HNHacker News
TopNewBestAskShowJobs

pinkmuffinere

3,655 karma · joined August 8, 2021

I like engineering, finance, and tech. My biggest project (as part of a team) is stingray resistant booties: https://mydragonskin.com/
submissionscomments
pinkmuffinere··on Rising sea destroys homes, erases beaches in California
Ya, I don't want to minimize the problem of sea levels rising, but the recent storms had massive swells (I believe partially because of the El Nino year). The sand on most of these beaches is not natural, it's pumped up to the shore to encourage tourism. It is normal for it to wash away during big swell, and not necessarily an indication of the sea level rising.

You could argue that the extreme El Nino this year is due to global warming, which is obviously related, but it's not quite the same as just rising sea levels.

pinkmuffinere··on Back and shoulder surgery is often worse than useless
That is kindof crazy! Imagine you are suspicious that something is awry in your soft tissue, in which case I think MRI is the main diagnostic (?). Then you don’t find out for ten weeks, plus some for analysis!
pinkmuffinere··on Italian parliament votes for return to nuclear energy
Wow, that's really fascinating to me! I am a bit cautious that the distribution of outcomes from nuclear may be skewed, so we've maybe avoided some of the "black swan" events so far simply because we don't have much nuclear deployed. I think the common example is "Picking up pennies in front of a steamroller"; you'll be totally fine until you totally aren't. However, I expect we'd also learn more and improve the safety as we deploy more, so I (somewhat naively) feel this is a good argument in favor of more nuclear.
pinkmuffinere··on AI safety is mostly a sex cult in Berkeley
Unfortunately I can't read the article (taken down), but I can read some of the underlying material shared by dang: https://news.ycombinator.com/item?id=49737985

I feel there's many interesting pieces to this -- Yudkowsky certainly has extreme views, it's clear there are a ton of poly and open people in this community, some people do have a religious veneration for Yudkowsky and "rationality" in general, and there's some smarmy folks in the mix. But I'm not sure any of that is a negative critique of the general movement. I suspect similar statements were true of the enlightenment, protestantism, and the beatniks? Marginalized views, free expression, and new ideas often go together in what I feel is an obvious pattern. I don't care if there's a bunch of sexually liberal people in this org, or some obsessive fan-devotion, so long as they're not forcing somebody into a situation they don't want. These people can live their lives however they want; I think that's quintessentially American in a way that's very positive and increasingly rare. And on the other side, there are many impacts from this movement that I'm quite happy with: I'm really happy that this movement has spawned/impacted GiveWell, the Long Now, Scott Alexander, Works in Progress, and Open Philanthropy, to name just a few. I'm also happy that it seems to have birthed the latest AI progress. Of course there are things that I find ridiculous (I didn't really like hpmor, their forum board is somewhat silly, and Roko's Basilisk[0] is just hilarious, lol).

If this piece was meant as documentation of a particular moment in time, then they've picked a really fascinating group to document! But it feels like it's really meant as criticism, that we should trust the whole group less because there are some sexually open fans of Yudkowsky. I'm not sure I can get on board with that. If sexually open fans of Yudkowsky want to build great things, I say let them!

[0] https://en.wikipedia.org/wiki/Roko%27s_basilisk

pinkmuffinere··on Why is the liver so weirdly regenerative?
This is still the normal in the middle east (and likely elsewhere)! In Turkey you call somebody dear to you "my liver" as a pet name.
pinkmuffinere··on Creatine uptake enhances antitumor immunity
Thankyou!! I'm sure if you have experience in the medical space this sort of thing seems obvious, but to me (non-medical), the original phrasing was so confusing. Your explanation makes sense.
pinkmuffinere··on Italian parliament votes for return to nuclear energy
To be fair, people will say many of those deaths are due to the evacuation rather than the meltdown, and they’ll say the evacuation was overblown — too _much_ regulation. However, I think you’re right that these can still be attributed to the meltdown, since the evacuations were undertaken to avoid the difficult-to-predict impact of the meltdown. In hindsight, we can see the evacuation wasn’t needed. But the reactor meltdown made an already-bad situation even more brittle, and in a disaster scenario i think we naturally prefer the evacuation, as opposed to the (low) chance of total destruction.

