I'm not a mirror life expert (though I have research experience in chemical biology), so I may be wrong here, but I am highly skeptical that mirror life would be a legitimate threat.
For starters, I think you're wrong about the natural evolution being too hard. Bacteria already use D-amino acids in cell walls and regulation (to avoid detection), but have not gone any further. If mirror life were really so easy and viable, every bacterium would be "cloaking" itself in a cloud of D-amino acids. And from there, they'd be using more and more D-amino acids and getting better and better at using them, etc etc. That's a pretty clean path.
Furthermore, it is often said that, at some point, both hands of life lived alongside each other, but our hand happened to win arbitrarily (lacking any natural advantage). If this were true and mirror life was, in fact, so dangerous, one would expect to see a "see-saw" of handed dominance (which has not happened). There are theories that our handedness has some inherent metabolic advantages and thus didn't just arise arbitrarily (Wei et al 2023, Ozturk & Sasselov, 2022). I tend to gravitate towards those. I suppose this is a bit of a post hoc assumption though.
Even then, if our dominance were arbitrary, I find it impossible to believe that a mirror bacterium could survive for a second in our "normal-handed" ecosystem. Such a bacterium wouldn’t be able to efficiently use most of the existing biosphere as food or infrastructure. You're right about achiral nutrients, but biomass has way too many ordinary-chirality proteins, sugars, nucleic acids, cell-wall fragments, lipids, etc. How would mirror life have any competitive advantage if they were living in a world of nutrients they couldn't use? A mirror bacterium’s enzymes would have incredible difficulty "eating" normal AAs, importing and metabolizing normal sugars, exploiting the normal metabolites leaked by other microbes, etc. I'm also not sure what you mean by "some bacteria don’t need food to grow." I don't think that statement is accurate at all.
Food is 100% needed, and a lot of it. If you wanted any chance at mirror life takeover, you'd have to start with an entire mirror-engineered self-sustaining ecosystem, which is next to impossible (very hard). Even if you DID do that, both ecosystems would much prefer to just be with each other (kind of like competitive exclusion), since they'd otherwise be surrounded by junk they couldn't use.
Recent research has pushed back against the feasibility of mirror life also. 14 days ago, the preprint Pla-Mauri et al. 2026 concluded that "the widespread establishment of mirror organisms in the extant biosphere is highly unlikely, highlighting the importance of ecological constraints in evaluating the risks and feasibility of synthetic life" (this has not been peer reviewed).
All that to say, I think the concern over mirror life is a bit of a hype wave. This article is, in my opinion, a little sensationalist and doesn't get into the weeds enough. Upon getting in the weeds, you notice the infeasibility. I don't think there's much real risk, and I don't think we need mirror life safety funding.
Nice comments. There's a lot of weeds to get into, but basically I think that if a bunch of leading experts with Nobel prizes are scared, it's at least a reasonable threat. Hard to know the exact odds.
Can't really speak to the first point. But my sense of the biology is that adding mirrored bits of bacteria doesn't do much good if the others bits aren't mirrored.
I don't think it matters much if both hands lived together alongside each other. Mirrorred versions of mirror bacteria would have evolved defenses against ordinary life. Normal biological life would be defenseless. The fact that one outcompeted the other in the distant past tells us little about adding an influx of mirror bacteria to an unprepared world that hasn't seen mirrored life for billions of years. And as you say, this hypothesis is speculative, so it doesn't affect the resultant probabilities much.
The reason that mirror bacteria would have an advantage is that they could bypass most biological defenses. Sure, they'd have trouble eating stuff, but it's not clear that enough of an impediment. And what do you make of the point about e-coli surviving on achiral foods.
But basically as I say the reason that I am worried isn't that I feel like I can go back and forth on every technical point forever. It's that a bunch of really smart people who know way more than you or I are scared.
Hi, coauthor of the referenced report on mirror life here! Thanks for engaging with this topic. A few thoughts on your arguments:
On the D-amino acid point: I don’t think bacteria using a few D-amino acids in cell walls or signaling creates a smooth evolutionary path toward full mirror biology. The D-amino acids bacteria use in cell walls and signaling are isolated substitutions handled by dedicated machinery like racemases that exists precisely to keep D-amino acids out of the ribosome. There's no gradient from "uses some D-amino acids peripherally" to "makes mirror proteins" without involving the core machinery, which would break with even a few substitutions. Chirality is less like a gradually optimizable trait and more like a hard systems constraint. This is why mirror life would have to be created in the lab.
