Can you give examples of curriculum elements that you think are aimed at the world of 20 years ago? The usual criticism I see is that school is barely connected to the needs of the working world.

Is there something special about 5GB?

What's wrong with streaming?

The remote host supports SCP and SFTP, but not SSH.

Part of them getting cheaper is becoming higher output, which means the same labor cost gets you more power. For example, in 2018 we got 360W panels while in 2024 we got 425W ones. But I agree this isn't the main component.

whether it is spending on transmission or generation is a fiction dictated by the regulator

That's the key place where we disagree: my understanding is that the "generation" charges are actual money leaving the utility for a competitive market, and this is a real division.

you assert that you know what the monopoly spends.

First, don't we know that? It's a public company and it has to report what it spends.

But more importantly, I do generally think getting a regulated monopoly like this to become more efficient is intractable, at least in the short to medium term.

Maybe we're meaning different things by "cost"? If a large monopoly spends $X to do Y then even if they're pretty inefficient in how they do Y I'd still describe $X as the cost. We might discuss ways to get the cost down closer to what we think it should be possible to do Y for (changing regulations, subjecting the monopoly to market forces in other ways, etc) but "cost" still seems like a fine word for it?

I don't know this area well, but my understanding is that the "generation" portion represents a market where different companies can compete to provide power, while the other portion is the specific company that has wires to my house operating as a regulated monopoly. So while I don't trust the detailed breakdown of the different monopoly charges (I suspect the company has quite a bit of freedom in practice to move costs between buckets) the high-level generation-vs-the-rest breakdown seems trustworthy.

How so?

At $1000/kW-hr it's (just barely) not worth even buying batteries to shift energy from daytime generation to night consumption, while at $700/kW-hr it definitely is worthwhile.

Doesn't this depend heavily on local utility rates, and so any discussion of crossover points should include rates? Ex: I'm at $0.33/kWh while a friend in TX is at half that.

releasing the people that CEA bars with nondisclosure agreements about that one episode with Leverage about which we unfortunately don't know more than that there are nondisclosure agreements

I don't know the details in the Leverage case, but usually the way this sort of non-disclosure works is that both parties in a dispute, including employees, have non-disclosure obligations. But one party isn't able to release their (ex) employees unilaterally; the other party would need to agree as well.

That is, I suspect the agreements are structured such that CEA releasing people the way you propose (without Leverage also agreeing, which I doubt they would) would be a willful contract violation.

Have you check if you are receiving some subsidies for it?

I think natural gas in the US is effectively subsidized by underinvesting in export infrastructure? This country produces a lot of gas.

Talking to a friend who works in the energy industry, this is already happening in Puerto Rico. Electricity prices are high enough that it makes sense for a very large fraction of people to get solar, which then pushes prices up even higher for the remainder, and it spirals.

If the cost of power generation were the main contributor to the overall cost of the system then I think you'd be right: economies of scale and the ability to generate in cheap places and sell in expensive places would do a lot to keep people on the grid. But looking at my bill (footnote [1]) the non-generation costs are high enough that if current trends continue that should flip; see my response to cata, above.

I'm not claiming here that it's currently cheaper, but that it will soon be cheaper in a lot of places. Only 47% of my bill is the actual power generation, and the non-generation charges total $0.18/kWh. That's still slightly more expensive than solar+batteries here, but with current cost trends that should flip in a year or two.

Looking at their breakdown (footnote [1]) it seems to be mostly the cost of getting the electricity to the consumer. Since they're a monopoly, there's not much getting them to be efficient here, operating a high-uptime anything is expensive, and MA is an expensive place to do anything.

That's a very different product, using UV inside HVAC systems as an alternative or supplement to traditional filtration. Because the delivery rate of HVAC as a fraction of all air in the room is so much lower than the fraction air above people in a high ceiling room, this is much less valuable.

Very roughly, the main ways people use UV to clean air to reduce spread of diseases are HVAC / in duct, far UV, and upper room. I'm only trying to talk about the last of these here.

The way you demonstrate that there are not long-term side effects is that we have very accurate ability to measure UV, and so you can show that the system being on vs off has a negligible impact on the amount of UV where people are. Long-term impacts would be downstream from this kind of easily detectable effect.

(I think this is very different for far UV, where you intentionally shine it in a way that does include the people. That is potentially a much better approach, because you can clean the air between people instead of only above them, but while the research on far UVC safety looks pretty good to me, it's a much harder system to gather safety evidence on.)

You do need to pay attention to what paint is on the ceiling and measure to verify that levels are low in the places people are, but pointing UVC up is something we've done safely for a long time in many places.