The deaths are due to the evacuation, and the evacuation is due to some combination of [poor judgement] and [reactor meltdown]. If we attribute all the deaths to poor judgement, then we’re asserting we could confidently predict a reactor meltdown would pose no danger. Some people might claim that, but it seems implausible to me.

pinkmuffinere··on Feds Target AI Critics as "Foreign Agents"
> “Alarming reports indicate that a network of foreign actors, led by the Chinese Communist Party (CCP), is attempting to manipulate U.S. policy and public opinion on data centers.“

Setting aside the politics, trump, ai-doom/boom, I just think this claim is fascinating. Can somebody steel man this? What’s the best evidence for this?

pinkmuffinere··on Italian parliament votes for return to nuclear energy
I agree coal power plants are much worse. As an aside, I’m not convinced that’s the bar we need to clear. I would think we need to be safer than solar power, since that’s the source that’s actually growing rapidly. If we’re going to add new power generation, I think it’s probably a trade off between nuclear and solar.

And anyways this isn’t the claim I was disagreeing with — the parent claim is that nuclear is regulated “beyond what's necessary to ensure safety”. I just don’t think that’s the case. As dramatic as it sounds, we regularly have nuclear disasters.

pinkmuffinere··on Italian parliament votes for return to nuclear energy
> Nuclear is brutally overregulated, far, far beyond what's necessary to ensure safety, and that's where most costs come from.

Is this really true? We seem to have a pretty large disaster perhaps every 15-20 years (3 mile island, Chernobyl, Fukushima), which makes me think that it is either correctly- or under- regulated. Of course coal and oil are worse, but that’s not exactly a high bar…

pinkmuffinere··on UK military jamming other nations' satellites to defend itself, BBC told
I prefer domestic free-range artisan game theory.
pinkmuffinere··on UK military jamming other nations' satellites to defend itself, BBC told
> Whatever that technology is, why hasn’t it proliferated into civilian life so that we all benefit from a resilient, non-GPS backup?

Once a stealth technology is integrated into civilian life, it's no longer stealth, and you lose the element of surprise if you have to use it in warfare. If somebody _does_ feel they have a clever secret work-around, the incentive is pretty high to keep it a secret.

pinkmuffinere··on Artificial intelligence now beats some of the best human forecasters
I finally got to reading this! Honestly so far it's more about politics than I wanted and less about Renaissance, but it's certainly well-written. I particularly like this line, lol:

> Mercer has asserted repeatedly that African-Americans were better off economically before the civil-rights movement. (Few scholars agree.)

pinkmuffinere··on Samsung is expected to more than double output of its HBM4 and HBM4E DRAM
> They may in principle have established the technologies to be self sustained but that’s only relevant and sustainable if they develop an economy based on domestic consumption rather than exports. Which they are struggling with as well.

Is this really true? Can't they just keep exporting "forever"? What requires them to find a domestic audience? I can think of some other countries that have done quite well mostly without finding a domestic audience (eg, Switzerland comes to mind, and frankly Europe might largely be in the same bucket).

pinkmuffinere··on Weeping whales: Stillborn humpback whale grieving documented
Are you trying to say that including the weasely claims makes the article better and more scientific? You are “possibly” right.
pinkmuffinere··on The American Religion of Self-Storage Facilities
> Swedish death cleaning is a decluttering practice from Sweden that emphasizes organizing your belongings before you die

Well I haven't died yet, so I suppose the time is right.

pinkmuffinere··on Artificial intelligence now beats some of the best human forecasters
lol this is a great quote, it's like a couplet from a standup set. It's humorous, unexpected, on closer reading it's possibly true, and then you see the source and immediately realize it must be correct.

Any article you'd recommend about Renaissance? I've always been curious but not curious enough to read "just anything"

pinkmuffinere··on Tell the speakers that you liked their talks
I _really_ like this advice, I will be filing it away for future use. Thankyou for your suggestion :P
pinkmuffinere··on AI recursive self-improvement might not come so quickly after all
Ah it's interesting they legitimately used 10k agents, I didn't realize that. I do agree that it's significant they solved the Millenium problem only once humans had made significant headway. I'm not sure I believe the relevant training data will be produced at a significantly lower rate due to AI -- Millenium problem solutions weren't generated at a very high rate before anyways. But I think all your points hold nonetheless. Thanks for elaborating on them!
pinkmuffinere··on AI recursive self-improvement might not come so quickly after all
To preface -- I try not to be dogmatic/politicized on AI, so I will genuinely consider your arguments! Please try to convince me. (indeed, I am the grandparent commenter)