On nutrients, I agree mirror organisms would face serious metabolic limitations. But the concern doesn’t require them to become dominant immediately or grow amazingly well, only that growth exceeds mortality somewhat. Plenty of central metabolites are achiral (like acetate, pyruvate, succinate, glycerol), and ordinary bacteria can grow on those alone. So it’s not obvious to me that a mirror organism would necessarily starve in every environment. Coupled with significantly reduced mortality, even slow growth could be sufficient for a mirror bacterium to become invasive. (See ch 8 of the technical report for references and a more detailed argument.)
On the Pla-Mauri preprint: it is great to see research in this space, but my impression is that it mainly studies constraints independently. Nutrient limitation alone may block invasion, but their own results also suggest predation removal can enable persistence. The concerning case is the combination: an organism that grows inefficiently but is also largely insulated from ecological controls. So in my opinion the authors overstate their conclusion that the establishment of mirror bacteria is unlikely.
Thanks for the comment:) As a protein scientist I’m a bit confused about the arguments about antibodies. We can raise antibodies against (almost) literally all classes of biomolecule. Why not D-proteins or inverted carbohydrates? We can inject llamas with all kinds of non-protein molecules of all chiralities, and reliably raise antibodies. In fact, way back in the 90s people reliably raised antibodies against D-peptides with no issues (https://pubmed.ncbi.nlm.nih.gov/8253750/). Is there an argument against this? I understand that proteases would be less efficient for processing these for display, but it seems impossible that no part of mirror life (esp. the carbohydrates and anything else vulnerable to oxidation, which often doesn’t care about chirality) could resist processing by the APCs.
Thanks for your question! Totally agree you can raise antibodies against mirror epitopes. BCRs don't have a chirality filter, and antibodies have been raised against D-peptides before. But those protocols usually supply an adjuvant to induce inflammation and (less often) a very immunogenic L-carrier for T-help. So they're artificially delivering the innate and/or T cell activation that a mirror organism is expected not to trigger on its own. The carbohydrate route could work, but it still requires inflammation dependent on stereospecific complement and lectin recognition upstream. They also usually trigger TI-2 responses which are low-affinity, IgM-dominated, and poor at memory (which are not good at clearing bacterial infections). Finally, there are also expected issue with mirror antigen loading on the MHC and TCR which further compounds T cell activation --> B cell responses.
Ok, fair enough! Would it be fair to say though that DAMPs would still be around to act as the ‘adjuvant’ in this scenario, just by the nature of a pathogenic infection inherently causing tissue damage? Also my understanding is that even if the mechanism’s not fully understood, D-polypeptides without conjugation can elicit an antibody response eg back in the day https://pmc.ncbi.nlm.nih.gov/articles/PMC1409157/?page=5. I understand the ‘less efficient in canonical mechanisms’ argument, but I mean, the same applies to the nutrition of these pathogens too - so it should be in tradeoff (both biologically and from a statistical risk perspective). Really appreciate your engagement here btw, and thank you for entertaining the thoughts of a non-expert!
On DAMPs: yes, tissue damage would release host (L-)DAMPs, so the innate "adjuvant" signal would probably eventually be there (I say 'eventually' as it's unclear when tissue damage would occur in the absence of typical virulence factors). Still, that signal mostly just activates the antigen-presenting cells. As DAMPs wouldn't be attached to the D-proteins, it probably doesn't supply the bit that's actually limiting, which is cognate CD4 T-cell help. For a B cell to class-switch, affinity-mature, and form memory, you need a peptide that's processed, loaded onto MHC-II, and recognized by the T-cell repertoire, and that's the step a mirror antigen can't really provide for itself. So the response would be fairly aspecfic.
The Sela paper you link is actually is a nice example of that. The D-copolymer with adjuvant did raise a primary response without a conjugate. But re-injection gave no boost, and the D-isomer was 50–1000× better than its L-twin at inducing immunological 'paralysis' (what we'd now call tolerance). So even with adjuvant on board, you get a non-boostable, tolerance-prone primary titer, not the affinity-matured, memory-backed response you'd need to clear something that's actively replicating.
