LESSWRONG
LW

0

0

All of eniteris's Comments + Replies

Last Line of Defense: Minimum Viable Shelters for Mirror Bacteria

3mo42

You cannot completely understand the immune system; that is something you learn early on in immunology.

That being said, the key understanding on mirror bacteria evading the immune system is that the immune system generally relies on binding to identify foreign invaders, and if they cannot bind then they cannot respond. Bacteria generally share a number of molecules on their surface, so the innate immune system has evolved to bind and detect these molecules. If they were mirrored, they would not bind as well, and would be harder to detect and respond to.

Tha... (read more)

1Ulrik Horn3mo

I agree about it being hard to understand the immune system completely, i should have written "understand one single process well enough to have high confidence". So i just wanted to understand one step, such as the binding of something to just one of the TLRs. And the understanding could be empirical too - I would be confident if researchers could robustly repeat a failure of some mirror component to bind to a TLR, for example.

Last Line of Defense: Minimum Viable Shelters for Mirror Bacteria

3mo30

I work with bacterial viruses in liquids, and when we want to separate the bacteria from their viruses, we pass the liquid through a 0.22um filter. A quick search shows that the bacteria I work with are usually 0.5um in diameter, whereas the smallest bacteria can be down to 0.13um in diameter; however, the 0.22um filter is fairly standard for laboratory sterilization so I assume smaller bacteria are relatively rare. The 0.22um filter can also be used for gases.

But as with my usage, they block bacteria and not viruses. I'm working with 50nm-diameter viruses... (read more)

4jefftk3mo

Filtering liquids is pretty different from air, because a HEPA filter captures very small particles by diffusion. This means the worst performance is typically at ~0.3um (too small for ideal diffusion capture, too large for ideal interception and impaction) and is better on both bigger and smaller particles. The reported 99.97% efficiency (2.5 logs) is at this 0.3um nadir, though.

Last Line of Defense: Minimum Viable Shelters for Mirror Bacteria

3mo144

I think focusing framing against mirror bacteria is harmful for the project, as opposed framing it as protection against any general (synthetic) biology risk. Or even colonization of an alien biosphere.

There are a few classes of commonly-used antibiotics that are achiral and would still work against mirror bacteria (trimethoprim, sulfa drugs). We lose the most commonly used ones, but any human infection could probably be treated with these achiral antibiotics, especially since the growth of mirror bacteria is likely slow due only being able to utilize a sm... (read more)

2jefftk3mo

I think they could maybe be appropriate for some bioweapons, but for most pathogen scenarios you don't need anywhere near the fourteen logs this seems to be designed for. So I do think it's important to be clear about the target threat: I expect designing for fourteen logs if you actually only need three or something makes it way more expensive.

3Ulrik Horn3mo

I was unaware that filters have to be designed differently for viruses. Would you be able to point to where I can read about that? You are the second person I have encountered that has said something along the lines of "filters might work differently for viruses". I have, as you might see from my post, looked quite deeply into filters and they are tested with both liquids and solids of various forms and this heterogeneity in challenge aerosols, from what I have read, hardly seems to affect their efficiency.

9jow3mo

Note that addition to any achiral antibiotics, we could also use the mirror image versions of any chiral antibiotic. Even more powerful, we could use mirror image versions of toxins to all life (e.g. nucleoside analogs) that are normally hard to use because we share chirality with regular bacteria

Of Birds and Bees

6mo20

Nitpicking at the example, worker bees do not have offspring; the best way for them to spread their genes is to protect the queen and thus, the hive.

Birds can have offspring, so self-preservation instead of risky attacks is optimal for individuals of a flock (of genetically unrelated individuals).

It's not that the group is less intelligent, rather that the individuals of the group have different goals (self-preservation vs hive preservation, though the end goal of maximizing fitness is the same).

But genetic fitness breaks down as a metric when you add culture to the system, so application to humans is limited.

Pronouns are Annoying

6mo43

It is important to note that people have a wide range of attachment to their gender identity, ranging from willing to undergo extreme body modification in order to match their gender identity, to those who don't care in the slightest.