I agree with the meat of your statement, but am very interested in the pre-emption, "they spent $30M do do something closer to a 1000 monkeys approach, rather than a singular inference being very intelligent". First, I think the $30M number is inflated -- that's what the general public would have paid, but presumably the internal cost is lower, perhaps it's more like $10M. But it is still expensive. Second, I'm curious if it's really the case that they did a 1000-monkeys approach? I haven't read much in-depth reporting about the proof, so it's totally possible I just don't know. What is it that they did which is more like 1000-monkeys? Also, I wonder if that distinction matters -- if 1000 monkeys can reliably make ground breaking proofs, and the approach generalizes to other tasks, I will happily become a circus owner. Maybe you're claiming that it won't yield other proofs? Or the proofs are too opaque to be useful to humans? Or it can handle proofs but not other tasks?

pinkmuffinere··on AI recursive self-improvement might not come so quickly after all
I made this same reply to another thread, so I'm sorry to say essentially the same thing twice, but -- the comment has a familiar structure, "it doesn't work for you because you used an [old / suboptimal / non-frontier] model. If you use X you'll see that it works". These sorts of claims push the onus back onto the other person (or in this case tfa), without really accepting the result, or taking on any work for yourself. It gets tiresome to retest with the newest model every other week. Is there any data you can provide to support your claim, or any result you can contribute here?
pinkmuffinere··on AI recursive self-improvement might not come so quickly after all
I think your reply has a somewhat familiar structure -- "it doesn't work for you because you used an [old / suboptimal / non-frontier] model. If you use X you'll see that it works". You might be completely correct! But these sorts of claims push the onus back onto the other person, without accepting any work for yourself. It gets tiresome to retest with the newest model every other week. Is there any data you can provide to support your claim, or any result you can contribute here?
pinkmuffinere··on Mexican student creates an acoustic fire extinguisher to put out fire in seconds
> I know the field of loudspeakers/acoustics very well, what I've said so far and will say now is all off the top of my head, and it's unlikely that it's in his head at all.

The wikipedia article (linked above, linked here for your convenience) [0], disagrees with your claim, and matches what the action lab says. Perhaps you can reconcile the difference, or you can correct the wikipedia article. The article says the bass reflex improves "the efficiency of the system at low frequencies", and "the efficiency can be substantially improved over the performance of a similarly sized sealed-box enclosure". My best guess is that are defining efficiency as something other than total_power_out/total_power_in.

[0] https://en.wikipedia.org/wiki/Bass_reflex

pinkmuffinere··on Mexican student creates an acoustic fire extinguisher to put out fire in seconds
tl;dr -- aside from the point about fuel, you mostly don't disagree with him. You have different definitions for the same terms he uses, and you draw your system in a different place than he draws his. Both of you make reasonable choices. A youtube video is not an academic paper, and cannot start by rigorously defining the system under discussion.

> he called oxygen the fuel for the fire... it's not the fuel for the fire

I think you're correct here, from chemistry definitions oxygen is not fuel, because breaking its bond consumes energy rather than releasing it. However, I don't think the host is trying to give perfectly precise chemistry definitions here, he's trying to explain how fire burns, and calling oxygen 'fuel' makes it clear to his audience that oxygen is required for combustion. I think your objection here is similar to objecting that voters in the US elect electors rather than candidates -- it's correct, but not necessary if you're explaining the concept of democracy.

> the port actually decreases the efficiency of the speaker by letting the negative bass note sine waves cancel out the positive sine waves from the front of the speaker, but this is done to tone down the "boominess" of the speaker at its resonant frequency, thus making the speaker more "neutral" sounding

I don't know much about bass reflex speakers, but I read the wikipedia page in decent depth, and it appears to me that you are wrong, or perhaps using a different definition than the field itself uses. From the wiki page: "This port enables the sound from the rear side of the diaphragm to increase the efficiency of the system at low frequencies as compared to a typical sealed- or closed-box loudspeaker or an infinite baffle mounting... Depending on the exact relationship between driver parameters, the enclosure volume (and filling if any), and the tube cross-section and length, the efficiency can be substantially improved over the performance of a similarly sized sealed-box enclosure." [0]. I bet I can see where your disagreement comes from though -- I think you're arguing that adding the reflex speaker cancels some of the power going through the main speaker, and therefore must result in decreased total_power_out/total_power_in (forgive me if I misunderstand). This is a sensible way to define efficiency, and I think you're correct that total_power_out is strictly decreased. However, there are also other sensible ways to define efficiency, eg, instead of total_power_out you might care about the perceptually-weakest frequency the speaker generates, to ensure the sound is "balanced" to a human ear. You might then define efficiency as num_spectral_bins*min_weighted_spectral_power_out/total_power_in. I have some experience optimizing perceived smoothness, and we used a metric very similar to this. Without fully knowing what you mean or what he means, I think you're disagreeing over terms, not correctness.