On the symmetry point I'd push back a little, because I don't think the two inefficiencies cancel out. What matters is limited growth combined with reduced clearance, not either one on its own. Bad nutrition could still leave net-positive growth (there should be enough achiral nutrients around), while really bad immunity could leave almost no clearance. Slow growth + almost no induced mortality can still result in invasion.
Again, thanks for the response! Interesting stuff. I do wonder how much environmental debris mirror life would pick up that would make it visible again to the immune system (soluble antigens from the environment and other humans), but I suppose that’s all speculation and better safe than sorry! I’ll read more about the stuff you’ve written - you’ve certainly intrigued me!
yep absolutely no reason antibodies can’t be developed against mirror molecules (with sufficient effort). however, in vivo, a robust antibody response relies on upstream activation of innate receptors and t cells, both of which are likely to be deficient.
After we create mirror life, mirror Jeff Goldblum will be there to tell us, ".dluohs yeht fi kniht ot pots t'ndid yeht ,dluoc yeht ton ro rehtehw htiw deipuccoerp os erew stsitneics ruoY"
Mirror bacteria might evade immune recognition but likewise their own immune systems wouldn't recognize things that might hurt them. Regardless, even if mirror bacteria were to infect a non-mirror organism, they would not be able to survive. Bacterial infections subsist off of the nutrients of their host, but these nutrients would be of the wrong chirality. I'm a molecular biologist btw.
Plausibly yes, but they’d be fighting an uphill battle. It’s also worth mentioning that although mirrored bacteria may evade some of our innate immune sustem defenses, they’d still be visibile to the adaptive immune system. They’d probably also be very easy to create vaccines for, with their anti-chiral molecules being quite distinctive. Thanks for the paywalled article btw.
The report has a whole chapter explaining why adaptive responses would be dramatically impaired (authored by some of the world’s leading immunologists).
You'd think if they were so worried about this stuff they'd make the paper freely accessible! I've learned to take my fellow experts' opinions with a grain of salt when it comes to things like this.
In short, the adaptive immune response would not necessarily be unable to recognize mirror ligands, the problem would be that it would have trouble getting switched on. T cell activation requires breakdown of proteins into small peptides and mirror molecules have been shown many times to not get cleaved by normal-chirality enzymes.
While I can't posit on the accuracy of the field, the linked paper's primary model is such a bacteria could subsist off the achiral nutrients in bloodstreams, growing until representing a lethal concentration in the bloodstream.
It's possible, but it would probably have to be specifically engineered to do this. I don't think there is much risk of an accidental lab leak of an otherwise benign strain turning into an unstoppable pandemic.
That is, mirror life presents new avenues for bioweapon development, but that needs to be part of a larger discussion on preventing bioweapon proliferation, not on stopping mirror research per se.
Mirrored biomolecules carry much potential for medicine and pharmaceuticals, which we lose out on if we decide to hold back research.
I don’t think anyone has proposed to hold back on mirror molecules research. In contrast, self-replicating mirror organisms hold little potential benefit and potentially enormous risks and that is where talks about governance are focused.
So I have a question, why wouldn’t a mirror bacteria be equally vulnerable to us and bacteria in our environment? Is the best way to counter a threat like this anti mirror bacteria that goes after it? Why wouldn’t this end up like an anti matter situation?
(Obviously we need a molecular biologist or something to weigh in on this, but I presume you’ve read more on it than me since I can’t say I’ve ever heard of this before)
I can't see a reason why anyone would go to the effort to build this. It's too indiscriminate and dangerous a weapon to be of use to anyone but omnicidal terrorists, who are not known for their worldbeating scientific faculties. AI is too useful to resist building, if it weren't the fact that it's very hard to build would be a reason not to worry.
Several authors of the article used to actively work toward building mirror life. They hoped it would help with production of mirror molecules, teach us about how cells function, or even have the cells themselves function as therapeutic agents.
That's true, but those are very flimsy reasons to build something that is so plausibly an existential threat. I agree if some scientists is stubbornly convinced that mirror life is not dangerous and must be researched for some reason, they should be dissuaded with all available means, but that just doesn't seem very likely at all. Perhaps if AI were to somehow ease the research burden massively it would be of greater concern?
This seems to me the kind of threat where building awareness is the opposite of what you want to do, since a lack of awareness is likely to lead to it not being built, and awareness is unlikely to lead to a solution.
