Posts
Comments
I'd be interested to know how many people flunked out of that internship because they couldn't pick it up, and to what extent people were pre-selected for the internship based on some estimate of their ability to pick it up.
This would mean that a hypothetical AI "uniformly" gaining capability on all axes would beat us at math long before it beats us at deception.
I'm pretty skeptical of this as an assumption.
If you want an AI to output a useful design for an aligned AI, that design has to be secure, because an aligned-but-insecure AI is not stably aligned, it could be hacked. Ergo, your oracle AI must be using a security mindset at superhuman levels of intelligence. Otherwise the textbook you'll get out will be beautiful, logical, coherent, and insecure. I don't see how you could make an AI which has that level of security mindset and isn't superhumanly capable of deception.
I've been attempting to use IFS for years without having read much more than brief summaries of it. This post put me on a much firmer footing with it and I was able to much more clearly categorize a bunch of things that have been happening over the past six months or so. Then over the weekend I had a low-level background internal screaming going on, and while my first couple rounds of attempts at resolving it only helped a little, I was finally able to isolate the issue and fix what turned out to be a massive misalignment. I have not felt this aligned in years.
So thank you very, very much for writing this.
For the specific case of weighing yourself, could you create a scale that only gives the positive reward, not the negative one? Like, it only tells you your weight if it's lower than yesterday, or better yet if the trend in your weight is downward over the past week? Maybe it displays a cheerful message and plays a soothing sound when you weigh yourself, and it emails you later at random if you've been losing weight.
My fiance likes it, supposedly it's been very true to the original manga. It might just be better than the original dub.
At this point, I won't be confident that i've been successfully preserved until ultra high resolution electron micrographs of my brain are in Amazon's S3 storage, replicated across multiple regions. Any storage that doesn't have redundancy doesn't count as safe.
Is Austin on the list? I work at a not-evil tech startup called SchoolAdmin that does school admissions software for a mix of public/private/charter schools. We're not hiring devs right now, but that might possibly change since we have a product manager coming in October. The company is REALLY not evil; we've had three different people come down with mental or physical health issues, and the president's mantra has been 'your job is to get better' in every case.
I could possibly also offer a place to crash, I've got a futon, a study it could be moved to, and already have cats.
Body armor only protects against weak curses like Somnium that are standard first year hexes. Stupefy, which is castable by first years, is not blocked by armor.
He got Lucius to agree to proposed next steps which we don't know about and secured an alliance with House Malfoy.
Also, when he added the bit about exonerating House Malfoy, he showed Lucious one wording, then wrote down his original wording. He might have changed something important, and Lucius might not have noticed.
Any chance you can refactor your recipe to replace bananas with something that keeps better? Otherwise, I was pretty pleased with it.
I just made my first batch of this and it was pretty decent. Not amazing, but definitely palatable. I had one big cup in place of dinner; I expected to want more but it was surprisingly filling. I started with the base recipe and added a scoop of vanilla whey protein, plus about a quarter cup of frozen chopped spinach.
I was expecting to have more issues with storing or measuring out the frozen OJ concentrate, but I just opened it, scooped some out with a spoon, put the lid back on, and stood it up in my freezer.
I finally started brushing regularly when I finally tried an electric toothbrush and began brushing immediately after I showered at night.
I tried Metamucil packets a while back to help alleviate side effects of the antidepressants I was on, and the trick with those was to drink them reasonably quickly after you added water, before they turned into gel. With the single-serving packets, you could try adding the fiber right before drinking.
I was browsing through discussions of that blog post on reddit, and someone pointed out that his formula has inadequate fiber. One batch has 1500ish calories from macronutrients (carbs/protein/fat), and only 5g of fiber. The daily recommended intake for fiber is 38g per day for an adult male, so if you're only consuming Soylent, you'll want to increase the fiber per batch dramatically.
It works great. I mentioned wanting a standing desk to my boss when I started a month ago, a couple of other people expressed interest, and he bought four of them, including one to use himself. It sits on the desk that's built into my cubicle, my laptop sits on the shelf, and the monitor sits on top. The boss had to send someone to IKEA to get another four.
We might need to get higher cubicle walls, though.
Switched to a standing desk at work. I have one at home and I mentioned the idea to my boss. He eventually found a post where the right $22 at IKEA lets you build a desk -> standing desk converter. It's basically an end table with a shelf screwed to the side; since you are making the screw holes yourself, the shelf is at whatever height is ideal for you. He decided to try it out also, along with a few other people, so we shall see how the grand experiment goes.