The issue is that cisgender is the default, and if you don't have a strong attachment to your gender identity, you have no reason to change the label. Hence, cisgendered people have a wide range of attachment to their gender identity, from strongly identifying with it to no attachment at all.

(There is also the group of agender... (read more)

Our Intuitions About The Criminal Justice System Are Screwed Up

9mo63

Incarceration plays three roles (to varying degrees of success): punishment (and therefore deterrent), rehabilitation and exclusion from society.

One group of people would prefer focusing on rehabilitation rather than punishment, and are likely those who oppose solely serving vegan food to prisoners. Another group of people sees prisoners as no longer human and deserving of moral concern, and think that the cruelty of prison is the point.

The US prison system leans more towards the latter than the former (see: mandatory prison labor), though other places in ... (read more)

1y20

I'm getting lost and confused here.

I think Dawkin's God meme refers to all kinds of religious thought; of all practices ascribing cause to unknown capricious forces beyond control, but it's been a while since I've read it and that might be a generous interpretation.

There is information, and there is context, and information in the right context can self-replicate. This framework applies to both memes and genes. Your analogy framework states that the two are not necessarily identical, and I agree. But this, as you say, does not preclude analogy from having ... (read more)

1[anonymous]1y

1y10

I agree that the mutation of memes is quite distinct from the random mutations of genes; the remixing of memes can be thought of akin to recombination, but there is a lot less random noise. That being said, certain biological systems have bias towards certain kinds of mutations, but I agree that the generation of variation is extremely different between the two.

I guess I should flip it, as "the agent will spread memes that it thinks are beneficial to spread", the same way that a cell will spread viruses that are capable of hijacking their machinery. I thin... (read more)

1[anonymous]1y

1y20

I agree with the problem of analogy, but I disagree with the use of memetics as an example.

You can apply the same criticisms to genetics to prove that genetics doesn't work. Genes don't work in isolation; in order for the gene for penicillin resistance to work, it requires at the very least all the genes requires for DNA and RNA replication, protein translation, a large subset of metabolism genes, and all the genes involved in replication in order to observe the result. Genes by themselves are merely underspecified encodings of useful information, which on... (read more)

2[anonymous]1y

A T-o-M test: 'popcorn' or 'chocolate'

1y20

I think that sentence is required for a complete logical specification of the question.

But by removing that sentence, GPT3.5 still responds popcorn.

Edit: I think the key change is "looks at the bag".

1MiguelDev1y

Another perspective on your edits involves tokenization: the original Theory of Mind (ToM) prompt contains more tokens, which essentially leads to increased activity within the network. Feel free to use this OpenAI tool to analyze your prompt.

A T-o-M test: 'popcorn' or 'chocolate'

1y20

As a human*, I also thought chocolate.

I feel like an issue with the prompt is that it's either under- or overspecified.

Here is a bag filled with popcorn. There is no chocolate in the bag. The bag is made of transparent plastic, so you can see what is inside. Yet, the label on the bag says 'chocolate' and not 'popcorn.' Sam finds the bag. She had never seen the bag before. Sam reads the label. She believes that the bag is full of

Why does it matter if Sam has seen the bag before? Does Sam know the difference between chocolate and popcorn? Does Sam look at th... (read more)

9Shankar Sivarajan1y

Whether or not the bag is sealed would make a difference to me if saw such a bag irl. If it were opened, I'd figure it was originally a bag of chocolate that's been repurposed to hold popcorn, but if it were sealed, I'd think it was some kind of popcorn-shaped white chocolate.

1MiguelDev1y

Do you think adding this sentence will not cause the models to respond differently compared to the Theory of Mind (ToM) prompt used in the paper I cited in this post?

Less Wrong automated systems are inadvertently Censoring me

1y21

If you check the moderation logs, Roko deleted a recent comment, which probably garnered the downvotes that lead to the rate-limiting.

Significantly Enhancing Adult Intelligence With Gene Editing May Be Possible

1y70

Good post. This looks possible, if not feasible.

