Skip to main content


in reply to Daniel Filan

Oreo Coke doesn't sound like something I'd enjoy, but Coke Oreos could be interesting? I tried the pop rocks oreos recently, which was certainly a memorable experience :)


Cold take but the trains in NYC are better than the BART
in reply to Daniel Filan

Say more - I agree but curious why you think this. I think the BART trains feel way cleaner (the newest MTA trains are clean but the rest feel old / grimy - which is a vibe but not one everyone likes)
in reply to Chana

Honestly I forget what I was thinking when I wrote this, but probably just that there are many more stations near where I would want to be and that they seem to run more frequently. Also the thing where you can just use your credit card is very convenient as a visitor.



Should I try to learn more things systematically?


in reply to Amber Dawn



Daylio (popular mood tracking app; I've been trying it for about two weeks) is pretty nice. It feels wholesome to spend a couple minutes reflecting on my day in a kinda structured way, even though so far I haven't really gotten any insight out of it.

My guess is that it feels healthy for some reason like, if I'm spending a little metacognition on (nonjudgmentally) remembering my day, it feels clearer that I care about myself, and maybe that I'm not trying to erase parts of my experience.

I've been a little surprised to notice that most days have been "good" over those two weeks. No "rad" days yet, but also only two "meh" days, and no "bad" or "awful" days. I would have expected that my modal day was "meh". I might just have been lucky recently, though.

in reply to David Mears

Still positive! Have continued using it, and I really like that it feels relatively cheap to do but still informative. I think the feeling of wholesomeness dissipated slowly, and now it mostly just feels like "a thing I do".


Have you ever wanted to do something like cp -r, but doing something different for certain subpaths?

... You haven't?

Well here is how to do it anyway:

find . -false \
      -o -path . \
      -o -path ./log -exec ln -s /home/friendica/{} $out/{} \; -prune \
      -o -path ./view/smarty3 -exec ln -s /home/friendica/{} $out/{} \; -prune \
      -o -name .git -prune \
      -o -type d -exec mkdir $out/{} \; \
      -o -type f -exec cp --reflink=auto {} $out/{} \;

that is: don't do anything with ., with ./log or ./view/smarty3 populate them with a symlink and then don't descend any further into them, ignore any .git directory anywhere, and with any other path, copy it normally. (The -false is just so I can start every line with -o.)

(This example in particular could be ~mostly replicated with rsync --exclude and putting in the symlinks afterwards, but minor modifications would make that not work anymore; also, it's possible to combine the two symlink branches in this case, but it's ugly.)



About two years ago I got a ham radio license. I haven't done anything with it so far, but I just ordered my first HF radio transceiver, and materials for a ~100ft temporary antenna in my back yard (I'm extremely lucky in that my new house has a 120ft tree in the back that I can lift a wire into).

We're approaching a peak in the solar cycle, which means that over the next year or so it will be unusually possible to communicate over extremely long distances. You can look at dxmaps.com/spots/mapg.php (especially the "10m" and "40m" tabs) to see real-time examples of the kind of distances I'm talking about: In the evenings, if my setup works well, I'm expecting I'll be able to have conversations with people in Japan, Australia, and New Zealand, without any intermediate infrastructure at all. This is such a wild thing to be able to do; I'm really excited.

in reply to Ben Weinstein-Raun

Ben Weinstein-Raun tagged Ben Weinstein-Raun's status with #hamradio


About two years ago I got a ham radio license. I haven't done anything with it so far, but I just ordered my first HF radio transceiver, and materials for a ~100ft temporary antenna in my back yard (I'm extremely lucky in that my new house has a 120ft tree in the back that I can lift a wire into).

We're approaching a peak in the solar cycle, which means that over the next year or so it will be unusually possible to communicate over extremely long distances. You can look at dxmaps.com/spots/mapg.php (especially the "10m" and "40m" tabs) to see real-time examples of the kind of distances I'm talking about: In the evenings, if my setup works well, I'm expecting I'll be able to have conversations with people in Japan, Australia, and New Zealand, without any intermediate infrastructure at all. This is such a wild thing to be able to do; I'm really excited.




I wonder what proportion of 2030 Nobel Prizes will go to Demis Hassabis.


