Comment by dbingham | Hacker News Reader

Comment by dbingham | Hacker News Reader

Honestly, I'm kind of surprised that any software engineer could be okay with us releasing anything genetically engineered into the wild.

We can't manage systems that we designed to be highly decoupled and very manageable with out introducing bugs or triggering unintended consequences (sometime pretty catastrophic to the system in question). These are systems we ought to know everything about, because we built them, and that we've intentionally designed in such a way that a fault in one part of it shouldn't affect other parts of it. And yet, we often find they do in ways we never would have predicted. I'm talking here, of course, about software.

Genetic code is a system we don't fully understand, that's vastly more complex than any of the software systems we've built, and that is highly coupled in ways we're only just beginning to understand. Tweak a gene that we're pretty sure does one thing, and it can have effects in completely unrelated systems we never would have predicted.

And yet, we're going to mess around in that system and then release it out into wild ecosystems (another highly complex and highly coupled system that we don't fully understand). Ecosystems that we don't completely control. We have no rollback ability for these releases. No way to undo it.

And we have no way to thoroughly test these changes impacts on the systems.

Even in a culture with the mentality "move fast and break things" we don't act this cavalierly towards the systems we're working in. It's absolutely insane to behave this way. And it's only a matter of time before it bites us in the ass in a truly catastrophic way.

You do know that humans have been releasing genetically modified organisms in the wild for thousands of years, right?

Selective breeding is not the same as transgenic modification. Selective breeding more closely mimics natural evolution, and while there is still a possibility or unintended consequences, that possibility is a lot lower. It's all genetic combinations that could have manifested in the wild, getting genes that already exist in species to activate.

With transgenic changes, you're talking about adding genes to species that don't exist in those species. Genes that have no effective pathway to make that move in the wild.

You know...not every concept needs to be interpreted through the lens of software development.

In this case, it is a very apt analogy. It's literally code. It's biological software.

And I have some experience on which to base that: I worked in a university genetics lab while I was in high school writing bioperl to track down instances of gene transposition from the mitochondria and chloroplast to the nucleus. The lab was using those instances as a way to track when species split in evolutionary history.

Genetic code is code. It's different from software code in many of the ways I laid out. But when we're talking about modifying it, or engineering it, the same principles apply.

No, your analogy is neither apt nor useful. It's an oversimplification that ignores the unimaginably more complex nature of ecosystems compared to software. The same principles do not apply.

The mRNA vaccines are genetically engineered, as is a lot of our agriculture. I see your point but that’s why this test was so heavily regulated by the FDA and had to conduct many trails before it was allowed to proceed.

Yeah, I'm not arguing for abandoning genetic engineering entirely. The potential for it change the world in beneficial ways is huge.

The mRNA vaccines are a great example where the risk is worth the reward and the possibilities of unintended consequences is relatively low.

But our agriculture is a great example of an area where we should exercising way more caution than we are. There are unintended consequences we're already seeing, and further consequences that are impossible to predict.

Reply on news.ycombinator.com