Posts
Comments
No, but these tweets describe the basic concept: extension of AlphaZero to train on theorems in Lean automatically converted from natural-language proofs.
Looking first at figure S3, nothing stands out. The most important figures here are A, B, and G-I. G-I are showing the % of sequencing runs that gave an unexpected base at I-Ppol cut sites, I-Ppol alternative cut sites, and 100,000 randomly selected sites. There’s a little bit of variation in the ICE samples at canonical sites, but I consider this pretty convincing evidence that the ICE system is not particularly mutagenic.
You should reconsider that conclusion. Double-strand breaks cause an increase in mutation rate. This is a known fact.
The experiment design is weird and the supposed results contradict a lot of other things.
see also phonons
I previously wrote a post that answers this question.
You only need a majority of voters, who would then vote for local government that would negotiate a mutually-beneficial deal. Not every single person.
People can already build a single normal house on a normal lot well enough. The people who want to massively increase density want big apartment complexes built. Big projects.
residents made the areas valuable by doing things they've since disallowed
It's not mainly the buildings that make an area valuable, it's who lives there. If there's a problem in how appreciation is distributed I'd say it's that non-resident property owners capture some value that they don't deserve to.
If developers need to compensate nearby property owners for negative externalities imposed on them, then who compensates the developer for the positive externalities they cause on properties the developer doesn't own? Because if the answer is "no one" then this is a pure disincentive on development
...governments do? It's common for big commercial/industrial projects to get big incentives from city and state governments.
Right now builders face an enormous number of veto points in any construction process someone once decided that this makes it possible to be used as an illegal apartment
People vote for ideological anti-development hardasses because other people kept getting bribed to look the other way while developers got special permission for an apartment next to your house, or because houses in the area got split by 10 guys.
If increasing density in a populated area has some costs to current residents but is worthwhile overall, it should be possible to pay off the existing residents to allow it. What I see instead is that developers consider that too expensive to make development profitable, but can bribe politicians/officials for less, and YIMBYs are useful patsies for this.
When I've mentioned this to YIMBY people, the response I've gotten was basically: "Those people in high-value suburban areas don't deserve to live there the way they do." But it's the existing residents who made areas valuable.
the reward for playing the game is more game ... It’s just the same thing over and over until you die. You don’t get out by winning; you get out by stopping.
What does this say about what it means for someone to be "qualified" for a position? What does it say about where you should expect to find the smartest people?
That has all been considered extensively before and this post isn't a good place to discuss it. Prizes have been found to be generally worse than patents and research funding.
Right. "Having a nucleus" is a pretty big difference.
these multiplex CRISPR edits of wood trees to have a 229% yield increase for paper-making
It's possible to use breeding techniques - and these days, you can do "molecular breeding" - to get a wide range of cellulose-lignin-xylose ratios. Anything outside the naturally existing range is going to be quite bad for plant growth/viability. Also, a 229% increase must have been from a particularly low baseline, much lower than anything currently used for making paper.
This is something I looked into a bit when I was pondering candidates for biomass -> chemical conversion.
Note that the "bridge RNA" method doesn't currently work in mammals. The team involved is trying to use AI to edit the enzyme involved so it does, but that might not be workable, due to differences in how DNA is packed and anti-virus defense systems.
What exactly is the difference that's needed from current large cruise ships? Is it size per se? Independent production of food and fuel? Production of trade goods?
In general, things with a higher value per mass have less price variation across countries, because transport costs are less important, but less competition and price transparency, because they're more specialized and lower-volume.
Maybe we should take a step back and consider what your higher-level goals are here. I'm not sure what they are. The possibilities that come to mind are:
- build large stuff out of pykrete because it's novel
- build large stuff out of ice because it's large
- establish a community separate from existing governments (seasteading)
- make more land so there's more land overall
If you want to argue against the very existence of the patent system, you should make your own post for that. This isn't the right place for that.
I think there's some underlying misunderstanding of material science here. You have some fibers, OK, and they have some amount of net strength. Distributing them in ice or whatever at a really low concentration doesn't increase the total amount of strength those fibers have. It's not better than putting those fibers together with less filler, unless:
- you specifically want only a little bit of extra tensile strength for a lot of material
- you want to keep viscosity of a thermoplastic or resin sufficiently low during processing
- you're getting much better dispersion at lower loading
Adding sawdust to ice, you're not going to get more additional strength per wood than just using good lumber or plywood.
something to be said for UHPC concrete megastructures if you can scale up and vertically integrate the manufacturing
it seems some people agreed with you about that https://www.youtube.com/watch?v=tIQrGfV9oA8
Engineering toolbox lists it as 20-40 MPa
...do you think all concrete has the same strength? Here is a paper with "concrete" that has 800 MPa compressive.
So once it gets to say -40 Celsius it has a compressive strength of around 60 MPa
- If you care about creep, ice at -20 C shouldn't have >1 MPa on it.
- Measured compressive strength of Pykrete was much lower. It took me 10 seconds to find this paper, there's some data for you. 20 MPa with 14% sawdust, but creep would obviously happen at lower stress.