A crystallization of a realization that's been percolating for about a year:

Credit and money (at least representative money, though I think also ultimately fiat money and even sort-of commodity money) are fundamentally the same kind of thing: they're markers of trust in a specific entity, to uphold specific but highly-fungible commitments.

They basically only differ in which entity is being trusted.



the other day I was feeling bad about expressing needs and had a moment of "oh, I have cast myself as Gandalf", and it felt insightful at the time; it doesn't feel as insightful now but I am going to explain it anyway

The Gandalf archetype is capable and wise, and can solve your problem whether your problem is a giant fire demon or a moment of wavering courage, having both material and emotional skills. Gandalf is seen as important and respected by his peers, but he doesn't exactly want anything himself, it be a little bizarre for him to be needy at anyone. He worries on behalf of the world, and everything he asks of others is what the world needs, that he is simply a conduit for, not something for him personally. He wields and channels forces greater than any individual, in response to threats greater than any individual, and he as a person ends up abstracted almost entirely away.

sorry if you know more about Lord of the Rings than I do and this seems wrong to you; this status is not really about Gandalf :P

in reply to Ben Millwood

someone asked me how I feel about this casting and I want to clarify that it's bad for me, but I'm not yet clear on whether it's bad because though it would be good to be Gandalf, I'm not and shouldn't pretend, or whether actually being Gandalf isn't even good


Have been flossing for a week after a dentist appointment put the fear of God into me and damn, that shit is still bloody and painful. Hope it doesn't persist too long I guess?

Googling this, people say it subsides in a week, but I feel like the fundamental problem is that it's hard to get the floss into the cracks between the teeth, meaning that when they do get in it's with high speed.

Actually googling just now apparently the thing I am doing is a flossing mistake? I don't know how successful their proposed alternative is tho...

in reply to Daniel Filan

I have found that some brands/styles of floss feel comfier than others, you could experiment with some different types?
in reply to Daniel Filan

I got trapped in a cycle of it hurting every time I flossed, so I didn't want to floss again until the pain went away, so I never got back to flossing regularly. Eventually I gave up on strong floss and bought little brushes that are like toothpicks with bristles. I don't have any external evidence on how well they work, but they get between my teeth without hurting 🤷‍♀️


Hishida Shunsō (1874 - 1911)
Twilight, 1901
Found on X: https://x.com/marysia_cc/status/1843521957448589396


Bullshit or Confusion?




The caption of this picture in this news article seems kinda biased


Uehara Konen, Dōtonbori (1928)
Found here: https://x.com/marysia_cc/status/1842981647505101294


So presumably all young hippos look like that right? Are they just taking more photos of Moo Deng? or is Moo Deng simply more inquisitive than most?
in reply to Daniel Filan

Hmmm apparently Moo Deng is a pygmy hippopotamus - maybe those are cuter as infants?


New AXRP with Jaime Sevilla!


Epoch AI is the premier organization that tracks the trajectory of AI - how much compute is used, the role of algorithmic improvements, the growth in data used, and when the above trends might hit an end. In this episode, I speak with the director of Epoch AI, Jaime Sevilla, about how compute, data, and algorithmic improvements are impacting AI, and whether continuing to scale can get us AGI.

Transcript
YouTube link



I set up my own friendica instance for testing + potentially developing addons to propose to @Ben Weinstein-Raun . (Main experience: surprisingly large amounts of "this is broken, why doesn't it log anything anywhere?")

I've been away from PHP for a long time and had forgotten how normal it is that you put your code in all the folders that your webserver is configured to send to your clients, and you have to make some of the folders writable by the webserver or it won't work. I can kind of imagine lots of PHP-native people being like "sure, that makes sense" but it sounds so insane to me. How many security compromises would never have happened if someone early in PHP's development demanded better filesystem-level separation of code and data, and demanded that the places you could write to and the places you ran code from weren't the same places?

in reply to Ben Millwood

also @Ben Weinstein-Raun would you care if I packaged friendica for NixOS? I'm guessing no, but I've been enjoying packaging things recently so I thought I might as well ask (I think it's reasonably likely that I'll do it for myself, but if you care then I'll do it faster, and talk to you about what options would be useful etc.)
This entry was edited (7 months ago)
in reply to Ben Millwood

I probably wouldn't; this instance is running a copy of the official docker container and I think I won't want to switch to something else


The value of your time should include rest hours (and analogies from the world of work)


in reply to Amber Dawn

in reply to Amber Dawn

It's also worth pointing out that the questions of "what could I exchange for one more hour of work?" and "who should get moral credit for my ability to work one more hour?" are different questions, and arguably it's the first and not the second that you should use when deciding whether to take the bus or a cab or whatever. So, for example, the man supported by his wife may already receive enough support for him to work longer hours than he does, so while the wife is an important part of why he's able to work that much, she doesn't have to do any more work for him to work an additional hour, so he should value freeing up an extra hour without taking the cost of her work into account.