It's been a few hours and I have already discovered that standing for long periods in tennis shoes is not quite the same as standing barefoot. If I don't adapt, I may need to buy new shoes.
I was on an SSRI so I'm not sure any of my experience is actually relevant to bupropion.
If your depression has an obvious cause, fix that instead. I was depressed because of grad school, and I got better when I graduated.
Are you sure you're not just building up a resistance/dependence? I tried anti-depressants but they eventually stopped really doing anything, I believe somewhere between 6 months to a year after starting them. I think resistance is pretty common.
Also, most anti-depressants take a while to kick in, so I suspect any day-to-day dosage changes are going to be more about withdrawal symptoms than anything else.
Yeah, memory is fallible as hell. This is why I love having conversation logs and why I have contemplated trying to figure out a way to log my entire life (so I could do that for real conversations as well).
It takes her a few seconds to remember the Asch Conformity Experiment and that was a long enough delay to be frightening.
Wizarding society likes to let people figure out the dangerous secrets for themselves. You don't tell people the dangerous secrets until they have proven themselves on the easier ones, you don't tell people the secret of potion invention because they might get turned into cats, etc.
Of course, Harry can violate that as he pleases if it is just a social convention, and Harry's guesses at principles seem\ to hold up far better than it seems like it should.
Neville is the most likely next target for an attack intended to separate Harry from his allies. Voldemort is probably too clever to try the same trick twice, though.
On the last point: Hermione almost certainly was false memory charmed twice; H&C would have removed the memory of their final conversation and replaced with with something innocuous at the same time as he implanted the false memory of casting the blood-cooling charm, so as to not leave a suspicious gap. He might also have implanted false memories immediately after the groundhog day attack, either to cover up the time gap or to not have Hermione wandering around with an extremely suspicious memory in her head (If Dumbledore or Snape had seen the memory of H&C, or his less-creepy disguise, they probably would have been fairly suspicious).
I'm not sure if anyone has commented on this, but I just noticed it while rereading the Self-Actualization chapters:
Hermione went to tremendous lengths to be her own person rather than just something of Harry's, including becoming a general and fighting bullies. Now she has sworn herself into Harry's service and house forever. That is really sad.
The dementors serve at least three purposes in Azkaban: they drain the magic from prisoners to render them helpless, they notify the guards when prisoners escape, and they chase down and incapacitate escaped prisoners and intruders. If Harry destroys 90% of the dementors, there probably won't be enough left for the first or third purposes. That would make Azkaban much less secure, and the perception of Azkaban's security would go down if there are hardly any dementors since the dementors are what make it infallible. Even just demonstrating that Dementors CAN be destroyed would probably force them to completely remake Azkaban to not depend on the dementors.
The dark side is presumably the result of the botched Horcrux creation ritual and is in some way an aspect of Voldermort's mind or soul. An AI might have different modules for emotions and computation, but a human mind is not so cleanly separated.
Ah, but Harry doesn't intend to kill Dementors in particular, he aims to eradicate death itself (destroying them indirectly) and he is NOT confident that he will accomplish that in his lifetime. A Dementor that pisses off Harry dies immediately, while a Dementor that doesn't will only die if Harry lives long enough to succeed.
Email sent!
I think people would appreciate knowing that they might still have a shot even if they haven't heard anything. You could maybe ask Eliezer to put a note that you're still wading through applications in his next HPMOR author's note. =P Otherwise, I think a mass email would not be too annoying to those who have already heard from you and very much appreciated by those who haven't.
Yeah, the early stuff in Jaynes is pretty comprehensible (the ideas are clear if not all the proofs). Intro stats classes tend to be very light on the proofs, though. They're very much "here's probability", not "here's why probability". I'll definitely reread Jaynes again before teaching, but I want to finish Bolstad and work through some of the problems before that.
I do need to read up on those; Jaynes talks about the implications of Cox's theorem but doesn't go into it directly, so I'm only vaguely familiar. Thank you for the reading suggestions. I did plan to talk about those issues in the introduction of the course. Bolstad has an intro section justifying the Bayesian perspective, as well.
I think I picked that particular set of justifications because educators in general don't care about mathematical proofs, they care about what will be useful for the students to know how to do; in biology, the point of knowing statistics is to be able to read and write scientific papers, and the vast majority of papers are written using frequentist statistics. Proofs will not convince them; the fact that top professors are using Bayesian methods might.
