All of Dr_Manhattan's Comments + Replies

Any updates on the API? (thinking of) Playing around with interesting ways to index LW, figure there should be something better than scraping

So how does one invest in China, as a country?

He was only a de facto mysterian: thought mind is so complicated that it may as well be mysterious (but ofc he believed it's ultimately just physics). This position is updateable, and he clearly updated.

A net saying "I'm thinking about ways to kill you" does not necessarily imply anything whatsoever about the net actually planning to kill you

Since these nets are optimized for consistency (as it makes textual output more likely), wouldn't outputting text that is consistent with this "thought" be likely? E.g. convincing the user to kill themselves, maybe giving them a reason (by searching the web)? 

I've been wishing for someone to write AI-singularity parallel of Bardbury's Martian Chronicles (which are pretty much independent sample/ simulations of how living on Mars could go)

Sharing a personal weird trick why not. I like falling asleep to light TV (via iPad). I watch short shows that a) I like and don't think are boring b) I have seen before. Usually 10 minutes into a 20 min show I'm ready (Futurama is my favorite for this + my meme game is much improved by this)

Was thinking about you! Glad you made it out. Feel free to DM if I can be of assistance

MIRI is bottlenecked more on ideas worth pursuing and people who can pursue them, than on funding

Ideas come from (new) people, and you mentioned seed planting which should contribute to having such people in 4-6 years, seems like still a worthy thing to do for AGI if anything is worth doing for any cause at all (given your short timelines). If you agree what's the bottleneck for that effort?

Related work: 
Show Your Work: Scratchpads for Intermediate Computation with Language Models
https://arxiv.org/abs/2112.00114

(from very surface-level perusal) Prompting the model resulted in 
1) Model outputting intermediate thinking "steps"

2) Capability gain

Koller & Friedman

They primarily & extensively statistical graphical models, not causality (but have a chapter on it)

Since comments get occluded you should refer to an edit/update somewhere at the top if you want it to be seen by those who already read your original comment.

Yes, I would! Any pointers? 
(to avoid miscommunication I'm reading this to say that people are more likely to build UFAI because of traumatizing environment vs. normal reasons Eli mentioned)

Double-masking

Wait, you can't get N95 or KN95 there? 

thought this was interesting

https://twitter.com/HealthAllegheny/status/1291466180386062340?s=20

Sounds like vaguely-good conclusions from my pattern-matching experience but very poorly argued, with much overloading of "impatience" and many cherry picked examples. Surprisingly bad quality from you. Also, "patience" is a great virtue, context matters a lot.

Ok. To clarify, one of them is to blame. Maybe it's not the CDC. History will tell.

Most obviously, blaming the CDC for the FDA and HHS not allowing 3rd party detection kits is somewhere between false and misleading.

Please support this claim. It seems obvious that they shat the bed (don't know which agency, let god sort them out for now, history and FOIA requests will sort them out in the future). It seems obvious from reading the news that many many local and commercial labs would have been ready with capacity a lot sooner than they are if FDA/CDC/HHS conglomerate got out of the way sooner.

It's quite plausible that this is due... (read more)

This is useful in case you have facing a choice of riding it out at home and going to a hospital with high probability of getting infected if you're not already. E.g. if you have fever chances are still high you're just experiencing regular flu, and should not go to the hospital, but if your oxigen starts dropping into the danger zone you need to go.

Potential method of coping: disinfecting room. Unpack the stuff in protective gear, then after unpacked blast it with UV light?

But it's possible we could even isolate there in the in-law unit.

I thought this was the pretty clear cut answer before you wrote it. Totally endorse. Wear masks on the flight if possible. Ask your parents to stock up or start sending prep packages there (Amazon, Costco delivers)

The first question for me is are people starving in Wuhan due to the outbreak?

Answer is no, as of now, though food situation is uncomfortable. (my wife has relatives there she's in contact with). Trucks come to apartment complexes and people pick up.

I'm not sure the analogy translates well to US though. For better or worse us people are less organized. Also large % population live in suburbs where such deliveries are not feasible.

OTOH we have an excellent general delivery system in Amazon, UPS etc.

I'm slightly worried.

There's some but not a lot of interest in this topic on LW; I have a mailing list with primarily rationalist types on the topic; PM me email address to be added

Data always says something unless it's randomly generated. At the very least Chinese data provides lower bounds on some things. You can get somewhat better estimates if you model their incentives (though the lying will greatly increase the uncertainty and complexity of any model)

There's the opposite effect: the early diagnosed cases tend to concentrate on the more obviously serious ones (more likely to die).

https://twitter.com/robertwiblin/status/1222577306515976192

It is also good to invest in improving ones’ immune system by health food, vitamins, light therapy, as it is our best protection of the virus. Evacuation into a cold county house would weaken the immune system.

