5 ms·
> The sources characterized that figure as a rough estimate because the company does not keep precise records on the number of animals tested and killed. How a
by neilv 4y ago
> The sources characterized that figure as a rough estimate because the company does not keep precise records on the number of animals tested and killed.
How are they doing this science without keeping records of such important information about scientific experiments and the subjects?
Or are they lying, or intentionally finding a way to do the experiments and manage subjects without keeping these records? (I could believe either. I've heard some surprising things from people involved in research on animals, including of primate facilities they know of being hidden/disguised/unmarked, where I'd never guessed. I assume the "delicate PR" of animal research is well-known by people who work in that space, and might be incentive to suppress PR-damaging records.)
- iancmceachern 4y agoTotally. For obvious animal welfare reasons it is standard practice in the industry to keep careful plans for these tests. You must prove that you absolutely need an animal model for the test, you must have a clear cut, documented procedure and design of experiments, etc. Every other company that does this kind of work does this.
- bpodgursky 4y agoThere's no legal requirement here, it's just standard IRB stuff. If you run a slaughterhouse, nobody is making you count the chickens you kill.
- iancmceachern 4y agoBut IRB is required by law. The accounts of what Neuralink did would never pass a legit IRB panel.
- telotortium 4y agoSource? Most universities have an IRB for non-human test subjects, and the federal funding requirements may require an IRB, but I can't find any federal requirements for an IRB on non-human subjects if you're not taking federal funding, even for FDA approval.
- iancmceachern 4y agoThank you, it's not an IRB in the case of an animal, it's a different acronym (though not a classic FDA TLA, three letter acronym). It's a IACUC, https://irb.utah.edu/guidelines/fda-requirements/animals.php#:~:text=The%20IRB%20reviews%20research%20that,IACUC%20and%20complete%20an%20application https://irb.utah.edu/guidelines/fda-requirements/animals.php.... So a similar process, similar requirement, just called something different.
- LatteLazy 4y agoThe scientists likely have exact records in their lab books etc. But there is no animal-accountant running a single, company wide, exactly accurate MI dashboard...
- iancmceachern 4y agoIn companies I've worked for there very much is an "animal accountant". They are also the "cadaver accountant". You can't just throw those things out with the morning trash, you can't just buy them on Amazon. You need permits, permission, documentation.
- happyopossum 4y agoPermits and permission for buying and disposing of lab rats and lab mice (which seem to make up the bulk of the 1500 number)? Hardly…
- iancmceachern 4y agoThis is reductio ad absurdum. And yes, you still need a permit for lab rats and mice. https://www.nyc.gov/nycbusiness/description/laboratory-permit-for-the-use-of-living-animals https://www.nyc.gov/nycbusiness/description/laboratory-permi...
- SamoyedFurFluff 4y agoHave you worked for a company or can you cite other experience beyond the person you’re replying to? You sound needlessly skeptical about a claim you actually have little knowledge to evaluate.
- caddemon 4y agoWell you also need to house the mice and do what you can to avoid unnecessary discomfort. In that process you will certainly be doing accounting to e.g. ensure you have enough cages and that they are being cleaned at a proper frequency. You will also be justifying the study protocol and as part of that an intended sample size. Sure if a mouse gets killed accidentally by a shaky undergrad it's not a major event, but you also can't just kill an order of magnitude more mice than the approved proposal without getting additional approvals. There are strict rules about this at every university I've seen and I imagine they developed because of government grant stipulations rather than out of the goodness of their hearts. It's a legit cost - vets need to be on staff and individual labs can be fined for breaking rules. I'd guess biomed companies end up with a similar structure normally due to how closely linked they are with the rest of the research machine. Having these centralized housing areas for mice from different projects, in addition to the IRB records of what was proposed, ought to give a pretty accurate number of mice on the department level. If the government asked for it to be calculated, I'd be shocked if the university didn't quickly turn around with a number. Regardless, it is bad scientific practice to not document the number of mice used. It's not even an animal welfare thing, you should be documenting the number of petri dishes you look at too. Not doing so could easily enable sketchy behaviors like p-hacking as well as increasing likelihood of honest mistakes. I'd also hope at a company like Neuralink they have electronic lab notebook functionality that would make it straightforward to pool numbers across individual researchers. Personally I didn't find the 1500 number so outlandish, as I assume Neuralink is targeting a very broad scope. However some of the other details certainly raise red flags about their operating procedures.