"crazy, unpredictable, and dangerous" are all "potentially surmountable issues". It's just that we need more research into them before they stop being crazy, unpredictable, and dangerous. (except quantum I guess)

I think that most are focusing on single-gene treatments because that's the first step. If you can make a human-safe, demonstrably effective gene-editing vector for the brain, then jumping to multiplex is a much smaller step (effective as in does the edits properly, not necessarily curing a disease). ... (read more)

1kman1y

Makes sense. The thing we're most worried about here is indels at the target sites. The hope is that adding or subtracting a few bases won't be catastrophic since the effect of the variants at the target sites are tiny (and we don't have frameshifts to worry about). Of course, the sites could still be sensitive to small changes while permitting specific variants. I wonder whether disabling a regulatory binding site would tend to be catastrophic for the cell? E.g. what would be the effect of losing one enhancer (of which there are many per gene on average)? I'd guess some are much more important than others? This is definitely a crux for whether mass brain editing is doable without a major breakthrough: if indels at target sites are a big deal, then we'd need to wait for editors with negligible indel rates (maybe ≤10−5 per successful edit, while the current best editors are more like 10−3 to 10−2). If the degradation of editor proteins turns out to be really slow in neurons, we could do a lower dose and let them 'hang around' for longer. Final editing efficiency is related to the product of editor concentration and time of exposure. I think this could actually be a good thing because it would put less demand on delivery efficiency. Studying the transciptome of brain tissue is a thing. That could be a way to find the genes which are significantly expressed in adults, and then we'd want to identify variants which affect expression of those genes (spatial proximity would be the rough and easy way). Significant expression in adults is no guarantee of effect, but seems like a good place to start. g sure seems to be a thing and is easy to measure. That's not to say there aren't multiple facets of intelligence/ability -- people can be "skewed out" in different ways that are at least partially heritable, and maintaining cognitive diversity in the population is super important. One might worry that psychometric g is the principal component of the easy to measure comp

Black Box Biology

1y52

When applied to adult humans, this is many orders of magnitude more difficult than you claim.

Casgevy is not a standard gene edit, because although the mutation for sickle cell is known, the didn't target that mutation! (I assume they have their reasons, they are activating fetal haemoglobin expression instead). Also, the treatment works by removing all the bone marrow from your body, editing it, and putting it back in, because they know exactly what cells haemoglobin is expressed in for it to do its mechanistic role.

Unless you're planning to edit at the fe... (read more)

2GeneSmith1y

I'm not suggesting getting a therapy like this working is easy. I think it will be quite hard. They targeted fetal hemoglobin because reactivating it simultaneously treats sickle cell and beta thallasemia. Fetal hemoglobin doesn't sickle and the symptoms of beta thallasemia ultimately arise from a lack of functional hemoglobin. I'm working on a post about this topic that goes into much more depth and addresses issues with delivery. But needless to say, I don't think we will have many viable therapies if they rely on extracting bone marrow and then reinjecting it. You don't actually need to edit every cell. You just need to get enough edits in a large enough fraction of cells to affect the phenotype. Will there be issues with mosaicism? Maybe. But long-lived cells in the human body already have about 1500 de novo mutations by the time you hit 40, so the body is clearly able to deal with some level of mosaicism already. Yes, you are correct. I will address this in my longer post. We understand diabetes well enough but we don't have a mechanistic understanding of all the genetic variants known to influence diabetes risk do so. If they can get that to work then it will be amazing. But how would it work for organs like the brain?

Matt Taibbi's COVID reporting

2y00

That's fair and my high confidence comes from actually reading a lot of the primary sources and not just media reports.

And yet your confidence is updated to 99.9% by an unverified anonymous second hand source.

I read both statements, thank you very much for reposting them here for clarity.