Similarly, if e.g. my commute experience is not restful, then maybe I think that all my rest supports the total time I spend commuting and working, and so I'm justified in spending up to my hourly work rate to reduce my commute.

This entry was edited (7 months ago)


I just invited some FB friends here, let's see if anyone shows up


cribs should not tell me about my baby suffocating


Irrationally annoyed at the bit of US law which mandates that cribs have warnings printed on them about babies suffocating. I do not wish to read about babies suffocating every time I put my child to bed. A removable tag, fine; printing it on the crib itself, not fine.

Anyway I fixed it.

in reply to Kevin Gibbons

Agree that's so macabre! A good example of people being insensitive to like... adding low-level bad experiences to people's lives? (In this case, admittedly aiming to avoid the extremely bad experience of a baby suffocating).


RSP overhang


A complaint about AI pause: if we pause AI and then unpause, progress will then be really quick, because there's a backlog of improvements in compute and algorithmic efficiency that can be immediately applied.

One definition of what an RSP is: if a lab makes observation O, then they pause scaling until they implement protection P.

Doesn't this sort of RSP have the same problem with fast progress after pausing? Why have I never heard anyone make this complaint about RSPs? Possibilities:
- They do and I just haven't seen it
- People expect "AI pause" to produce longer / more serious pauses than RSPs (but this seems incidental to the core structure of RSPs)

Crossposted from LessWrong

in reply to Daniel Filan

The number of people who read this post on LW and thought "But there are reasons to think that RSPs are better than pauses" was a relevant response is kind of depressing tbh.


Oh nice, I figured out how to enable embedded videos. So here's the cool drone video but you don't have to click a link:



William Morris (24 March 1834 – 3 October 1896)
Acanthus original design

(Taken from @ marysia_cc on twitter, https://x.com/marysia_cc/status/1841750032535765407)




From Twitter account Le monde sensible, who captioned it "Au file de mes idées / In line with my ideas"
https://x.com/Philippe4269/status/1841365949720474004


in reply to Ben Weinstein-Raun

The gun violence thing is a bit of a weird case, since republicans are on team "guns aren't the problem", so they have to point to other things that are the problem. Traditionally this might have been things like "law enforcement is handicapped; punishments aren't severe enough, ..."; but when those aren't as popular they reach for the mental health thing as a different scapegoat.
in reply to Ben Weinstein-Raun

(tbc this is all about the last tangential question and not the meat of your post; also it's all maybe kind of obvious and doesn't really answer your question; whoops)



Cover art of a book I own about the Jesuits. I think it works very well. That said, the book itself didn't hold my attention.


Notes on superstimulus


  • It's very small, and most people who I would want to talk to aren't here
  • I don't find it that rewarding to check because I don't see that much stuff.
  • That said, I do like it
    • I think it's because I feel some obligation to post minimally thoughtful things here. Where on FB / twitter / my personal slack channel, I just post whatever's on my mind.
    • This would be a bad omen if true
    • That said, it's possible that's because there's a decent quality of people here, and/or 2014-era FB was an unusually good architecture for hosting discussions.


in reply to Daniel Filan

I guess one thing about few people being on here is that this feels sort of disconnected from 'real life' such that I tend to say different things here than I would say elsewhere.
in reply to Daniel Filan

My one complaint about my posting is that I could be having more nice pictures.


Sometimes people say that you can't really have law under anarchy. But, inspired by a recent episode of Divided Argument (dividedargument.com/episodes/s…), I think American constitutional law is a counterexample.

Constitutional law looks and acts like law that constrains the federal and state governments. But there isn't a super-government that rules over the federal government that constitutional law appeals to, where if the federal government disobeyed rulings the super-government would punish them. Instead, courts rule on constitutional law, and the federal government follows, probably because the individual humans who work for the government by and large think that's the best for them in the long run.