To add to Bentham’s point, several authors of the article used to actively work toward building mirror life. So if nothing was said, well-meaning researchers could have eventually created this.
Reading through bits of the linked 250 page article, it certainly doesn't spell certain doom or gloom either. Though I think mirror autotrophs, say released in the ocean, seems like the scariest candidate in this paper!
I certainly can see why teams are unable to pierce the hypothesis, since it rests on so many hard to know unknown factors and requires a massively scoped model. Perhaps as large and filled with questions as the various abiogensis hypotheses.
On the other hand, being it's speculating on a very large unknown, I think that's why it's both so scary but also perhaps not entirely convincing yet either. (and who doesn't love a good apocalypse scenario?) Historically, fields that could not directly observe anything being hypothesized over were usually quite wrong once new observation technologies opened the field up.
But reading the paper, uncertainties about all the various forms of fitness come off in totality that while a super-fit mirror life is certainly plausible, mirror life being less fit or unfit also aren't unlikely.
However, if our ultimate environmental catastrophe is a novel life of our own creation, I find it at least a more entertaining ending - gods who cursed themselves by their creation - than environmental pollution or nuclear weapons.
I'm not a mirror life expert (though I have research experience in chemical biology), so I may be wrong here, but I am highly skeptical that mirror life would be a legitimate threat.
For starters, I think you're wrong about the natural evolution being too hard. Bacteria already use D-amino acids in cell walls and regulation (to avoid detection), but have not gone any further. If mirror life were really so easy and viable, every bacterium would be "cloaking" itself in a cloud of D-amino acids. And from there, they'd be using more and more D-amino acids and getting better and better at using them, etc etc. That's a pretty clean path.
Furthermore, it is often said that, at some point, both hands of life lived alongside each other, but our hand happened to win arbitrarily (lacking any natural advantage). If this were true and mirror life was, in fact, so dangerous, one would expect to see a "see-saw" of handed dominance (which has not happened). There are theories that our handedness has some inherent metabolic advantages and thus didn't just arise arbitrarily (Wei et al 2023, Ozturk & Sasselov, 2022). I tend to gravitate towards those. I suppose this is a bit of a post hoc assumption though.
Even then, if our dominance were arbitrary, I find it impossible to believe that a mirror bacterium could survive for a second in our "normal-handed" ecosystem. Such a bacterium wouldn’t be able to efficiently use most of the existing biosphere as food or infrastructure. You're right about achiral nutrients, but biomass has way too many ordinary-chirality proteins, sugars, nucleic acids, cell-wall fragments, lipids, etc. How would mirror life have any competitive advantage if they were living in a world of nutrients they couldn't use? A mirror bacterium’s enzymes would have incredible difficulty "eating" normal AAs, importing and metabolizing normal sugars, exploiting the normal metabolites leaked by other microbes, etc. I'm also not sure what you mean by "some bacteria don’t need food to grow." I don't think that statement is accurate at all.
Food is 100% needed, and a lot of it. If you wanted any chance at mirror life takeover, you'd have to start with an entire mirror-engineered self-sustaining ecosystem, which is next to impossible (very hard). Even if you DID do that, both ecosystems would much prefer to just be with each other (kind of like competitive exclusion), since they'd otherwise be surrounded by junk they couldn't use.
Recent research has pushed back against the feasibility of mirror life also. 14 days ago, the preprint Pla-Mauri et al. 2026 concluded that "the widespread establishment of mirror organisms in the extant biosphere is highly unlikely, highlighting the importance of ecological constraints in evaluating the risks and feasibility of synthetic life" (this has not been peer reviewed).
All that to say, I think the concern over mirror life is a bit of a hype wave. This article is, in my opinion, a little sensationalist and doesn't get into the weeds enough. Upon getting in the weeds, you notice the infeasibility. I don't think there's much real risk, and I don't think we need mirror life safety funding.
Take everything I just said with a grain of salt.
Nice comments. There's a lot of weeds to get into, but basically I think that if a bunch of leading experts with Nobel prizes are scared, it's at least a reasonable threat. Hard to know the exact odds.
To discuss in a bit more detail.
Can't really speak to the first point. But my sense of the biology is that adding mirrored bits of bacteria doesn't do much good if the others bits aren't mirrored.