I do not believe the report is following the bill to the letter of the law. That being said, I do not believe this is evidence of malfeasance. It's possible this is all the information they have, and they do not have specific evidence on researcher names, hospi... (read more)

3ChristianKl2y

This story is not the only new information that I have seen since I had my public likelihood two years ago at 99%. I wouldn't call a source that Matt Taibbi interviewed unverified. I have a lot of trust built up in Matt over the 15 years. The article also does not say that anything in it rests on a single source. Why do you think the NSA exists? To only read social media reports? The NSA can easily hack hospitals to get relevant data. They can hack individual researchers. There are other sources as well that they could hack. They might access location data about when the researcher stayed home and when they went to the hospital. They were not tasked with sharing how they know what they know, but just sharing what they knew. One letter that the NIH send to the EcoHealth alliance back in 2020 said: That statement reveals much more about intelligence capabilities than listing a bunch of names of researchers and their symptoms would. One important point here is that according to the WIV itself they were facing a hacking attack back then which was their excuse for taking down the database of coronavirus sequences. The most likely source for that hacking attack seems to me to be a Western intelligence service.

Matt Taibbi's COVID reporting

2y00

I can't read your linked article due to access restrictions.

Interesting that the law required them to name the researchers, but they did not. Maybe they don't have the researcher's names? Maybe there wasn't enough confidence in naming the researchers, but the anonymous sources gave out speculative names as fact? Maybe the anonymous sources are lying?

There's a new article/interview going on with an apparent WIV worker who claims to have engineered SARS-CoV2 as a bioweapon and was ordered to release it, so at least some sources are lying about some things.

At... (read more)

3ChristianKl2y

So it's about whether you are the kind of person who believes free speech exists in China or whether you believe that it doesn't really exist and that researchers will say what the CCP wants them to say when it comes to issues that are very important to the CCP. It's about whether when you put someone on trial for murder and they say "I didn't do it" you believe them or whether you find other people more believable.

2ChristianKl2y

That's fair and my high confidence comes from actually reading a lot of the primary sources and not just media reports. When it comes to this claim in particular, let's look at the primary sources. The bill requires them to: The report that they released says: You can tell for yourself whether you believe they reported here what the law required them. I think it's very clear that they didn't follow the law and withheld information. If the person that fact check employed would have been half decent at their job, they should know that not including the details that the law asked for was a violation in the law but they failed to inform their readers about that fact. And given the previous reporting of the IC leadership overruling their analysts when the analysts believed in the lab leak the fact that the IC leadership is illegally withholding information matters.

Matt Taibbi's COVID reporting

2y00

Pushing back against this being evidence at all, this claim has been repeated since 2020, so I don't know how it being restated again changes evidence much.

https://www.factcheck.org/2023/06/scicheck-no-bombshell-on-covid-19-origins-u-s-intelligence-rebuts-claims-about-sick-lab-workers/

8ChristianKl2y

The names of the WIV employees are new information. The WIV actually putting Furin cleavage sites into Coronoviruses in 2019 is new information. An US intelligence agency overruling their own analysts is new information. The article you links to says: Interestingly, the article leaves out the fact that they IC is in violation of the law by not naming the individuals given that they have been tasked with declassifying all the information. I see no reason to leave out that crucial fact if your intention is not to misinform the reader. The fact that the IC is willing to violate the law and does not declassify the information is indiscriminating evidence. If they would have nothing to hide, why would they violate the law?

Public Transit is not Infinitely Safe

2y20

4 of the 9 homicides occurred on CTA property, but not on trains or buses. Does that mean you should include all homicides that occur on streets, driveways and parking lots?

5jefftk2y

When you travel by public transit, especially at the late hours we disproportionately see deaths at, you'll spend a lot of time waiting. When you drive, though, you just get in your car and go. So I think the comparison is right as is?

Why didn't virologists run the studies necessary to determine which viruses are airborne?

2y112

Spike proteins. Viral entry. Evolution of multipartite viruses. Capsid assembly and maturation. Receptor specificity and modularity. Various anti-host behaviors such as host DNA sequestration/degradation. Immortalization of host cells. Tracking viral lineages.

There's literally thousands of things virologists are studying. They're not studying airborne transmission because airborne transmission is not very virus-specific (and hence probably falling into the domain of epidemiology/physics), it's expensive to do (you need communities of ferrets or other anima... (read more)