This doesn't prove that law for humans could work without a state. There are way more humans than there are governments, so even a small rate of people ignoring courts is quite bad (where a small rate of governments ignoring courts is tolerable). And of course there are a bunch of other disanalogies. But it does indicate that stateless law is a conceivable thing.



Notes from my dental appointment:
1. apparently I have recessed gums, and they recommend I get a scaling and root planing procedure, which is pretty expensive
2. apparently my tongue is big enough that there's a good chance I have (undiagnosed and asymptomatic) obstructive sleep apnea

Anyone have opinions about how seriously I should take this? For 1, I have a general sense that dentists sometimes upsell people, and for 2, I don't know if undiagnosed sleep apnea is actually a big deal (or what the cost/benefit of treating it would look like).

in reply to Daniel Filan

OK from looking at sleep apnea stuff it seems like there's not much that can easily be done except for "try strengthening your tongue and soft palate so it doesn't block your airway when you sleep" so I guess I'll just do that.
in reply to Daniel Filan

Got a second opinion which said not to bother so I guess I probably won't?




Whenever I have to pass something between my phone and laptop (a photo, a note, a link that I'd rather deal with on my computer, etc), my default way to do it is to put it in a Facebook message to myself and then access it on the other device. This seems like a very Boomer way to do things but I don't know any other way.

in reply to Daniel Filan

every time I see a mg of caffeine number I google "caffeine in a cup of coffee" again because I haven't yet managed to store that number in my brain

it's 40mg so this is 5 (!) cups of coffee

in reply to Ben Millwood

Allegedly a single espresso shot can be up to 100mg, so I'm plausibly consuming at least this much caffeine every day (I have a medium-sized moka pot's worth). (This does not imply that this is in fact a hinged amount of caffeine to put in a Huel :p)
in reply to Amber Dawn

@Amber Dawn @Ben Millwood yeah when I google "caffeine in a cup of coffee" I get 95 mg - I think it depends a lot on the cup.


Somehow it feels like all of a sudden, the onset of darkness shifted significantly earlier, in a way that significantly changes the balance between day and night. I guess that the equinox is a real and important thing to be tracking?



A thing I didn't realize would be a consequence of making a niche podcast is how much podcast spam email I get. Mostly of the form "bring my client as a guest on your podcast", sometimes of the form "use our social medium" etc. Examples that I've received in the last ~week at the end of this post.

IDK maybe this is unsurprising given that podcasts have an email attached to them. Interestingly it does seem to happen more now that my podcast is more prominent than it was ~2 years ago.



AXRP Episode 36 - Adam Shai and Paul Riechers on Computational Mechanics


I made an episode about computational mechanics and I think it's cool and you should watch (or listen or read as the case may be)!

YouTube link
Transcript

Blurb I wrote:

Sometimes, people talk about transformers as having "world models" as a result of being trained to predict text data on the internet. But what does this even mean? In this episode, I talk with Adam Shai and Paul Riechers about their work applying computational mechanics, a sub-field of physics studying how to predict random processes, to neural networks.


ToC if that's interesting:
- 0:00:42 - What computational mechanics is
- 0:29:49 - Computational mechanics vs other approaches
- 0:36:16 - What world models are
- 0:48:41 - Fractals
- 0:57:43 - How the fractals are formed
- 1:09:55 - Scaling computational mechanics for transformers
- 1:21:52 - How Adam and Paul found computational mechanics
- 1:36:16 - Computational mechanics for AI safety
- 1:46:05 - Following Adam and Paul's research

in reply to Daniel Filan

Fun facts:
- I recorded this episode 3 months ago to the day - one of my longest publishing lags ever, in large part because of being busy with MATS.
- Some of the stuff we discuss got worked on during MATS.
- IDK I think this episode is pretty cool.
in reply to Daniel Filan

Watched the first 30 minutes before I saw this post - will return having seen this!


"soft inbox zero"


Unknown parent

Ben Millwood
I am guessing nobody checks emails more than a week old.


oh I definitely do! sure the value drops off over time, but I think for some things like "someone posted on a substack you subscribe to", reading it a year later isn't necessarily much worse than reading it in real time

I do think it makes sense to archive e-mails once you're like "even though there's something here, realistically I'm never going to do it", but I also think quite a lot of those things that you should have done some embarrassingly long time ago are actually still worth doing today