I don't think it matters much if both hands lived together alongside each other. Mirrorred versions of mirror bacteria would have evolved defenses against ordinary life. Normal biological life would be defenseless. The fact that one outcompeted the other in the distant past tells us little about adding an influx of mirror bacteria to an unprepared world that hasn't seen mirrored life for billions of years. And as you say, this hypothesis is speculative, so it doesn't affect the resultant probabilities much.
The reason that mirror bacteria would have an advantage is that they could bypass most biological defenses. Sure, they'd have trouble eating stuff, but it's not clear that enough of an impediment. And what do you make of the point about e-coli surviving on achiral foods.
But basically as I say the reason that I am worried isn't that I feel like I can go back and forth on every technical point forever. It's that a bunch of really smart people who know way more than you or I are scared.
Hi, coauthor of the referenced report on mirror life here! Thanks for engaging with this topic. A few thoughts on your arguments:
On the D-amino acid point: I don’t think bacteria using a few D-amino acids in cell walls or signaling creates a smooth evolutionary path toward full mirror biology. The D-amino acids bacteria use in cell walls and signaling are isolated substitutions handled by dedicated machinery like racemases that exists precisely to keep D-amino acids out of the ribosome. There's no gradient from "uses some D-amino acids peripherally" to "makes mirror proteins" without involving the core machinery, which would break with even a few substitutions. Chirality is less like a gradually optimizable trait and more like a hard systems constraint. This is why mirror life would have to be created in the lab.
On nutrients, I agree mirror organisms would face serious metabolic limitations. But the concern doesn’t require them to become dominant immediately or grow amazingly well, only that growth exceeds mortality somewhat. Plenty of central metabolites are achiral (like acetate, pyruvate, succinate, glycerol), and ordinary bacteria can grow on those alone. So it’s not obvious to me that a mirror organism would necessarily starve in every environment. Coupled with significantly reduced mortality, even slow growth could be sufficient for a mirror bacterium to become invasive. (See ch 8 of the technical report for references and a more detailed argument.)
On the Pla-Mauri preprint: it is great to see research in this space, but my impression is that it mainly studies constraints independently. Nutrient limitation alone may block invasion, but their own results also suggest predation removal can enable persistence. The concerning case is the combination: an organism that grows inefficiently but is also largely insulated from ecological controls. So in my opinion the authors overstate their conclusion that the establishment of mirror bacteria is unlikely.
Thanks for the comment:) As a protein scientist I’m a bit confused about the arguments about antibodies. We can raise antibodies against (almost) literally all classes of biomolecule. Why not D-proteins or inverted carbohydrates? We can inject llamas with all kinds of non-protein molecules of all chiralities, and reliably raise antibodies. In fact, way back in the 90s people reliably raised antibodies against D-peptides with no issues (https://pubmed.ncbi.nlm.nih.gov/8253750/). Is there an argument against this? I understand that proteases would be less efficient for processing these for display, but it seems impossible that no part of mirror life (esp. the carbohydrates and anything else vulnerable to oxidation, which often doesn’t care about chirality) could resist processing by the APCs.
Thanks for your question! Totally agree you can raise antibodies against mirror epitopes. BCRs don't have a chirality filter, and antibodies have been raised against D-peptides before. But those protocols usually supply an adjuvant to induce inflammation and (less often) a very immunogenic L-carrier for T-help. So they're artificially delivering the innate and/or T cell activation that a mirror organism is expected not to trigger on its own. The carbohydrate route could work, but it still requires inflammation dependent on stereospecific complement and lectin recognition upstream. They also usually trigger TI-2 responses which are low-affinity, IgM-dominated, and poor at memory (which are not good at clearing bacterial infections). Finally, there are also expected issue with mirror antigen loading on the MHC and TCR which further compounds T cell activation --> B cell responses.
Ok, fair enough! Would it be fair to say though that DAMPs would still be around to act as the ‘adjuvant’ in this scenario, just by the nature of a pathogenic infection inherently causing tissue damage? Also my understanding is that even if the mechanism’s not fully understood, D-polypeptides without conjugation can elicit an antibody response eg back in the day https://pmc.ncbi.nlm.nih.gov/articles/PMC1409157/?page=5. I understand the ‘less efficient in canonical mechanisms’ argument, but I mean, the same applies to the nutrition of these pathogens too - so it should be in tradeoff (both biologically and from a statistical risk perspective). Really appreciate your engagement here btw, and thank you for entertaining the thoughts of a non-expert!
