4 ms·

I have friends who have worked and currently work at Neuralink, and I agree with much of the sentiment I see in the comments about Musk. I also am confused abou

I have friends who have worked and currently work at Neuralink, and I agree with much of the sentiment I see in the comments about Musk. I also am confused about the long term viability of the company. BUT, I hate to see this sort of animal rights hack-job amplified on HN. You can recognize it as such by the radical animal rights orgs quoted and the minimal contextualization or detailing. It seems like the main criticisms are (1) in 2 procedures out of 1500 , there were surgical targeting errors, and (2) many procedures are done in parallel rather than sequentially. With regard to (1), if that were really the case, 0.1% errors would be much better than human surgeons. I'm sure that the mistake rate is higher, but that still doesn't mean anyone is behaving unethically? With regard to (2), there's a huge problem in medtech around small sample sizes. It's always easy when you have N=4 to say "Oh, we should have understood this at N=1." But the reality is that a host of N=1 conclusions have been subsequently debunked. If the team wants to replicate failures to understand them, that should be applauded. There's also some discussion of wound glue - translating, it sounds like some sutures opened up, which caused an infection, and they found a vet who was happy to blame the brand of vetbond they used. There are many things wrong with Neuralink, but personally, one might that the major tragedy here is that I'm sure that the pigs that ended up with the wrong spinal implants didn't subsequently get turned into pork chops and bacon. And the reason for that life-wastage is the frustrating ability of animal rights groups to get their message amplified.

fabian2k 4y ago
The two experiments mentioned involved 61 animals, so it's not 2 out of 1500 but at least 61 out of 1500. And I think it would not be unreasonable to assume that there are more cases than the ones explicitly documented in the article, given the immense time pressue Musk seems to be creating here.
- ckemere 4y ago
  The quote I see says, "staff accidentally implanted Neuralink’s device on the wrong vertebra of two different pigs during two separate surgeries". The other issue mentioned in the Reuters article apparently has to do with testing different sizes/generations of their devices. Whether it's appropriate to sacrifice more animals to make sure the FDA is not confused by a submission is arguably not an ethical issue you can blame Neuralink for.
  - fabian2k 4y ago
    > Reuters identified four experiments involving 86 pigs and two monkeys that were marred in recent years by human errors. and > The mistakes leading to unnecessary animal deaths included one instance in 2021 when 25 out of 60 pigs in a study had devices that were the wrong size implanted in their heads My figure added up the wrong events here, but the 86 pigs and 2 monkeys is the better figure anyway. And it's not that it would confuse the FDA, if you implant the wrong device it reduces the information you get from the experiment as it is not exactly comparable to your other experiments. The article clearly states that they implanted the wrong size, so they made a mistake somewhere.
    - fastball 4y ago
      The experiment involved 86 pigs, but 86 pigs were not killed. The 2 monkeys didn't die either.
SketchySeaBeast 4y ago
I'm confused by your numbers - it's not 1,500 procedures, it's 1,500 animals, and we have quotes like " Reuters identified four experiments involving 86 pigs and two monkeys that were marred in recent years by human errors", so I have no idea where your 0.1% number comes from. Even just the 25 pigs that had the wrong implants installed brings that number up to greater than 1%.
iancmceachern 4y ago
In regards to #1 yes, but in a human surgeons case it would be called malpractice and the surgeon would likely face consequences, and have to pay out damages to the victim(s)
xracy 4y ago
I can't tell if you're so anti-animal rights activist that you're defending what you see as a reasonable use of animals, OR if you're so pro-Elon that you're willing to defend whatever. I'm reading you as the former. So let's take animals off the table. I work for a company, and we have a high-value limited resource that's well-regulated. Instead of doing methodical tests for teh product we're trying to develop with this resource, we've ruined 1500 samples of the resource. This seems unthinkable to me. Especially coming from an actual engineering with any kind of real-life resources. "Measure twice cut once" is the adage in anything with construction, because you don't want to needlessly waste any kind of resource. This is a software approach to hardware, and if they were destroying gold beyond repair, I get the weird sense you'd be less okay with it... > In all, the company has killed about 1,500 animals, including more than 280 sheep, pigs and monkeys, following experiments since 2018... 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. It's even wilder to consider they don't actually know how many animals they've killed. On any basic construction project, you should be able to account for all of the valuable resources you used. It seems unthinkable to me that you wouldn't have the same level of accounting for living things especially given the regulations around it. But Elon don't love himself some regulations. I'm guessing you feel similarly from your other comments...
- ckemere 4y ago
  My annoyance is with the articles presentation. It leads to the kind of analogies that you're making. I think the most clarifying thought I've had is that people (and the article) are imagining that Neuralink is doing a clinical trial. Instead, they're developing at least 5 different bleeding edge technologies which are the needed precursor for making the device that one day might lead to a clinical trial. Imagine 5 teams with dozens of people working on different pieces that each require multiple rounds of in vivo testing. 1000 rodents in 4 years with 200 people all working independently is not particularly wild. And to clarify, "they don't know the number" is not accurate. Rather, the USDA requires exact reporting of the number of Animal Welfare Act protected animals (non rodents) and round numbers of rodents. These are published by the USDA. Neuralink may also report precise numbers as part of their annual report to NIH or AALAC, but these are not public. My frustration is that this is the scale of a small university's (like mine) research, and it's a shame to have actual mistakes linked to context-less numbers to make the situation appear bad.

