Hacker News new | past | comments | ask | show | jobs | submit login
A step toward systems that can assess competence using neurological data (wsj.com)
104 points by em1305 on Oct 3, 2018 | hide | past | favorite | 66 comments



I think it’s really detecting who has more connections in a given region.

That being said, I was born with a condition which leaves the sections of my brain loosely coupled. Usually this leads to dyslexia, schizophrenia, retardation, etc.

Luckily, the mutation that runs in my family is not the kind that lowers your IQ (it increases it, with a ~30% schizophrenia chance).

Because of that, I assume this scan would detect me having low skills. When I’m reality I’ve always excelled (as has my whole family). My only symptom is I don’t read linearly (kinda bounce all over the page).

I hope this doesn’t become a thing, else it’s kind of like judging job fitness based on genetic markers (I.e. race)


I think people tend to set up a false dichotomy between absolutely no genetic knowledge whatsoever and full on Gattaca. Basketball is a good example here. The desired genetic expression is quite visible in that game, and indeed most basketball players are quite tall. Yet on the other hand there have also been professional 5'3" players. Genetic information can tell you that, all other things being equal, one person is probably going to be more capable at some task than another person. But in reality it's never the case that all other things are equal, and so genetic expression is just yet another component that can ideally simply help create a more accurate profile of who you might be hiring.

On the other hand there is one very paradoxical risk. Consider fields where people work to marginalize the magnitude or relevance of inherent differences between people. Imagine you're hiring for a basketball team and think people are mostly about equal and any differences can be made up with training and effort. What are you going to do? Simply hire all the tallest people! And the same would be true in any field where individual exceptionalism is not emphasized. It's quite a peculiar notion that focusing on 'blank slate' equality would be what would lead to Gattaca, but I can find no flaws in the logic.


> I hope this doesn’t become a thing

It's useful in that it's another way to find companies you don't want to work for. The whiteboard interview being another signal.


How'd you learn that you have this condition?


That is extremely fascinating. Do you have any insight and/or speculations on why looser networked neurons end up benefiting you rather than harming in your particular case?


No, it detected activity in the motor cortex for surgions performing fine motor skills tasks


Could you tell me something more specific? I have a high IQ but my grandma had schizophrenia. My cousin has some other mental problems and dyslexia. I wonder if my offspring would have elevated chance of mental illness.


I took a genetics test- it was very clear my brother and mother have it. Me to a lesser degree


Did they name this mutation/condition? Something I could google to find out what kind of genetics test should I do.


Interesting. What specific mutation/SNP[s] is it?


Very curious about this as well!


I'm traveling and can't pull up my genetics test, it's a marker that impacts the corpus collusum and makes it very weakly connected. This leads to the increased risk of several diseases, and a likely increased IQ. It's passed down in the X chromosome. For us it was identified in brain scans, then I got everyone to take genetics tests which confirmed.

From there, I'm guessing you can maybe find something on Google. It's an active area that's still being explored too. From ancedotal evidence we do have schizophrenia run in the family, and we also regularly break 130 on IQ tests - so take it for what you will


Thank you! It may be of interest to lots of people here in HN (and elsewhere) with similar phenotypic profiles; so if you ever have a chance, please post more details if you don't mind! It could benefit a lot of people.


which condition are you referring to?


This is not a substitute for diversity jobs.


The Science Advances article is http://advances.sciencemag.org/content/4/10/eaat3807

"Seventeen surgeons and 13 medical students participated in this study." According to Table S1 they were classified as:

- 8 expert, mean age 35, 700 laparoscopic procedures in average

- 9 novice surgeons, mean age 31, 60 laparoscopic procedures in average

- 7 trainees and 5 controls, medical school year 1-4, 0 laparoscopic procedures.

The article has 8 authors.



