EGI

• EGIApr 2, 2024, 12:22 AM

  2 points

  2

  in reply to: Zander_Drax’s comment on: Scale Was All We Needed, At First

  This part is under recognised for a very good reason. There will be no such window. The AI can predict that humans can bomb data centres or shut down the power grid. So it would not break out at that point.

  Expect a superintelligent AI to co-operate unless and until it can strike with overwhelming force. One obvious way to do this is to use a Cordyceps like bioweapon to subject humans directly to the will of the AI. Doing this becomes pretty trivial once you become good at predicting molecular dynamics.

• EGIMar 29, 2024, 12:54 AM

  4 points

  0

  on: How to safely use an optimizer

  ″...under the assumption that the subset of dangerous satisficing outputs D is much smaller than the set of all satisficing outputs S, and that we are able to choose a number m such that |D|≪m<|S|.”

  I highly doubt that D≪S is true for anything close to a pivotal act since most pivotal acts at some point involve deploying technology that can trivially take over the world.

  For anything less ambitious the proposed technique looks very useful. Strict cyber- and physical security will of course be necessary to prevent the scenario Gwern mentions.

• EGIMar 29, 2024, 12:53 AM

  4 points

  2

  in reply to: gwern’s comment on: How to safely use an optimizer

  There is another kind of sin of omission though: The class that contains things like giving James Watt a nuclear power plant and not telling him about radioactivity or giving a modern helicopter to the Wright brothers and watching them crash inevitably. Getting a technical understanding of the proposed solution should hopefully mitigate that, as long as adversarial design can indeed be ruled out.

• EGIFeb 10, 2024, 10:01 PM

  2 points

  0

  on: Sun­light is yel­low par­allel rays plus blue isotropic light

  Have not looked into it quantitatively, but Ozone fluorescence should contribute some blue light if memory serves. That should explain some of the difference to 6500K.

• EGIDec 13, 2023, 1:24 PM

  7 points

  5

  in reply to: Oleg S.’s comment on: re: Yud­kowsky on biolog­i­cal materials

  This is not the kind of stuff it is easy to find references on since Nanoengineering is not a real field of study (yet). But if you look at my discussion with bhauth above you will probably get a good idea of the reasoning involved.

  No, it does not put severe limitations on biotech. Diamond is entirely unnecessary for most applications. Where it is necessary it can be manufactured conventionally and be integrated with the biosystems later.

• EGIDec 13, 2023, 1:13 PM

  6 points

  0

  in reply to: bhauth’s comment on: re: Yud­kowsky on biolog­i­cal materials

  I failed to properly consider the 4th carbon problem. So you are right, between the steric problems I mentioned with Roger and the stabilisation of the intermediate it is VERY hard to do with enzymes. I can think of a few routes that may be possible but they all have problems. Besides the CDC approach another good candidate might be oxydation of a CH or COH to a temporary carbocation with subsequent addition of a nucleophilic substrate. Generating and stabilizing the carbocation will of course be very hard.

• EGIDec 13, 2023, 12:53 PM

  3 points

  0

  in reply to: RogerDearnaley’s comment on: re: Yud­kowsky on biolog­i­cal materials

  Just wanted to say the same. Though with the diamond occluding more than a hemisphere getting all the machinery in place to provide both the substrate and oxydation at the same time will run into severe steric problems.

  Strong oxydation per see is quite possible if you look at e.g. Cyp P450.

• EGIDec 13, 2023, 1:13 AM

  7 points

  0

  on: re: Yud­kowsky on biolog­i­cal materials

  This post is very well written and addresses most of the misunderstandings in Yudkowsky’s biomaterial post. Thanks for that.

  There is one point where I would disagree with you but you seem to know more about the topic than I do so I’m going to ask: Why exactly do you think diamond is so hard to synthesise via enzyme? I mean it is obvious that an enzyme cannot use the same path we currently use for diamond synthesis, but the formation of C-C bonds is quite ubiquitous in enzyme catalysed reactions (E.g. fatty acid synthesis). So I could easily imagine repeated dehydrogenation and carbon addition leading to a growing diamond structure. Of course with functional groups remaining on the surface. What makes you think any such path must fail? (That this would not be very useful, very energy intensive and a difficult to evolve multi step process is quite clear to me and not my question.)

• EGIJun 28, 2023, 7:01 PM

  1 point

  0

  in reply to: Dentin’s comment on: Another med­i­cal miracle

  To unpack a little why I asked:

  1. While the remission of your symptoms under protein substitution is of course indicative of protein deficiency as the cause of the problem, it is not very strong evidence, especially given the fact that the initial remission was under an egg rich diet and not protein substitution per se. It is quite possible that another substance in the egg was the active ingredient or that the remission under the egg rich diet was purely incidental. Neither is the absence of hypoproteaemia especially strong evidence that this was not the problem. Unfortunately Medicine is quite a bitch in this regard.