Thanks for these questions!
On DAMPs: yes, tissue damage would release host (L-)DAMPs, so the innate "adjuvant" signal would probably eventually be there (I say 'eventually' as it's unclear when tissue damage would occur in the absence of typical virulence factors). Still, that signal mostly just activates the antigen-presenting cells. As DAMPs wouldn't be attached to the D-proteins, it probably doesn't supply the bit that's actually limiting, which is cognate CD4 T-cell help. For a B cell to class-switch, affinity-mature, and form memory, you need a peptide that's processed, loaded onto MHC-II, and recognized by the T-cell repertoire, and that's the step a mirror antigen can't really provide for itself. So the response would be fairly aspecfic.
The Sela paper you link is actually is a nice example of that. The D-copolymer with adjuvant did raise a primary response without a conjugate. But re-injection gave no boost, and the D-isomer was 50–1000× better than its L-twin at inducing immunological 'paralysis' (what we'd now call tolerance). So even with adjuvant on board, you get a non-boostable, tolerance-prone primary titer, not the affinity-matured, memory-backed response you'd need to clear something that's actively replicating.
On the symmetry point I'd push back a little, because I don't think the two inefficiencies cancel out. What matters is limited growth combined with reduced clearance, not either one on its own. Bad nutrition could still leave net-positive growth (there should be enough achiral nutrients around), while really bad immunity could leave almost no clearance. Slow growth + almost no induced mortality can still result in invasion.
Again, thanks for the response! Interesting stuff. I do wonder how much environmental debris mirror life would pick up that would make it visible again to the immune system (soluble antigens from the environment and other humans), but I suppose that’s all speculation and better safe than sorry! I’ll read more about the stuff you’ve written - you’ve certainly intrigued me!
yep absolutely no reason antibodies can’t be developed against mirror molecules (with sufficient effort). however, in vivo, a robust antibody response relies on upstream activation of innate receptors and t cells, both of which are likely to be deficient.
After we create mirror life, mirror Jeff Goldblum will be there to tell us, ".dluohs yeht fi kniht ot pots t'ndid yeht ,dluoc yeht ton ro rehtehw htiw deipuccoerp os erew stsitneics ruoY"
David Lynch tried to warn us as well... https://www.youtube.com/watch?v=BL57-9171pk
Mirror bacteria might evade immune recognition but likewise their own immune systems wouldn't recognize things that might hurt them. Regardless, even if mirror bacteria were to infect a non-mirror organism, they would not be able to survive. Bacterial infections subsist off of the nutrients of their host, but these nutrients would be of the wrong chirality. I'm a molecular biologist btw.
Your second claim is innacurate! there are many achiral nutrients such as glycerol, glycine, etc. that bacteria could plausibly feed on.
https://www.science.org/doi/10.1126/science.ads9158
Plausibly yes, but they’d be fighting an uphill battle. It’s also worth mentioning that although mirrored bacteria may evade some of our innate immune sustem defenses, they’d still be visibile to the adaptive immune system. They’d probably also be very easy to create vaccines for, with their anti-chiral molecules being quite distinctive. Thanks for the paywalled article btw.
The report has a whole chapter explaining why adaptive responses would be dramatically impaired (authored by some of the world’s leading immunologists).
You'd think if they were so worried about this stuff they'd make the paper freely accessible! I've learned to take my fellow experts' opinions with a grain of salt when it comes to things like this.
Here is a non-paywalled link to the article, which is hosted on the Mirror Biology Dialogues Fund website for accessibility: https://www.science.org/stoken/author-tokens/ST-2327/full .
In short, the adaptive immune response would not necessarily be unable to recognize mirror ligands, the problem would be that it would have trouble getting switched on. T cell activation requires breakdown of proteins into small peptides and mirror molecules have been shown many times to not get cleaved by normal-chirality enzymes.
the report is freely available
the paper in science is paywalled due to publisher policies but can be read in full in the supplementary materials
While I can't posit on the accuracy of the field, the linked paper's primary model is such a bacteria could subsist off the achiral nutrients in bloodstreams, growing until representing a lethal concentration in the bloodstream.
It's possible, but it would probably have to be specifically engineered to do this. I don't think there is much risk of an accidental lab leak of an otherwise benign strain turning into an unstoppable pandemic.
