4 ms·

Over the last few years I've seen many comments on Hacker News that are pro-GMO. The thing is... I don't understand how any respectable developer could support

Over the last few years I've seen many comments on Hacker News that are pro-GMO. The thing is... I don't understand how any respectable developer could support programming DNA with current technologies. You know how they code new organisms? They rip out code from one organism and then they randomly insert it somewhere in the DNA code of another organism. They do this thousands of times... 99.999% of the time the cell dies. So they keep trying to compile random variations of code and it keeps crashing. But let's say they get lucky one time and the code compiles. The cell lives, and then they use that cell to create a new organism. If that organism has the one property they wanted to port over from the new DNA code, they call it a success. They don't do any unit testing, they can't possibly test every scenario this new organism would be used in production, and yet no one cares. Would any software developer use a process like this in a production environment? No - it's obvious this way of coding would would cause drastic unintended consequences. With GMOs you get malformed proteins, new chemicals/toxins, and interbreeding with other plants that could cause even more unintended consequences. Usually I'm all for technological progress. But until we understand DNA inside and out and can precisely edit just the code we want, GMOs seem very dangerous.

ucaetano 11y ago
I'm not "pro-GMO". Not against it either. "Would any software developer use a process like this in a production environment?" Would any engineer use software development techniques in building a bridge? Absolutely not. Hold on, I'll build a mock bridge, run a test truck on top to see if it holds. If it doesn't, I'll just build a new one. Don't compare disciplines which have enormously different requirements. With all dues respect, the world isn't just software engineering.
- super-serial 11y ago
  Sorry I didn't mean to label you. I still think the analogy holds because DNA is so much like source code. If software keeps eating the world maybe everything else will be software too... well I can only hope.
  - dragonwriter 11y ago
    > I still think the analogy holds because DNA is so much like source code. DNA isn't much like source code at all. A better, though still far from precise, analogy would be base-four data which can, in some environments, configurations, and contexts, be interpreted as machine code.
dragonwriter 11y ago
Nice scare analogy for GMOs. The problem is that if you do the same thing to traditional breeding, its even worse: rather than ripping out code that you know performs the function you are looking for and inserting it into another code base, randomly or not, traditional breeding (and especially the modern, accelerated forms) is a variety of different combinations of randomly mixing code from codebases that each have some of the features you want and inducing random bit flips throughout the code base, testing for the features you want, throwing out code that doesn't have them, and hoping you don't have unnoticed unintended consequences. (And natural breeding is pretty much the same as traditional breeding, but without the selection for traits desired by humans.) That's not to say traditional breeding or natural breeding is bad, but it is to say that if we reason at the level of your scare analogy, we just would avoid food entirely, rather than avoiding GMO-food specifically; you don't really provide any reason to think that GMO food is worse than the alternatives.
- super-serial 11y ago
  I think there's a reason nature doesn't allow a frog to breed with a plant. You can call it a scare tactic... but with GMOs we're doing things that would be impossible to try with traditional methods.
  - dragonwriter 11y ago
    > You can call it a scare tactic... I can, because that's exactly what the software development analogy is: a scare tactic that relies on ignorance of how the alternatives would work if looked at the same way. There is no method available for crop development that is analogous, in the manner that things were framed upthread, to good software development. (And there won't ever be, unless, through experience with the use of the gene editing tools we have now, we develop much better tools.) > but with GMOs we're doing things that would be impossible to try with traditional methods. With "traditional" (which mostly means non-GMO -- the methods that fall under that label now are often as new, or newer, than the GMO methods) methods we are also doing things which would be impossible with the traditional methods available before the time when we started doing GMO. That's not really surprising, the advance of technology usually involves doing things impossible with earlier technology.
angersock 11y ago
We've been modifying plants through cruder means for thousands of years, killing the ones we don't like and planting the ones we do. It's a vastly slower process, relying on natural mutations, but it's the same principle. The real problem with GMOs involves patents and copyright--not necessarily anything to do with the plants themselves.
- akamaka 11y ago
  Doesn't that fact argument work against GMOs? Cross-breeding has been done for thousands of years, so it's a very well understood technique, whereas doing it in a lab is has not even had half a century of experience behind it.

ucaetano 11y ago

I'm not "pro-GMO". Not against it either. "Would any software developer use a process like this in a production environment?" Would any engineer use software development techniques in building a bridge? Absolutely not. Hold on, I'll build a mock bridge, run a test truck on top to see if it holds. If it doesn't, I'll just build a new one. Don't compare disciplines which have enormously different requirements. With all dues respect, the world isn't just software engineering.

super-serial 11y ago
Sorry I didn't mean to label you. I still think the analogy holds because DNA is so much like source code. If software keeps eating the world maybe everything else will be software too... well I can only hope.
- dragonwriter 11y ago
  > I still think the analogy holds because DNA is so much like source code. DNA isn't much like source code at all. A better, though still far from precise, analogy would be base-four data which can, in some environments, configurations, and contexts, be interpreted as machine code.

dragonwriter 11y ago

Nice scare analogy for GMOs. The problem is that if you do the same thing to traditional breeding, its even worse: rather than ripping out code that you know performs the function you are looking for and inserting it into another code base, randomly or not, traditional breeding (and especially the modern, accelerated forms) is a variety of different combinations of randomly mixing code from codebases that each have some of the features you want and inducing random bit flips throughout the code base, testing for the features you want, throwing out code that doesn't have them, and hoping you don't have unnoticed unintended consequences. (And natural breeding is pretty much the same as traditional breeding, but without the selection for traits desired by humans.) That's not to say traditional breeding or natural breeding is bad, but it is to say that if we reason at the level of your scare analogy, we just would avoid food entirely, rather than avoiding GMO-food specifically; you don't really provide any reason to think that GMO food is worse than the alternatives.

super-serial 11y ago
I think there's a reason nature doesn't allow a frog to breed with a plant. You can call it a scare tactic... but with GMOs we're doing things that would be impossible to try with traditional methods.
- dragonwriter 11y ago
  > You can call it a scare tactic... I can, because that's exactly what the software development analogy is: a scare tactic that relies on ignorance of how the alternatives would work if looked at the same way. There is no method available for crop development that is analogous, in the manner that things were framed upthread, to good software development. (And there won't ever be, unless, through experience with the use of the gene editing tools we have now, we develop much better tools.) > but with GMOs we're doing things that would be impossible to try with traditional methods. With "traditional" (which mostly means non-GMO -- the methods that fall under that label now are often as new, or newer, than the GMO methods) methods we are also doing things which would be impossible with the traditional methods available before the time when we started doing GMO. That's not really surprising, the advance of technology usually involves doing things impossible with earlier technology.

angersock 11y ago

We've been modifying plants through cruder means for thousands of years, killing the ones we don't like and planting the ones we do. It's a vastly slower process, relying on natural mutations, but it's the same principle. The real problem with GMOs involves patents and copyright--not necessarily anything to do with the plants themselves.

akamaka 11y ago
Doesn't that fact argument work against GMOs? Cross-breeding has been done for thousands of years, so it's a very well understood technique, whereas doing it in a lab is has not even had half a century of experience behind it.