For 50 MPa concrete you basically need to add 1% of additives that are maybe $3500/ton.
For much higher strengths you start needing expensive stuff, eg silica fume.
Cooling at -100C is a lot more expensive.
- I saw 14% for Pykrete. $10 to $20 a ton * 14% is $1.40 to $3.80.
- Rice husk is very different.
- Most importantly, you can't get stuff collected for free, or shipped for $10.
It's not about making it "harder" or "easier to get patents". That's not what increasing the time that reviewers have to look at patents does.
There's also a distinction between reducing the procedural difficulty of applying for patents and reducing the "real" difficulty of making a legitimate novel invention. You don't seem to be making that distinction, but surely you agree it's relevant?
straw can be substituted for sawdust
would need to be finely ground
maybe you are paid to take it off their hands
No, straw / corn stover / etc is not free. Sugarcane bagasse is only worth about its value as fuel, tho.
mostly just have to pay for shipping which is between $10 and $20. Then you mix with water so you're down to $1 per cubic yard
check your math there
Pure Ice is up to 25MPa under compression
Not if you care about creep.
maybe you can get to 50MPa with the right Pykrete
No, fiber reinforcement mostly affects the tensile strength.
Concrete is listed here at $170 per cubic yard
that's poured, not raw material costs
UHPC is about 10x that
No. Maybe for overpriced proprietary mixes with >100 MPa, but we were talking about 50 MPa concrete.
Pykrete is much weaker than concrete. You need to look up actual numbers for both tension and compression, in MPa. Concrete with >50 MPa compressive is common now.
Pykrete is not 50+ times cheaper than concrete. Sawdust is now something like $40/ton + say $20/ton delivery. At 14%, that's $8.40/ton for the sawdust alone. Portland cement is like $130/ton and you only need 15% for concrete, so around $20/ton for the concrete component. Then you need aggregate, but with Pykrete you also need to deal with freezing it, and that's relatively difficult because it has low thermal conductivity.
Would you patenting more things actually lead to you doing something to turn the patented inventions into products that you currently aren't doing?
That's not the right question, because I'm not the one with the power to turn them into products. There are people who have power - money, executive positions in companies, government positions, media influence, etc - that could be used to accomplish (part of) that. The question is whether it would lead to them doing something to turn the patented inventions into products.
Most of the cost of patents usually comes from patent lawyers, anyway. Yes, the US patent office will probably have to raise its fees at some point, but it's also had trouble hiring enough examiners, and not just because of the salaries offered.
I could imagine some billionaires being willing to pay a lot of money for cloned extinct animals, if they can do it.
Varda, they're trying to make ritonavir crystallize in a different way, but even if they can do that in space, and even if there isn't an easier way to do it...the crystals aren't the active form of the drug, it has to dissolve before it can do anything. If you want it to dissolve faster you can use smaller crystals, and if you want it to dissolve slower you can encapsulate it. It's totally meaningless. Earlier, I think they were trying to make ZBLAN optical fibers, but the only reason they were supposedly better in space was because there were fewer particulates than in labs without air filters, and also they're not actually better than current optical fibers in practice.
I suppose I'd say that without astronomical observations showing accretion disks and gravitational lensing without emission within an event horizon, the existence of black holes would be theoretically justified by general relativity but we wouldn't be able to make strong statements about GR holding in such extreme conditions.
You should decide if you believe it or not.
No, I don't think people should start by deciding if they think that, eg, black holes have internal structure or not. That's backwards.
I don't consider that bound a "result", just a "part of a hypothesis" or "implication of a speculation". The word "result" means, to me, something that follows from the data of experiments.
The Bekenstein bound? That doesn't make any testable predictions, it's just a calculation of some theoretical implications of a theoretical model of black holes. I don't see why I should count that as evidence of anything in particular.
It's not certain that there's a good reason to try to quantize gravity in the first place. The Standard Model says other forces have carrier particles, but the whole reason that's the dominant view is because W/Z masses were successfully predicted, and I don't think it can be definitively said that the forces exist because of the particles rather than particles of those masses being (briefly) stable because of those forces.
There are some other proposed approaches as well.
The main point of this post wasn't to explain superconductors, but to consider some sociology. What I thought people in your position would do is click on the "Cooper pairs" link and see the references to phonons in that Wikipedia article. Phonons are definitely a real thing, broadly applicable enough that neither of the articles you mentioned are related to superconductors. I previously wrote about phonons here.
"Pink noise" refers to a noise spectrum, but "1/f noise" often refers to a source of noise also called "flicker noise".
Anyway, I'm glad you got something out of my post, and your reply might be helpful to readers here.
The introduction was setting up an apparent contradiction to be resolved later. It's obvious that any apparent contradiction must have a way of being resolved.
Walking and swimming animals don't have near-zero benefit from sensing magnetic fields. And some of them - but only a fraction - do have some (seemingly less-accurate) way of sensing magnetic fields.
Primates have 2 arms and 2 legs because they evolved from animals with 4 legs and bilateral symmetry.
