This seems to tacitly assume that the outliers on either side have equivalent weight with respect to whatever is being investigated, while the explicit premise behind this proposal is that in this case they do not.
I take your point, as demonstrated here, that even straightforward written language can be completely misunderstood, and that emoji might help alleviate that, though how they could help in the particular case here is not clear to me.
"However, there are niche applications where the much higher costs of computing in space could be justified. Examples include ... active collision avoidance in the increasingly crowded low Earth orbit."
It seems that in most ports they are required to hold and treat it (and, perhaps to a lesser extent, in inshore waters.) Fouling the places they take their customers to is ultimately bad for business.
A two meter cubed volume would be sufficient to hold that much. I would guess that cruise ships store more urine than that, though presumably not separated from everything else humans dump into toilets.
With the qualification 'in orbit', I imagine it is.
I have something I want to remove rust from, so I have been thinking about sand blasting. Thus, this thought popped into my head: how spectacular would it be to use aluminum powder? (To be clear, I am not going to try this.)
My wife’s reaction to this was “You guys…” but I know she would absolutely want to watch if someone was going to try it.
I would guess you need something rigid and dense to generate enough pressure, though I don’t know if that rules out your suggestion. It does not have to be a ball.
The claim that disasters happen to older plants is not refuted by the observation that lessons learned are applied to the whole fleet.
One might object that there is selection bias in the original claim, due to the slowdown in construction of recent plants, but that is a separate issue. A more thorough investigation of the causes of all events leading to a significant degradation of safety margins would be needed to determine whether and how older designs are inherently more risky and whether that risk can be adequately mitigated given the constraints imposed by their design.
The fact that, prior to Chernobyl, there were several foreshadowing incidents with RBMKs which should have raised serious concerns, suggests that 'lessons learned' isn't much of a reason to be satisfied with the status quo.
This will serve as an interesting empirical test, then: will LLMs do better with Vera than with Go or other languages? The testing so far seems inconclusive (https://github.com/aallan/vera-bench), but the authors make this interesting observation:
"No LLM has ever been trained on Vera. There are no Vera examples on GitHub, no Stack Overflow answers, no tutorials — the language was created after these models' training cutoffs. Every token of Vera code in these results was written by a model that learned the language entirely from a single document (SKILL.md [https://veralang.dev/SKILL.md]) provided in the prompt at evaluation time."
If LLMs do much better with Vera (or something like it) than with traditional languages, we may be entering a time when most machine-written code will be difficult for humans to review - but maybe that ship has already sailed.
Full disclosure: I didn't figure this out myself, I got it from Ms. Vale's review.
I agree that the term "welfare trap" is a loaded one. This looks to me to be a case of refusing to look through the telescope in case they might see something they do not want to.
On the other hand, an accurate digital simulation of a mechanical calculator really does calculate. The "a simulation is not the real thing" objection breaks down when the function is information processing, on account of information's substrate independence.
One of her points is that there are various pesky consequences for AI companies if AI becomes to be seen as conscious, such as what the paper calls the "welfare trap": if AI systems are widely regarded as being conscious or sentient, they will be seen as "moral patients", reinforcing existing concerns over whether they are being treated appropriately. This paper explicitly says that its conclusion "pulls the field of AI safety out of the welfare trap, [allowing] us to focus entirely on the concrete risks of anthropomorphism [by] treating AGI as a powerful but inherently non-sentient tool."
I cannot, of course, speak about this particular incident, but a person inclined to skip procedures expressly implemented to avoid the problem which occurred, or who ignores clear warnings that a problem is developing, is a liability, not a trained asset.
As hinted at by its 2017 postscript, this article is a mess of incommensurable comparisons.