My expectation was that I would replace a mediocre frequentist statistics lecturer with an excellent bayesian statistics lecturer within the same class. The class that I TA is taught by multiple professors, and at least one of them teaches from a Bayesian perspective. Professors have ridiculous academic freedom; one professor covers only basic t-tests, while another professor covers everything from linear regressions to the KS test to chi-squared to two-way ANOVA, and it's still the same course listing. So long as the students aren't complaining about failing, the university does not care. The students can try to sign up for a different professor, and will do so if they hear another prof is easier, but they still have to take the class, so even harder professors still have full sections (especially if their version has a reputation for being very useful/educational).
So, assuming that I would be hired to teach statistics and could choose to teach either frequentist or Bayesian, I see very little point in teaching frequentist. I could also reach vastly more students lecturing than I could via tutoring, probably 80ish vs 10ish.
I think the students that are interested in learning Bayesian stats should have the option available; I think there are probably a fair number of students who are smart, savvy, and motivated enough to sign up for a stronger stats course but aren't quite good enough to teach it to themselves.
I think I would almost rather not teach statistics than teach straight frequentist. I am really sick of teaching kids stuff that I know is suboptimal. I mean, I could do a good job of it, but raising the waterline isn't worth being miserable.
This paper consists of some vague simulations, followed by wild speculation. I'm pretty sure it's bunk (speaking as a computational/cell biologist). It would be pretty easy to test, as well, as disrupting microtubules AT ALL would completely destroy memories if he is correct.
Here's another idea: Draco uses his Patronus to tell the assembly he forgives the blood debt. Harry can use his own Patronus to beg Draco to do this.
If "A potion spends that which is invested in the creation of its ingredients." then what the heck goes into an Animagus potion? Something that's been transfigured a lot?
Edit: It doesn't seem like the method for becoming an Animagus is described in canon, so the potion aspect might be new to MoR.
If potion invention is slow, Harry must have gotten the light potion from a book, since I don't think there's enough time between battles to do serious potion research safely between classes and homework, even for Harry's 30 hours a day. If he can invent potions that fast, he potentially has a huge number of instant win conditions available (that's what I really meant, that rapid potion invention would be a huge pain in the ass to write around). I think at this point it's clear that Harry probably does know enough to invent potions, but not without probably months or years of experimentation per new recipe. If he didn't know enough to be dangerous he wouldn't have freaked out Flitwick.
Harry comments at some point that "He'd noticed the correlation between the effort Professor Quirrell expended and the time he had to spend 'resting'." (74)
Harry notices after Azkaban that Quirrell looks older (65).
What I meant was that it seems like Quirrell has spent more and more of his time active using his body as little as possible. Maybe we've just seen it more because he's hid less from Harry? In the most recent battle he talked and made the tiniest possible shrug but otherwise didn't move at all. When he was grading papers he did it purely by magic as well. Whenever he can let his body sit around and not move, he seems to try to do that.
My theory is that potions which don't involve magical ingredients are obscure because they're usually less powerful and because they require a greater investment of energy from the creator to do the reshaping (explaining why Harry doesn't do very much in that battle). Given that Flitwick and McGonagal had suggestions of books to make at all after hearing what Harry wanted, it seems very likely that such potions do exist, just not in the standard textbooks. It seems very likely that Harry got his potion out of a book, because potions research is dangerous and presumably very time consuming, and because Harry with the ability to invent potions would be powerful enough to wreck the story.
The catatonia appears to be getting worse and worse over time. Channeling strong magic through Quirrell accelerates the decay. I suspect he'll crap out as a host by the end of the school year, and that's with Quirrell being reasonably conservative of his energy.
I'd like to point out that after Azkaban, when Quirrell tries to talk Harry into his next plot, Harry refuses by citing what Hermione and Draco would say. Quirrell sits there and thinks for a really long time, and asks if Harry really cares about what they think. My guess is that right then and there is when Quirrell decides to take them out.
What evidence is there that H&C isn't just Quirrell wrapped in an illusion?
There's no need for Hermione to have cast the lethal hex. She wins the duel, then the real perpetrator stuns both of them, hexes Draco, and then memory charms Hermione into thinking she did it. However, if that's the case, unless the perpetrator then used Hermione's wand to cast the hex, checking what spells her wand had cast would reveal something fishy.
Why are we proposing the H&C is not clever and powerful?