Thanks!

35min is pretty optimized if I'm starting with "all components loose in a bag" -- it used to take me almost an hour. It's a lot of stuff that needs to get plugged in: https://www.jefftk.com/p/rhythm-stage-setup-components

But setting up a pedalboard so I don't need to manually connect dozens of things (this post!) is also optimizing the process.

Thanks; fixed!

Following Emma's advice on FB I exploded the tracks, healed splits, sorted them by take, and then comped manually ignoring reaper's features that are supposed to make this easier. It worked great, and I now have a rough mix ready for Lily's feedback!

Nit: your "Duplo" diagram shows a ratio of 1:3 which would make it a (hypothetical!) Triplo. Real Duplos are 1:2, and (discontinued) Quatro are 1:4 with Lego and 1:2 with Duplo.

(This is all in each dimension, so the overall volume ratio is 1:8)

The benefits of breastfeeding may be somewhat lower for the median reader of this blog than the median late 1990s Belarusian. You probably have access to cleaner water, more/better medical care and newer formula that includes things like DHA, ARA, and more nucleotides.

Why somewhat? It's plausible to me that even just the lack of DHA would give the overall RCT results.

On priors I'd expect formula to be worse: there are strong evolutionary reasons for breast milk to be the ideal food, while formula is limited by our capability to understand what babies need. If we were missing something important in formula to where babies were dying early we'd notice, but as the effect size gets smaller the chance we notice goes way down. But running studies and updating our practices based on them is still a factor, and DHA is mandatory (in the US) now. I'd still guess that an RCT today would show IQ impacts, but this is based on priors about how hard this problem is, our civilizational capacity, and the incentive structures for formula manufacturers, and not primarily on an RCT in quite a different environment.

Does this fall apart for sufficiently large donors? Could OpenPhil do all their giving via a DAF?

That seems unlikely to me, at least in this case. FarmKind doesn't look to be listing the kind of very small charities that would be in danger of failing this test.

@Raemon I've been playing around with "you climb up, you climb down".  This communicates essentially the same thing, but as "this is how it works" and not as a promise about the climber's abilities.

What’s your rough assessment of AI risk now?

I think it's pretty important and I'm glad a bunch of people are working on it. I seriously considered switching into it in spring 2022 before deciding to go into biorisk

Also, how many people work on your current project? If you left, would that tank the project or be pretty replaceable?

We're pretty small (~7), and I've recently started leading our near-term first team (four people counting me, trying to hire two more). I think I'm not very replaceable: my strengths are very different from others on the team in a highly complementary way, especially from a "let's get a monitoring system up and running now" perspective.

(I must admit to some snark in my short response to Mako above. I'm mildly grumpy about people going around as if alignment is literally the only thing that matters. But that's also not really what he was saying, since he was pushing back against my worrying about dragons and not my day job.)

I don't think inferring a probability anywhere near as high as 20% is justified. If, conditional on finding dragons the value of the knowledge is lower than the harms of being known as a dragon believer, then you shouldn't go check no matter how low your prior.

I agree, but I strongly disagree with @Shankar Sivarajan that if a person does this in some areas then they shouldn't "claim to be 'truth-seeking' in any way".

For people disagree-voting [edit: at the time the parent was disagree-voted to -7], I'd be happy to see arguments that I should switch from trying to detect bioengineered pandemics to alignment research.

I don't think refocusing my main efforts on the alignment problem would make humanity safer.

Operationally it means that I'm not trying to find out the truth one way or the other. If I come across arguments I ignore them, if someone asks if they can explain it to me I say no, I try not to think about it, etc.

the reasons outlined in this post (i.e. "if I checked and it's not the way I thought, that would necessitate a lot of expensive and time-consuming updates to what I'm doing").

Wait, that's not what I'm trying to communicate in the post. If learning that dragons existed would precipitate major updates, it will very often be worth investigating their existence. Instead, it is highly-political low-payoff topics I am intentionally agnostic.

A general commitment to seeking truth doesn't obligate one to investigate every possible question! I can be quite committed to seeking truth in some areas, while intentionally avoiding quite unrelated ones.

One certainly shouldn't claim to be truth seeking in areas where one is are intentionally agnostic, but that's part of why I'm writing this post: so I can later link it to explain why I have chosen not to engage with some question in some area.

Good point! Yes, we will need to clean it in place.

I think it is okay as is, but if they do manage to break it I will redo it with something to spread the force out.

That is very impressive!

