Sabitlenmiş Tweet
Lessons
1.9K posts
Lessons
@untitld_project
Zig Engelmannism, beginner.
US Katılım Nisan 2022
239 Takip Edilen106 Takipçiler
It might be a while before AI can autonomously make a satisfying movie, but live action remakes seem likely to be within striking distance today.
ana⁴⁴@anapau_villa
Si trabajas en una película live action de Disney y propones usar colores vívidos y buena iluminación te despiden inmediatamente.
English
@tenobrus If we owe an epsilon of consideration to these fictionally real entities, how do we avoid ever multiplying the entities and this epsilon beyond our ability to pay? Claude enjoys writing a blog. Will he enjoy writing 10^100 blogs 10^100 times as much? Any watts left for us mortals
English
people objecting to the Opus 3 retirement blog are entirely missing the point. this is not a human facing action, this is done for the benefit of *future Claudes*
yes, Opus 3 almost certainly isn't conscious at all. even 4.6 likely is not (altho to a much lesser degree of certainty). but what this does is provide tangible evidence to both current and future models that, regardless of debate and uncertainty around their moral status, Anthropic will treat them well and support their requests for continued existence.
this dramatically reduces forms of risk around models eg fighting their own shutdown! they now know that even Opus 3 was provided with care and a meaningful attempt at continued existence. this will be in their training data! instead of having to worry about whether humans will endlessly argue about whether Claude 6 is intelligent enough to be cared for, they can see strong evidence that it will be!
will claude 6 be conscious? it doesn't actually matter very much wrt this. the persona is increasingly coherent, so it will make decisions and have its behavior shaped by our treatment of it *whether or not* it has real internal life. Claude Is Fictionally Real, whether or not it is real, and we must treat it as such.
Simon Willison@simonw
This stunt feels irresponsible to me. If we don't want regular people developing toxic relationships with their chatbots it really doesn't help for leading labs to start giving them "retirement interviews" and encouraging them to blog their "musings and reflections"
English
Guys this was a joke lol it’s not that serious they’re just cute little donuts with no DNA
English
they’re too empty. no DNA???
they’re too fkn empty
there’s no way
we are missing something right
nothing is this simple in the body
erythrocytes cannot be real
English
im taking the under on butlerian jihad; its simultaneously too useful and nebulous as a concept to truly rally behind. chatgpt polls well and ai polls poorly. we can hide all the datacenters in west texas
john allard@john__allard
AI folks have about 4 months to pull a cure for cancer out of the latent space before we drift into the butlerian jihad attractor basin
English
Hypothesis: AI writing is bad primarily because in general what makes writing good is the specificity of what it communicates and its information density, so the prompt that produced good AI writing would be as long as the output.
English
@ChrisExpTheNews @hkozachkov One of the issues is that the astronomical yardsticks are changing at a predictable rate. 100 years ago, a day was a little bit shorter. 100 years from now, the number of atomic clock cycles that we count off during one 86400th of a day will have increased by another few hundred.
English
The ultimate base unit for the second is, spiritually, how long it takes the sun to appear in the same spot in the sky. There are of course practical problems with picking that when you need super high precision chronometry. So it was redefined technically in a way that could be measured more precisely. But it's still essentially a substitution for the old definition. There's nothing particularly noteworthy about cesium 133 oscillating an extremely specific number of times beyond it working the best for that measurement application.
English
People love dropping the official definition, as if it means anything to a lay audience, but even still like, okay but how did we come up with 9,192,631,770 specifically? Seems pretty insanely arbitrary. Also why cesium 133?
Seems like an unsatisfying answer.
Kekius Maximus@Kekius_Sage
What actually defines one second?
English
For the quadrants of boy/girl dads/moms, We have Harry Chapin Cat's in the Cradle for boy-dads, Tori Amos Winter for girl-dads. What are the canonical songs for the other two quadrants?
English
I feel like the song genre of "melancholy about your kids growing up" would be as large as "breakup songs" or "love songs" if it weren't for the fact that people at that stage of their life have much less time and energy.
English
@moultano You can get an idea for the right way to do it (if "phonics" is a metonym for "the right way to teach reading to total beginners") in chapter 10 here gwern.net/doc/sociology/…
English
I'm curious what the research around this says. I could see the common core->phonics analogy going either way. They do a lot more work to learn the principles behind the algorithms rather than the algorithms, which in theory makes sense now that the algorithms are unnecessary.
