth per request amortized over the whole server (ie, you assume all experts are being concurrently used by some user during the forward pass, then on average the bandwidth required for one forward pass for any individual request is just the size of one expert). deepseek also have some innovations around kv cache and compressed attention which allow for further reductions in memory bandwidth which means that the thing that’s actually bottlenecking inference, (memory bandwidth) is significantly lower than for qwen 3.8, which has been optimized for running 1 instance ~= 1 user.
The power imbalance will ensure that this is unlikely to happen, in much the same way that people generally don’t trust that massive corporate entities have their best interest at heart but might have some regard for their colleagues wellbeing.
I trust that my colleagues are not going to do stupid things with their accounts, I cannot say the same for agents, and making their own accounts that you’re still responsible for seems like you’ve just moved the problem rather than actually resolving it, since ultimately, you’re still responsible for their actions.
Pareto optimal dominant vs a human for the same task, not an unreasonable framing but that assumes that it can actually do the task, which the op was arguing it couldn’t at all. Which, I suppose you could model as the utility of task completion % as being non linear. I have heard many people argue that the nature of work is messy and complicated and many things they do could not easily be emulated or automated. I do wonder how many of those activities are actually something that are connected to a companies ability to generate revenue or are just the messy interactions between people.
I’m not sure that I agree entirely with your framing here, yes, you do at some point need to correct your assumptions with external evidence. But clearly there have been many individuals throughout history who have had incredibly out sized impact in their respective fields as a consequence of the quality of their reasoning.
Python is mostly glue around C/C++ numerical code for the things I’ve used it for, it’s the inverse of lua. In both cases you have a language that’s easy to manipulate tacked on to one that’s fast because time spent writing code is more expensive than time spent executing it.
Never high and a significant number were donated to ukraine specifically because a ground launched pgm has never been a high priority for the US since the USAF exists.
I mean, we could, the issue is that it would be politically untenable. Controlling the straight means controlling Iranian territory, which means a ground invasion and occupation, which means ~2 million deployed personnel until you set up a compliant regime (because Iran, is absolutely massive). Which is why literally every other president has chosen to negotiate with them or keep them at arms length.
There’s fan noise, but the acoustic engineers at Apple have done a very good job of making it be pretty much not noticeable to someone who has noise sensitivity (me).
the only reason to use cython (in my experience) is to write quick numpy extensions without a tonne of extra baggage from having a complete compiled extension in C/C++/Rust. Often, you explicitly want to avoid the compiler offering you any kinds of additional checking because you want to get maximum throughput. Most of the functions I’ve written in cython explicitly opt out of bounds checking etc since the memory access is sequential and bounded by construction and very obvious when you mess things up by writing simple unit tests. Given that, it doesn’t seem useful to me to add any kind of memory safety as additional overhead. YMMV I mostly do scientific computing and signal processing.
I don’t think any of what you just said is true. Code quality matters even with testing even up to proof of correctness using formal methods, because working software is a moving target and quality is a measure of how easy / expensive it is to continue to track that target going forward.
I take the authors point about empires being necessarily cosmopolitan, but it’s not as though it’s a new phenomenon that people have a myopic view about what empire is or how they should operate. (Eg, Russification) Its always been true that empires use various means of social engineering to reduce differences in populations to engender a more uniform polity over various time scales (because it’s beneficial to the entity to do so, eg: Spanish and Italian are both just bad latin) It’s also often true that ethnic cleansing does take place. (If empire was just about getting the already existing social structures to work for you, america would look very different). It’s not true that empires are purely extractive on the periphery, otherwise they would only ever be loose confederations with a tax system and a hegemonic central power. Which is sometimes true, but is not what most empires do. Even saying that empire is just about reifying one polity over another doesn’t hold together all that well on closer inspection. It doesn’t explain why, for example, the Roman Empire shifted from being culturally italic and centered in Rome to being centered in Constantinople and culturally Greek. How could the periphery take over the center if it’s a purely extractive enterprise by one group over its periphery?
I guess the point I’m making, is that empires are very complicated, and trying to reduce them to being solely a thing imposed by one group onto a periphery in order to extract resources is almost as reductive as to say they only exist through colonization and genocide.
While it’s reasonable to say that people get their view of the world through media, I think it’s fine that media can contain simplifications of reality. Otherwise I would have to play Europa Universalis, which is boring.
not only that, but what all of this misses is the fact that the specification that they used was a product of writing the software, not the other way round. It would be much, much more difficult to first write a correct specification than to first write a correct program, because programs are formal systems that can be executed to observe their correctness. You can’t do that with a specification.
The only barely sane rationale that I’ve heard for wanting datacenters in space is that they would give space-x low latency signal processing capacity it would need to turn Starlink into a real time Passive SAR constellation.
While you’re not wrong, this is also simply mechanical, the context window is the root problem here, if the llm doesn’t happen to have read implementation n of a thing that it needs that session, well now you have n+1 implementations.
One would hope, though I suspect they may want to make things a bit flashier than “we made the audio transcription on the keyboard not terrible” in the changelog given the amount of work that’s gone in.
Yeah, apple will be optimizing a model to work on ANE and then turn it into a native app. My only hope is that it has a reasonable api so that I can use that as a generic input source across iOS / macOS that’s equivalent to the ubiquity of the keyboard.