Mail-Order CRISPR Kits Allow Anyone to Hack DNA • Hacker News Board

Re: Mail-Order CRISPR Kits Allow Anyone to Hack DNA

#41

Post by fsiefken » Sat, Nov 11, 2017, 1:38 PM UTC

1001101 wrote: post #21

ilamont wrote: post #4

Finally, what about the nightmare scenario: Is CRISPR so easy to use that we need to worry about biohackers—either accidentally or intentionally—creating dangerous pathogens? Carroll and others think that the danger of putting CRISPR in the hands of the average person is relatively low. “People have imagined scenarios where scientists could use CRISPR to generate a virulent pathogen, ” he says. “How big is the risk?…

> It's hard for "normal" people to imagine how sociopaths or those with destructive aims will behave when given flexible tools like this. Given enough time, you just need one misanthropic individual. Edit: and it doesn't just need to be CRISPR. It could be AI, fusion, you name it. The power of the individual to impose their will on larger and larger numbers of people and parts of the planet keeps increasing with tech…

Remember the guy in Snow Crash who drove around with his brains wired to a fusion bomb.

Re: Mail-Order CRISPR Kits Allow Anyone to Hack DNA

#42

Post by dTal » Sun, Nov 12, 2017, 2:21 PM UTC

maratd wrote: post #34

ben_w wrote: post #14

Earlier quoted context omitted.

Back in the days of Mac OS classic, there was a tool called ResEdit. You could use it to explore and modify apps. Or the core OS. While it was running. It warned you this was a bad idea, but I did it anyway. It took ten years before I saw a computer lock up that hard again. That’s how I see humans DIY hacking their own genome. If you’re lucky you’ll get what you were after. If not, you now have sexuality transmissibl…

> If you’re lucky you’ll get what you were after. No, if you're lucky, nothing happens. The majority of the other alternatives are just death. Just imagine a complex codebase and then you start changing parts of it randomly. What are you going to get? Either nothing or more likely, some sort of a crash which in biology means death. What's the probability that you'll get a feature or fix a bug doing that? Almost 0.

What about a complex codebase with nearly 8 billion subtly different running copies? Your copy has a bug, so you find a copy that doesn't and monkey-patch the assembly. Have you fixed the bug? Unless it was in some initialization code that won't get run again, your odds are pretty good.

A lot of genetic disorders are simply "whoops, your gene for X protein/enzyme/whatever is corrupted". I'd sure be tempted to patch that in myself.

Re: Mail-Order CRISPR Kits Allow Anyone to Hack DNA

#43

Post by ben_w » Sun, Nov 12, 2017, 5:16 PM UTC

maratd wrote: post #34

ben_w wrote: post #14

Earlier quoted context omitted.

Back in the days of Mac OS classic, there was a tool called ResEdit. You could use it to explore and modify apps. Or the core OS. While it was running. It warned you this was a bad idea, but I did it anyway. It took ten years before I saw a computer lock up that hard again. That’s how I see humans DIY hacking their own genome. If you’re lucky you’ll get what you were after. If not, you now have sexuality transmissibl…

> If you’re lucky you’ll get what you were after. No, if you're lucky, nothing happens. The majority of the other alternatives are just death. Just imagine a complex codebase and then you start changing parts of it randomly. What are you going to get? Either nothing or more likely, some sort of a crash which in biology means death. What's the probability that you'll get a feature or fix a bug doing that? Almost 0.

“Lucky“ in the sense that the tutorial you followed didn’t miss a step, that you followed it perfectly, and that you correctly understood what it did. How often do people copy the wrong thing from stackoverflow, for example? Or miss an important step, or the last character of text?

I rather doubt many people will try to hack their own DNA by keyboard mashing a new sequence.

Re: Mail-Order CRISPR Kits Allow Anyone to Hack DNA

#44

Post by ben_w » Sun, Nov 12, 2017, 5:20 PM UTC

bayesian_horse wrote: post #10

I am not afraid of what "biohackers" can do with CRISPR. It's very hard to come up with "pandemic viruses", even for nature, let alone for scientists, or even hobbyists. At worst, the capabilities CRISPR and other techniques from synthetic biology give to genuine researchers will offset any mischief a hobbyist can do.

Nature doesn’t ”want” lethal pandemic diseases. It’s evolutionary suicide to wipe out your own host.

Re: Mail-Order CRISPR Kits Allow Anyone to Hack DNA

#45

Post by ben_w » Sun, Nov 12, 2017, 5:26 PM UTC

bayesian_horse wrote: post #39

TheOtherHobbes wrote: post #18

Earlier quoted context omitted.

The hard part is protecting your side against the effects. If you're not worried about protecting a population, pathogen production and delivery becomes rather simpler.

Protecting anyone hasn't historically been a concern for terrorists at all. Islamic terrorists, in particular, hit entirely Muslim targets all the time. And no, even then, production and delivery isn't easy. There are no pathogens which are both highly infectious and highly pathogenic, both because those traits are mutually inhibiting, and such a pathogen would already have run its course in the world. Any attempt to…

“Islamic” is probably the wrong level to group them. I would be surprised if many people in Daesh are cool with their own terrorists killing lots of Daesh supporters no matter how many Shia Muslims they kill without losing sleep.

Re: Mail-Order CRISPR Kits Allow Anyone to Hack DNA

#46

Post by bayesian_horse » Mon, Nov 13, 2017, 4:56 PM UTC

ben_w wrote: post #45

bayesian_horse wrote: post #39

Earlier quoted context omitted.

Protecting anyone hasn't historically been a concern for terrorists at all. Islamic terrorists, in particular, hit entirely Muslim targets all the time. And no, even then, production and delivery isn't easy. There are no pathogens which are both highly infectious and highly pathogenic, both because those traits are mutually inhibiting, and such a pathogen would already have run its course in the world. Any attempt to…

“Islamic” is probably the wrong level to group them. I would be surprised if many people in Daesh are cool with their own terrorists killing lots of Daesh supporters no matter how many Shia Muslims they kill without losing sleep.

I did mean the group of terrorists who refer to some version of Islam (as by their interpretation), and regardless which subtype in this group, they all seem to be killing a lot of muslims, again without particular regard of which kind of muslim.

I admit this is an overgeneralisation, and there is a lot of subtle details with significance, but my point was simply that terrorists often don't care about their "partisan" civilians.

Re: Mail-Order CRISPR Kits Allow Anyone to Hack DNA

#47

Post by bayesian_horse » Mon, Nov 13, 2017, 5:04 PM UTC

has2k1 wrote: post #23

bayesian_horse wrote: post #10

I am not afraid of what "biohackers" can do with CRISPR. It's very hard to come up with "pandemic viruses", even for nature, let alone for scientists, or even hobbyists. At worst, the capabilities CRISPR and other techniques from synthetic biology give to genuine researchers will offset any mischief a hobbyist can do.

The democratisation/commoditization of a technology increases the probability that any given one capability of that technology will be discovered [1]. If what you think of as "very hard" is rooted in historical and present capabilities, those with angst on the issue are uncertain how long it will remain "very hard". Given that there is radical change, I do not think that sense of dread is misplaced. [1] This is no a…

"Hard" in evolutionary terms means that there are dozens or hundreds of steps, which are individually very unlikely, compounding to the improbability of them happening.

Evolution solves these things by brute force, but only if there is an evolutionary pressure. "Pandemicity" is at best an accident, from an evolutionary standpoint, and so is lethality. Both happening at once, thus, is very hard.

An engineer could arrive at a solution faster, given that he knows the steps. But knowing which steps to take is actually the hardest part of synthetic biology.