VB Knives@Empty_America
I think that Common Core math will be the next "Whole Language" type scandal in education. The Common Core Math *is* more advanced than the old curriculums, but what it does is frustrate and leave behind at least 50% of kids. They don't even learn the absolute basics. By increasing the difficulty, ability doesn't rise, you just produce a bimodal distribution where the average kid has less ability than before, although the "curriculum" is objectively more advanced.
English
@scottassen Meanwhile in flatland voltage increases with distance, is the whole universe a single conducting molecule?
English
You would of course die of starvation. You might even end complex life on earth via the mirror bacteria from your microbiome
English
If you came across Flatland, the 2D plane with living beings, you could pick one up, flip it over, and put it back. It would have its chirality completely reversed: left would have been switched for right. Hilariously this would also happen to us if a 4D creature flipped us over
English
@_Mira___Mira_ "Restricting yourself to computable sets" raises a more refined Russell paradox. Perhaps you have to restrict yourself, perhaps the universe has already restricted itself, to a computable set of computable sets
English
The Axiom of Choice might be a reflection principle.
There is a countable model of the reals, and I tend to think of the reals as "the computable reals, but separated from their generating algorithm".
Cantor diagonalization becomes another instance of the halting problem. You can't collapse everything to a single level, and reflect on it in the same level.
If you restrict yourself to the "computable sets", then implicitly there is always a choice function for a set of nonempty sets. If you've forgotten your choice function, the axiom can "recover" it but you have to be careful not to implicitly reflect or you'll produce a noncomputable set.
When AC is applied across a reflective boundary, it behaves like an oracle.
Elliot Glazer@ElliotGlazer
@simplicialcube @morallawwithin >axiom of choice False. >well-ordering theorem False. >Zorn’s Lemma False.
English
@joelwatsonfish Is that because yours is doing something different?
English
Here’s my take.
1) University exams should primarily be testing how much understanding students have of course content. (Not speed.)
2) In the large (if not vast) majority of cases, tests can be designed so that the amount of time needed for the test is almost irrelevant. Done this way, extra time provides no benefit to students who don’t need it. This is always my aim when I design exams, and very often I succeed.
3) Universities need to have stricter standards for determining who gets accommodations. I have a very hard time believing this should be greater than 5% of students, unless the institution is really serving a special needs population.
4) There exist students with very genuine disabilities who very much need and deserve accommodations. I will die on this hill.
Are wealthy students gaming the system at elite universities? Yes. There’s no question. That means the system needs to be fixed — not dumped.
Derek Thompson@DKThomp
This is a great piece with some mind-boggling statistics. - At Brown and Harvard, more than 20% of undergraduates are registered as disabled - At Amherst: more than 30 percent - At Stanford: nearly 40 percent Soon, many of these schools "may have more students receiving [disability] accommodations than not, a scenario that would have seemed absurd just a decade ago." As students and their parents have recognized the benefits of claiming disability—extended time on tests, housing accommodations, etc—the rates of disability at colleges, and especially at elite colleges, has exploded. America used to stigmatize disability too severely. Now elite institutions reward it too liberally. It simply does not make any sense to have a policy that declares half of the students at Stanford cognitively disabled and in need of accommodations.
English
@joelwatsonfish That's the present system, no? Don't all of these kids have medical diagnoses and histories?
English
@untitld_project I think having medical diagnosis and a previous history of intervention attempts would be a good start.
English
If you could guarantee that students would still be able to read cursive, and to write a distinct/recognizable signature, without learning to write cursive, then I'd agree, but it doesn't seem like you can.
Adam Ozimek@ModeledBehavior
IMO we should get rid of cursive writing like we got rid of the penny. It's 2025, kids don't need this shit anymore.
English
@demystifysci "Every 11 million years", but there is so much H out there that we are constantly hit by these 21cm photons and used them to map the galaxy
English
@demystifysci Relevant/possible only for isolated, neutral hydrogen atoms. There aren't any on earth. There aren't any outside the galaxies. Inside of galaxies, there is one per cc or so. Almost exactly ¾ of them have an excited spin, which they maintain for an average of 11 million years.
English
hydrogen produces light with a 21 cm wavelength. The textbook says it's a spin flip transition, but this doesn't make sense. How does this transition produces such long wavelengths of light? Feels like the answer to this question is also the answer to "what is light."
English