> but in any case, what he says about the "bulk" movement of air (he should probably call it net, i mean all air is bulk by nature) there is no way that the speaker pumping air in and out of the cavity could turn into net movement in one direction... the air coming out was directionally focused from the cone? and sucking back in was amorphous and somehow not detected by his device?

I think you're agreeing with him here, his explanation matches yours. We can break a full period of oscillation into two sections: a high pressure section (the air molecules being pushed away from the speaker), and a low pressure section (air being pulled back). The high pressure section travels _coherently_ through the air cannon, mostly moving air molecules in a single direction (as you say, it is "directionally focused from the cone"). The low pressure segment pulls air back into the air cannon, but it is _not_ coherent (as you say, it is "amorphous"). During the low section, molecules can flow towards the air cannon from the side of the cannon; they do not have to come from directly in front of the air cannon. If you rewatch that section I think you'll see you're not disagreeing with him. To be fair, more drawings would help communicate it more clearly. You're correct that if you draw your observation box big enough, there can be no net movement from something wiggling back and forth, but I think he's drawing his observation box over the fire. If you look at that system (the air above the fire), you will see net movement.

[0] https://en.wikipedia.org/wiki/Bass_reflex

pinkmuffinere··on Mexican student creates an acoustic fire extinguisher to put out fire in seconds
This channel (The Action Lab) is great, it sets an amazingly high standard for clarity of explanation, and for practical combination of experiments and theory, all while being appealing to a broad audience. Most educators would be extremely proud to produce work that's accurate to the standards of a university professor, while remaining appealing to middle-schoolers. I can't imagine what more you want from the content.
pinkmuffinere··on OpenAI fought dirty on career-making math problem
Wow I wonder how much of that differential is them operating at an extreme loss, and how much is the relative cost of their new models vs old. If a significant portion of that is net losses, I'm very concerned for their business model :|
pinkmuffinere··on OpenAI fought dirty on career-making math problem
I don't think I'm discounting the human breakthroughs, the cost of employees, or the previous attempts. All of that is significant work, and has lead us to a scenario where the marginal cost to solve one millenium prize problem appears to be ~$10M. I feel like you're taking a very "to bake an apple pie you must first create the universe" perspective on this; I make no claim about how much money was spent in the past, on whether it was well invested, etc. Regardless of how we feel about it, it's already been done, right?
pinkmuffinere··on OpenAI fought dirty on career-making math problem
Setting aside the disagreement, I was very interested to see the net pricing of the discovery:

> All told, the week-long effort consumed 300 billion output tokens — $22.5 million worth of compute, if charged at current Astra rates.

> The Navier-Stokes existence and smoothness problem is one of the seven Millennium Prize problems — a set of major unsolved math problems, each carrying a $1 million bounty

I know openAI isn't solving these problems in order to make profit, but it's interesting to guess how close we are to these things becoming profitable. Eg, if you think their public pricing for Astra is ~2x as expensive as their internal price, then they lost ~10M on net for this proof. That's not profitable, but it is much better than I would have expected, which is exciting for the other Millennium prize problems! Of course the fundamental approach (which they may have plagiarized from Buckmaster and Alpoge) might have added cost to that as well. Nonetheless, I wouldn't be too surprised if they're all solved within the next 3 years!

pinkmuffinere··on Tell HN: OpenAI brings back 5 hour limit for plus and business standard users
Ya, setting aside any judgement about what's "right", this strategy doesn't seem profitable. There's no real moat between one model/provider to another, switching is relatively easy. I don't understand how this is supposed to work for them.

It strikes me as similar to UPS / USPS / Fedex -- everyone uses the mail, and they mostly use whichever is cheapest for their requirements. I don't think there's much loyalty to specific services, and people are happy to switch between the options

pinkmuffinere··on Americans Without Degrees Are Having One of the Best Job Markets in Years
> Degree is over-valued as a life goal in a time where we no longer really value tertiary education the way we should

What does “the way that we should” mean here? Are you disagreeing with the free market valuation of a good? If not the equilibrium created by supply and demand, what do you propose instead?

← PreviousPage 2 of 34Next →