That is, mirror life presents new avenues for bioweapon development, but that needs to be part of a larger discussion on preventing bioweapon proliferation, not on stopping mirror research per se.
Mirrored biomolecules carry much potential for medicine and pharmaceuticals, which we lose out on if we decide to hold back research.
I don’t think anyone has proposed to hold back on mirror molecules research. In contrast, self-replicating mirror organisms hold little potential benefit and potentially enormous risks and that is where talks about governance are focused.
"No one's trying to stop research on nuclear power, we just want to stop people from building nuclear power plants. They're dangerous don't you know!"
LOL some kid ran a debate case on me involving mirror life -> extinction once!
So I have a question, why wouldn’t a mirror bacteria be equally vulnerable to us and bacteria in our environment? Is the best way to counter a threat like this anti mirror bacteria that goes after it? Why wouldn’t this end up like an anti matter situation?
(Obviously we need a molecular biologist or something to weigh in on this, but I presume you’ve read more on it than me since I can’t say I’ve ever heard of this before)
I can't see a reason why anyone would go to the effort to build this. It's too indiscriminate and dangerous a weapon to be of use to anyone but omnicidal terrorists, who are not known for their worldbeating scientific faculties. AI is too useful to resist building, if it weren't the fact that it's very hard to build would be a reason not to worry.
Several authors of the article used to actively work toward building mirror life. They hoped it would help with production of mirror molecules, teach us about how cells function, or even have the cells themselves function as therapeutic agents.
It is astonishing how such spins on Pascal's Gambit are taken much more seriously than the original one. :)
That's true, but those are very flimsy reasons to build something that is so plausibly an existential threat. I agree if some scientists is stubbornly convinced that mirror life is not dangerous and must be researched for some reason, they should be dissuaded with all available means, but that just doesn't seem very likely at all. Perhaps if AI were to somehow ease the research burden massively it would be of greater concern?
IIRC the AI 2027 scenario involved the AIs developing mirror life to kill us all.
Things I worry about more than mirrored life (but I read Frank Herbert's "White Plague" at a formative age) - **This** https://www.bbc.co.uk/worldservice/sci_tech/highlights/010117_mousepox.shtml. **intersecting with this** https://www.science.org/content/article/how-canadian-researchers-reconstituted-extinct-poxvirus-100000-using-mail-order-dna **or this** https://www.ncbi.nlm.nih.gov/books/NBK84435/ **given this** https://royalsocietypublishing.org/rstb/article/375/1812/20190572/31288/Variola-virus-genome-sequenced-from-an-eighteenth. TL;DR someone with published smallpox genomes, a gene synthesis machine, and lab skills could theoretically recreate a virus. Mousepox experiments showed engineered pathogens can become far deadlier than natural strains, even defeating vaccines (this was done by accident). The technical barriers to "resurrecting" eradicated diseases and making them far deadlier are now lower than the biological barriers to stopping them.
This seems to me the kind of threat where building awareness is the opposite of what you want to do, since a lack of awareness is likely to lead to it not being built, and awareness is unlikely to lead to a solution.
It's important to taboo research in the scientific community that would let people build it.
To add to Bentham’s point, several authors of the article used to actively work toward building mirror life. So if nothing was said, well-meaning researchers could have eventually created this.
Reading through bits of the linked 250 page article, it certainly doesn't spell certain doom or gloom either. Though I think mirror autotrophs, say released in the ocean, seems like the scariest candidate in this paper!
I certainly can see why teams are unable to pierce the hypothesis, since it rests on so many hard to know unknown factors and requires a massively scoped model. Perhaps as large and filled with questions as the various abiogensis hypotheses.
On the other hand, being it's speculating on a very large unknown, I think that's why it's both so scary but also perhaps not entirely convincing yet either. (and who doesn't love a good apocalypse scenario?) Historically, fields that could not directly observe anything being hypothesized over were usually quite wrong once new observation technologies opened the field up.
But reading the paper, uncertainties about all the various forms of fitness come off in totality that while a super-fit mirror life is certainly plausible, mirror life being less fit or unfit also aren't unlikely.
However, if our ultimate environmental catastrophe is a novel life of our own creation, I find it at least a more entertaining ending - gods who cursed themselves by their creation - than environmental pollution or nuclear weapons.
Certainly not a guarantee, just a big threat.