The terms of the challenge state that she can't tell anyone about it before or after the duel or it goes to the Wizangamot. So, no, presumably she can't tell Harry about it.
Heck, she might have severely injured Draco by accident, rendered basic medical care, and then just left, because she can't tell anyone. If someone found Draco unconscious and half-dead later, and they figured out Hermione did it and left him, that would look like attempted murder.
Especially when the person in question has been fighting a lot of bullies lately AND is royally pissed off.
It's possible that H&C never did figure out an effective lever. In that case, he might have given up on Memory Charming her (requires the target's defenses to be lowered, at least in the case of an experienced Auror and possibly for a very pissed off first year) and just oblivated her. If he'd managed to memory charm her, I don't think she'd have been so freaked out. She also wouldn't have 'lost track of time', she'd have had a perfectly reasonable legitimate excuse put in place.
Also, have we seen Quirrell use Legilimancy at all? If we have, that's an argument for H&C not being Quirrell, because if you've thoroughly read someone's mind you it shouldn't take that many tries for a groundhog day attack.
When you're going to Obliviate the target anyway, there's little downside to letting some frustration slip through. I don't think that necessarily counts as a screwup.
I think you would need a remnant of the destroyed dementor itself, not just a cloak a dementor happened to be wearing when you killed it, and I don't think dementors leave anything behind when you kill them.
The fact that the light was impossible to Finite suggests that Harry did tap the energy of the acorns. It's implied that the magical cost to the creator of making a potion is a minor cost to reshape the components. So, the potion taps the light stored in the acorns, and Harry's magic is tapped only to do the reshaping. Probably most magical potions use the magic of the magical ingredient to do most of the reshaping work, so the user only has to invest a tiny bit of magic, while a potion not involving any magical ingredients might require much more input from the creator for the reshaping. That would explain why Harry is drained, but also why the light can't easily be dispelled.
The other critical limitation on potions is that you must known the stirring pattern and the recipe in general. Figuring out the stirring pattern is the sort of thing that gets you permanently turned into a cat. So, Harry does not have god-mode because he doesn't have the time or expertise to do all the potion experimentation necessary to invent new potions without blowing himself up; he's limited to potions with known (but possibly obscure) recipes.
Just because every potion in the two textbooks Harry looked at involved magical ingredients doesn't mean all potions require a magical ingredient. As I read it, Harry found the potion he used in a more obscure book suggested by Prof. McGonnagal or Flitwick, probably something like a wilderness survival guide. Converting acorns into a beacon would be pretty helpful for getting found by search parties.
Does that include biology? If so, I will be rather annoyed that no one warned me.
If you go straight from undergrad to a PhD program, you won't have a Masters on record, which means you'd have to drop all the way down to applying for bachelor's level positions.
Plus which, if you're not going to leverage the PhD, why would you spend an extra four years of hard work and low pay to get it? Just get a Masters instead.
Having a graduate degree somehow makes my resume look less attractive than just having an undergrad.
In many cases you have to pay PhDs more. If you can find someone who can do the work who doesn't have a PhD, you save money by hiring them. In many fields, there is much more need for people to do Masters level work than to do PhD level work, so there are more jobs available at the Masters level.
grad or professional schooling can be more time/effort/work than it's worth.
A PhD in biology in the program I mastered out of involves working anywhere from 50-80 hours a week, takes 5-6 years, and typically pays around $30k a year at best. If you could have instead made $50k at a real job, graduate school has an opportunity cost of $100-120k.
Graduate school is bad for you. Most of my friends are in graduate school and they are noticeably less happy and more crazy than they would be otherwise; one friend is going gray at 27 from stress. I was chronically depressed almost the entire time I was in graduate school, and it only went away when I graduated.
One of my fellow teaching assistants last semester was a first year graduate student, taking two graduate level core curriculum classes, trying to start up his research, and also trying to be a teaching assistant for biostatistics when he had never taken any statistics. Both of my fellow teaching assistants this semester are just getting Masters, but they're both taking two classes, TAing biostatistics, and trying to finish their Masters theses.
Do undergraduate research first before making any decisions about graduate school. PhD programs are almost entirely research. You take some classes at the beginning, which weeds out the idiots, and then the entire rest of the program is research. If you don't thoroughly enjoy research, you will hate graduate school.
My standard advice is to not do a PhD unless you have to (because the career you want requires one).
It looks like they're meeting on Sunday: http://lesswrong.com/r/discussion/lw/6bz/meetup_houston_tx/