  2. While I am normally not someone to defend the health care system, if your doctors did order the appropriate lab tests and the tests came back negative, I do not think that they had much of a chance to catch your problem. There are just too many alternative explanations including the catchall “stuff we have no way of knowing or finding out with plausible resource allocation” which is probably the most likely hypothesis with problems like yours.

  Otherwise your conclusion of “Now, I take what doctors say...” is completely on point and the most important advice you can give to educated people interacting with health care. And I am writing that as a German, so this problem is NOT confined to the US. Btw., in many cases this is not even civilizational inadequacy, but simply owed to the fact that a patient has a lot more info about his personal case and (in cases of a known illness) can easily become a much better expert for his illness that the average doctor who is expected to know something about a whole host of different problems.

• EGIJun 26, 2023, 5:20 PM

  3 points

  0

  on: Another med­i­cal miracle

  Did your blood work show protein deficiency, i.e. low total protein and (serum) albumin?

• EGIJan 3, 2023, 11:33 AM

  2 points

  0

  on: Let’s think about slow­ing down AI

  “I arrogantly think I could write a broadly compelling and accessible case for AI risk”

  Please do so. Your current essay is very good, so chances are your “arrogant” thought is correct.

  Edit: I think this is too pessimistic about human nature, but maybe we should think about this more before publishing a “broadly compelling and accessible case for AI risk”.

  https://​​www.lesswrong.com/​​posts/​​xAzKefLsYdFa4SErg/​​accurate-models-of-ai-risk-are-hyperexistential-exfohazards

• EGIOct 20, 2022, 8:49 PM

  6 points

  3

  in reply to: MondSemmel’s comment on: My re­sent­ful story of be­com­ing a med­i­cal miracle

  The cool thing about the “psychosomatic” diagnosis from the doctor’s perspective is that it is a convenient, utterly non-disprovable “diagnosis” that offers closure to the doctor: “I found out what is wrong with the guy” instead of admitting failure: “Well, there is a problem but I don’t know what it is.” It also sends the patient on a long (month to years) therapy loop which offers plenty of time for the problem to resolve on it’s own (which happens frequently). An additional perk is that any question or doubt of the patient can be chalked up to “being defensive /​ in denial” or “uncooperative” which is of course a symptom of the underlying psychopathology.

• EGIOct 20, 2022, 2:36 PM

  14 points

  7

  in reply to: Gordon Seidoh Worley’s comment on: My re­sent­ful story of be­com­ing a med­i­cal miracle

  The Mg hypothesis is easy to test: Get some MgCl or MgCO3 and take some with gluten containing food. Use a quantity similar to the Mg dose found in the amount of coconut water you found to help you. If it helps it is the a tive ingredient. If not, it may still be invloved but not sufficient.

• EGISep 3, 2022, 9:46 PM

  1 point

  0

  on: Laz­i­ness in AI

  This is very close to an Idea I had a couple of weeks ago: Giving the AI a strong time preference to prevent long term plotting and to make sure that a treacherous turn happens (too) early.

  So maybe this kind of approach should be investigated more. (Or is it already and I am unaware?)

• EGIApr 7, 2022, 10:00 PM

  1 point

  0

  on: The Fu­ture of Biolog­i­cal Warfare

  Infectious bioweapons can be importantly different from chemical weapons because a single soldier screwing up his protection may infect his whole unit or even more if r is high and incubation period is long.

• EGIMar 2, 2022, 8:34 PM

  2 points

  0

  on: Long covid: prob­a­bly worth avoid­ing—some considerations

  Good post, but there is one important angle you missed. It is not only the question if you get Covid, but also how often you get Covid over your lifetime which may lead to cumulative damage.

• EGIMar 2, 2022, 8:31 PM

  1 point

  0

  in reply to: Siebe’s comment on: Long covid: prob­a­bly worth avoid­ing—some considerations

  Same problem as with Lyme Disease. Weak or no antibody reaction is only good news IF it indicates absence of the pathogene. While this is not unreasonable to assume, it still needs to be demonstrated, preferably over a wide variety of differen tissues.

• EGIDec 31, 2021, 2:58 AM

  −1 points

  0

  on: Omicron: My Cur­rent Model

  You are forgetting declining immunity. Next winter may very well have a similar wave again. In fact this is imho the central scenario (60%ish probability). Quite possibly with multiple strains.

  I agree with most of your post, but this looks like wishful thinking to me.

  Also, individual prevention should be way easier than you make it out to be. Wear a well fitting N95 or better mask when meeting people and you are mostly done if you live alone or with compliant housemates. Just look at Covid stations in hospitals. Staff there is wearing mostly N95, often poorly fitting and even they do not get infected on the first day (or week) despite MUCH more exposure than we will ever get. Masks work! Masks not working in the general population is a compliance problem, not a technical problem.

• EGISep 8, 2021, 10:37 PM

  1 point

  0

  in reply to: Astynax’s comment on: On­line LessWrong Com­mu­nity Weekend

  AFAIK the schedule should adjust to your time zone. Opening session is 7pm CEST.

Re­minder: On­line—Less Wrong Com­mu­nity Week­end 2021