How much can one "improve" one's immune system by these methods in a short time? Is there any data to back this up?

In general agree with the rest. In worst-case scenario ability to self-isolate for a while ("bug in" in prepper lingo) seems worthwhile.

What do you use to listen to pdfs?

On the one hand this post does a great job of connecting to previous work, leaving breadcrumbs and shortening the inferential distance. On the other hand what is this at the end?

But one thing I'm pretty sure won't help much is clever logic puzzles about implausibly sophisticated Nazis.

I have no idea what this is talking about.

[edit] It also seems like this is the sort of thing that marketing pretty strongly encourages misrepresentation of. "All children are above-average," in that the restaurant wants to present itself as serving healthy, cheap, tasty food, while also paying its employees well and having good returns for its investors. But several of those variables are in direct tension with each other, and there's not great language for speaking publicly about the tradeoffs you're making. 

Couple of reasons spring to mind:

• Marketing leverages the Hal

The way I understand the objection it that YC promotes "building great products", which attracts (a lot of) certain kinds of founders, but in fact YC is optimizing for something else (primarily described in Black Swan Farming, confirmed by other sources). I believe they are quite value-additive to the companies they accept, but attract more founders than if they were "honest about their optimization function", where some founders could have been better off engaging with other VCs on possibly better terms.

>> Fiddly puttering with something that fascinates you is the source of most genuine productivity. (Anything from hardware tinkering, to messing about with cost spreadsheets until you find an efficiency, to writing poetry until it “comes out right”.) Sometimes the best work of this kind doesn’t look grandiose or prestigious at the time you’re doing it.

Hmm, I use to spend quite a bit of time fiddling with assembly language implementations of encryption code to try to squeeze out a few more percent of speed. Pretty sure that

I think in part these could be "lessons relevant to Sarah", a sort of a philosophical therapy that can't be completely taken out of context. Which is why some of these might seem of low relevance or obvious.

Fiddly puttering with something that fascinates you is the source of most genuine productivity. (Anything from hardware tinkering, to messing about with cost spreadsheets until you find an efficiency, to writing poetry until it "comes out right".) Sometimes the best work of this kind doesn't look grandiose or prestigious at the time you're doing it.

http://paulgraham.com/genius.html seems to be promoting a similar idea

I'll claim LW priority for pointing to the idea (but not to elaborating it in a post) https://www.lesswrong.com/posts/isSMDR8rMr5pTzJK5/example-of-poor-decision-making-under-pressure-from-game?commentId=YzvDcA357NboxD2fE :)

Ross https://web.stanford.edu/~shachter/ uses something like this to score answers to (appropriately) Decision Analysis homework questions. (Don't remember the exact rule, but the intent was the same)

What does it imply for things like AI governance and global coordination on x-risks?

I've read the article a while ago, and vaguely concluded there should be some implications here (but largely uncertain about the direction or magnitude, being a non-expert). Interested to hear what people think (esp. people who concentrate on policy)

How about we let go of success, but keep doing challenging stuff anyway, just for the fun of it?

This sort of feels like Feynman's attitude, despite him being extremely successful.

Also notable: NVIDIA trained a half-order-of-magnitude larger model https://nv-adlr.github.io/MegatronLM?utm_campaign=NLP%20News&utm_medium=email&utm_source=Revue%20newsletter

Seems way off from the actual release; any post-mortem?

A decent solution to the "who should you be accountable to", from the wisdom of the ancients (shows thought on many of the considerations mentioned)

When in doubt, remember Warren Buffett’s rule of thumb: “… I want employees to ask themselves whether they are willing to have any contemplated act appear the next day on the front page of their local paper—to be read by their spouses, children and friends—with the reporting done by an informed and critical reporter.”

I think the Hypothesis is not about Open Threads specifically

Tracking employment/location and publishing/conference attendance records of researchers will probably be good source data for this.

Small typo:

Hence it has no motivation to manipulate[d] humans through its answer.

IN MICE

I somewhat overlooked this line and yes, it's a nod in the right direction

Based on the transcript this does not sound like a FOOM discussion (as in rapid self-improvement) other than mentioning "group learning" by autonomous cars, which is maybe somewhat related. Also the pregnancy ad story is much more about pattern recognition with lots of data than any serious AI.

Basically JP is, in this area, a complete layman (unlike Gates, Musk, or, from the other side, Pinker) whose opinion counts for little and not talking about FOOM anyway.

Is this much different from Scott Adams' advice https://dilbertblog.typepad.com/the_dilbert_blog/2007/07/career-advice.html

of

if you want something extraordinary, you have two paths:

1. Become the best at one specific thing.
2. Become very good (top 25%) at two or more things.

?