fabian2k 4y ago

The two experiments mentioned involved 61 animals, so it's not 2 out of 1500 but at least 61 out of 1500. And I think it would not be unreasonable to assume that there are more cases than the ones explicitly documented in the article, given the immense time pressue Musk seems to be creating here.

ckemere 4y ago
The quote I see says, "staff accidentally implanted Neuralink’s device on the wrong vertebra of two different pigs during two separate surgeries". The other issue mentioned in the Reuters article apparently has to do with testing different sizes/generations of their devices. Whether it's appropriate to sacrifice more animals to make sure the FDA is not confused by a submission is arguably not an ethical issue you can blame Neuralink for.
- fabian2k 4y ago
  > Reuters identified four experiments involving 86 pigs and two monkeys that were marred in recent years by human errors. and > The mistakes leading to unnecessary animal deaths included one instance in 2021 when 25 out of 60 pigs in a study had devices that were the wrong size implanted in their heads My figure added up the wrong events here, but the 86 pigs and 2 monkeys is the better figure anyway. And it's not that it would confuse the FDA, if you implant the wrong device it reduces the information you get from the experiment as it is not exactly comparable to your other experiments. The article clearly states that they implanted the wrong size, so they made a mistake somewhere.
  - fastball 4y ago
    The experiment involved 86 pigs, but 86 pigs were not killed. The 2 monkeys didn't die either.

SketchySeaBeast 4y ago

I'm confused by your numbers - it's not 1,500 procedures, it's 1,500 animals, and we have quotes like " Reuters identified four experiments involving 86 pigs and two monkeys that were marred in recent years by human errors", so I have no idea where your 0.1% number comes from. Even just the 25 pigs that had the wrong implants installed brings that number up to greater than 1%.

iancmceachern 4y ago

In regards to #1 yes, but in a human surgeons case it would be called malpractice and the surgeon would likely face consequences, and have to pay out damages to the victim(s)

xracy 4y ago

I can't tell if you're so anti-animal rights activist that you're defending what you see as a reasonable use of animals, OR if you're so pro-Elon that you're willing to defend whatever. I'm reading you as the former. So let's take animals off the table. I work for a company, and we have a high-value limited resource that's well-regulated. Instead of doing methodical tests for teh product we're trying to develop with this resource, we've ruined 1500 samples of the resource. This seems unthinkable to me. Especially coming from an actual engineering with any kind of real-life resources. "Measure twice cut once" is the adage in anything with construction, because you don't want to needlessly waste any kind of resource. This is a software approach to hardware, and if they were destroying gold beyond repair, I get the weird sense you'd be less okay with it... > In all, the company has killed about 1,500 animals, including more than 280 sheep, pigs and monkeys, following experiments since 2018... 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. It's even wilder to consider they don't actually know how many animals they've killed. On any basic construction project, you should be able to account for all of the valuable resources you used. It seems unthinkable to me that you wouldn't have the same level of accounting for living things especially given the regulations around it. But Elon don't love himself some regulations. I'm guessing you feel similarly from your other comments...

ckemere 4y ago
My annoyance is with the articles presentation. It leads to the kind of analogies that you're making. I think the most clarifying thought I've had is that people (and the article) are imagining that Neuralink is doing a clinical trial. Instead, they're developing at least 5 different bleeding edge technologies which are the needed precursor for making the device that one day might lead to a clinical trial. Imagine 5 teams with dozens of people working on different pieces that each require multiple rounds of in vivo testing. 1000 rodents in 4 years with 200 people all working independently is not particularly wild. And to clarify, "they don't know the number" is not accurate. Rather, the USDA requires exact reporting of the number of Animal Welfare Act protected animals (non rodents) and round numbers of rodents. These are published by the USDA. Neuralink may also report precise numbers as part of their annual report to NIH or AALAC, but these are not public. My frustration is that this is the scale of a small university's (like mine) research, and it's a shame to have actual mistakes linked to context-less numbers to make the situation appear bad.