We generally don't require our kids to wear shoes. The exceptions are places that require them (ex: school) and places where it's seriously unsafe (broken glass all over)

The title is what I say to little kids when they have climbed somewhere and want me to carry them down. I would prefer to be using phrasing that was literally true, but overall this feels good enough to me? Do you have suggestions for better phrasing?

That is a classic question of population ethics (LW). The author is writing from a totalist perspective (which I think is by far the most common view on LW) while you seem to find a person-affecting perspective clearly correct.

I would expect that to work fine as an adult on my own, but if I'm the one who gets to the street first it would be confusing to my kids.

We are eating food cooked at home, which is a lot cheaper then frozen meals and much cheaper than restaurant food. We have a dinner rotation with housemates (and a neighbor) where we take turns cooking dinner, so Julia and I each usually cook once a week.

What I cooked last night:

• 3lb sausages, bought on sale @ $2/lb
  • I often buy meat on sale and keep it in our chest freezer
• 2lb pasta @ $1.20/lb
• 2lb tomatoes @ $2/lb
• 2lb onions @ $2/lb
• 2 bell peppers @ $2/each

Usually there'd also be a vegan or vegetarian option, but we didn't happen to have anyone who needed that any last night. If we did I probably would have baked some tofu, 1lb @ $3.50/lb.

Overall cost, assuming I'd included the tofu, is ~$25 when you count the small amount of amortized spices, cooking oil, etc. This would be dinner for about 10 people, and several lunches. About $2/meal. $732 is ~$5/person/day, and lunch and breakfast are generally much lighter and cheaper than dinner.

My lunch is also provided by my work, and once a week we have dinner with my extended family at my dad's house. I'm also not counting the electricity used to run the stove, but I think that's probably too small to matter?

A much larger missing expense is the opportunity cost of each adult's time cooking, but I can combine cooking with watching the kids to some extent and it's also something we enjoy.

I only use this setup when I'm away from my desk. If I'm in a convenient location I can just plug into an external monitor at the right height.

(somehow) make bread that was likely to last longer

You could decrease the water content and not leaven it, making it into hardtack. This will last for years if kept dry. But it's very unpleasant to eat, and I don't think people would buy it or bakers would make it unless required to.

Whoops; fixed!

I think it's most likely contamination within the lab doing the nucleic acid extraction, since talking to them they do work with lentiviral vectors.

Setting aside economics or technology, would it in principle be possible to detect a variant of concern in flight and quarantine the passengers until further testing could be done?

There are two pretty different scenarios:

1. Initial detection: if you don't already know whether there's something out there, you'll need to do metagenomic sequencing or something similar to identify the pathogen. This is the part of the problem that the NAO is trying to solve. While I haven't looked into the absolute-minimum-sequencing-time portion of the space deeply, my understanding is if you want a reasonable cost-per read you need to use a sequencing method that (counting both the preparation and the sequencing machine running) takes multiple days. So not a good fit for per-flight testing.

2. Containment: we've learned about a pathogen somehow (ex: someone with unusual symptoms, metagenomic sequencing) and we're trying to keep it from spreading. Now we can use a targeted method, such as qPCR, where there are stand-alone speed-optimized options here that take under an hour (ex: KrakenSense). In this case, the question is, how do you get the samples to test? Ideally you'd get everyone to give a sample before boarding, which you could do a pooled test on while the plane was in flight, but that requires infrastructure and cooperation with the originating country.

you're saying that your prelim results show that 0.2% of the sampled population would need to have at some point in the past been infected for the variant of concern to be detectable?

That's correct. While detection is fundamentally based on the people who are currently shedding copies of the virus, but our modeling counts "time" in terms of the progress of the infection through the population.

suggesting this would be deployed in airports rather than municipalities. So the plan has changed?

We're also exploring arport monitoring, but airplane blackwater tanks not terminals. Preliminary data from pooled tank samples (you collect between the truck that sucks it out of the planes and the dumping point) looks very good.

infected travelers/day in an airport setting to get .2% of the wastewater being from them

Sorry to keep harping in this, but 0.2% of wastewater from people who've ever been infected (cumulative incidence) not currently infected (prevalence). While shedding is primarily about prevalence (though varying over the course of the infection) for evaluating a system we generally think cumulative incidence is more informative because it tells you much more and how far along the pandemic is.

Technically it's 0.2% cumulative incidence not 0.2% prevalence, but depending on the assumptions you make about how long infections last and how quickly they spread they're usually in the same ballpark.

Many SeaTac travelers do not defecate, so your effective sample size is smaller. Possibly too small for this to work well. This modeling is generally assuming larger sewersheds, like municipalities.