Right, but presumably sensing of magnetic cell clusters developed in one location first, probably the beak because that wouldn't require specific pigmentation patterns. Then that structure was duplicated in the eye and developed pigments. It's possible the genes originally came from horizontal gene transfer, maybe an endosymbiont.
Signals coming from mechanoreceptors doesn't mean they're perceived as a "tactile sense". Different senses produce nerve impulses the same way; they're just interpreted differently.
Designing new technology seems to be what I'm best at, but I've often asked myself:
What's the point of new technology if all it does is give institutions more room to be corrupt?
Computers got much faster, but software and web pages bloat until they're slow, until there's enough of a problem that management gets pushback and has to make changes they otherwise wouldn't. More technology can improve the results of a small amount of effort, but any amount of technology multiplied by zero effort from leadership towards a goal is zero.
This post probably should have mentioned that Paul Nakasone is a former NSA director.
A more cynical interpretation of this news is that it represents a deal that gives OpenAI favorable access to US government contracts and some protection from safety-related regulation in exchange for ensuring the NSA will have access to user data going forward.
Also, polycrystalline diamond (PDC) bits are pretty common these days.
That's theoretically possible, but how do you install them? Power them? Deal with abrasive particles and lubrication issues?
What people have decided is more practical is: have a big bucket in the hole, and have compressed air blow the cuttings into the bucket. Then it's periodically lifted up and emptied. But liquid drilling fluid has other advantages, like balancing pressure down in the hole.
Of course, with microwave drilling you can't use liquid and need a large mass flow of air for cooling.
When rock gets very hot in a microwave beam, it expands. Pressure could get very high. Will it quickly flow outwards, meaning that removing the dirt isn't needed?
I'm not sure what that's supposed to mean.
will use a less efficient form of 1 hole geothermal when you know that most geothermal plants use something more efficient?
You're suggesting fracking between 2 holes? With supercritical water? In rock that's hot and pressurized enough to have a little plastic flow?
You guess at a way things might be done, do calculations, and declare it impractical. (The calculations are for a naive / stupidly designed version of the tech that is indeed impractical.)
I am probably better at designing drilling systems than the people at Quaise Energy. In general, my problem has been overestimating how competently startups are doing something. I've actually designed multiple things for drilling, such as electric DTH hammer systems; I just don't post such things on my blog.
Your arguments in “energy payback” apply to any form of geothermal energy.
No:
- Drilling normally uses much less energy than vaporizing rock.
- Using liquid drilling fluid uses much less energy than compressed air for deep holes.
- Pumping out existing hot water isn't limited by thermal conductivity, and conventional geothermal does that.
All of your calculations assume that the rock has to be vaporized. I don’t see why it wouldn’t suffice to melt the rock, or even just heat it quickly enough that it shatters while remaining in the solid state.
Quaise Energy is a startup that is specifically about vaporizing rock with microwaves.
Melting rock doesn't help get it out of a hole. (What are you going to do, try to pump it?) And thermal cracking is pointless when you can use a normal drill. But yes, you can drill rock while keeping it in the solid state.
I don't think so. How about if what people want depends on the current weather, or if people end up wanting whatever they were able to get in the past?
- Lobbyists influence politicians, who decide policies based on politics. They then find economists who will justify the things they want to do. Those economists become high status, and other economists imitate them.
- Appointment of economics faculty has been influenced by big donors with ideological agendas.
- The predictions have been bad, even beyond the difficulty of predicting market movements. "Transitory inflation", for example.
- Macroeconomics is mostly about making a model and working out its implications, and selects for people good at doing the math involved, but there are no pressures towards making sure the assumptions are actually correct.
- see this post
People have been trying to do this with a tiny worm. That project has not yet been successful.
I understand why MIRI has Yudkowsky, Bourgon, and Soares as "spokespeople" but I don't think they're good choices for all types of communications. You should look at popular science communicators such as Neil deGrasse Tyson or Malcolm Gladwell or popular TED talk presenters to see what kind of spokespeople appeal to regular people. I think it would be good to have someone more like that, but, you know...smarter and not wrong as often.
When I look at popular media, the person whose concerns about AI risks are cited most often is probably Geoffrey Hinton.
There are rounded N95 masks, too. The important part is probably the behind-the-head straps.
The greater force means a better seal. But that seal comes from your face deforming to fit a rigid mask, and your face would prefer not to be deformed.
It's better to use soft polyurethane or foam to seal around the edges, and that's what some P100 masks do. But because they have higher upfront costs, they have cultural associations with industrial work and gas masks, while people use disposable N95s for things like woodworking. Also, the typical design with 2 flat filters to the sides makes it harder to talk to people while wearing them than disposable N95s do.
Here is a typical disposable P100 mask. It adds a foam seal around the rim, and it's much more expensive than a N95 mask. I doubt the production cost is that much higher, and in theory, the foam seal could be separate from a replaceable filter.
Yes, but Zvi's earlier posts were more positive about Altman. I just picked a relatively recent post, written after the board fired him.