Awesome... another loss of privacy at the hands of convenience and data for ignorant decision-makers



Either some scientist must have broken their wrist facepalming over this heading, or else it must be a scam. How have we suddenly advanced X00 years in our understanding of the brain, to the point where our scans can reliably tell the difference between a brain expending slightly more energy in a vaguely defined volume and a brain fine tuned to deal with a specific problem? Of course, business owners would pay a great deal of money to make their hiring practices even 1% more "accurate," no matter how many candidates they end up rejecting because the computer says "No."


surprised the article didn't bring this up, but anyways, let's defuse this entire idea with a very simple thought experiment so that we'll never have to think about it again.

the tech they have measures brain activity in different regions. great. the article mentions in experienced personnel there is less activity which would indicate conscious planning and more activity which would indicate doing. or maybe it's backwards. it doesn't really matter. what it's really saying is that people who are more experienced are more efficient at performing a task on a neurological basis, which is measurable.

but the efficiency of a brain circuit at performing a task is irrespective of whether that task turns out correctly or incorrectly. you can learn how to do things incorrectly. your brain's efficiency for performing a task in a certain way increases the more you do things that way, regardless of whether the outcome is what is desired.

so, here's the thought experiment: we have one master surgeon who is known for getting great patient outcomes via consistently meticulous, premeditated, and intentional application of tradecraft, and another surgeon who is awful, having killed half of their patients. the awful surgeon is very experienced -- just as experienced as the master surgeon, in fact.

the master's perfectionism leads the master to see small flaws in each of their executions, driving them to do better next time via careful applications of past lessons. in contrast, the awful surgeon thinks they're pretty good, but that the patients they've had who died were more or less beyond helping or perhaps that another member of the surgical team did something wrong. the awful surgeon never really learns from mistakes, and spends most operations thinking about dinner rather than thinking about how to help the patient.

who does the brain scan say is more skilled?

on the basis of the science described in the article, it's the awful surgeon every time. the awful surgeon doesn't put in as much conscious effort, leaving them to rely on the incorrect patterns that they've learned and never engaged with, which is what the scan will see. in contrast, the master surgeon's attentiveness will be perceived as inefficiency and lack of skill, even if the corresponding motor regions are as efficient as the awful surgeon's.

to put it differently, this isn't a "we can do more technology and get around this" so much as a fundamental drawback of the tech itself. you'd need to introduce other data to even start to make a case for the scans as corresponding to skill, and even then there are plenty of other issues.


There are all sorts of ways that correlation between brainscan images and skill might not be causation. This might just be how familiar a surgeon is with the setting, so an experienced but incompetent surgeon shows as "skilled" and an inexperienced but competent surgeon shows as "unskilled".

I don't know if this using machine learning but one might see the whole machine learning field as resulting in a rush to correlation-based "cargo-cults" of various sorts (see "racist AI" and etc).

(wsj paywall articles are annoying)


It only detected a activity on a motor skills task in the motor cortex.

This has not been generalized to cognitive skills beyond, "the subject is inactive"



If you are novice of course you have to think more and things don’t come automatic. That’s basically how they are measuring this.


Numchuku skills, bow hunting skills, computer hacking skills...


I would choose a brain scan over whiteboard interviews any day.


In the end you will be getting that brain scan while you whiteboard.


Can’t read the full article due to paywall. Two important points:

For anything team based, skill, beyond basic competence and trainability, doesn’t matter very much. A team full of mediocre team players is going to perform better than a team of selfish rockstars. Implication: focus on soft skills, organization, facilitation rather than picking the best people with brain scans.

Job aids are a really significant factor - see the checklist manifesto. It doesn’t matter how good the surgeon is, they’re still human. Things like checklists, automated code analysis/testing/fuzzing, best practice review processes, pair programming, and agile can make or break projects regardless of the skills of those involved.

Another implication: as a manager, you have a lot of control over how your team performs. When you say that you need to hire Johnny because he has the best brain scan, you’re totally ignoring the fact that Sally can totally learn to do the work.


Click on “web” by the link to the article


I forget the policy on HN re: talking about paywalls, but incognito mode got me to the article without problems.


If I'm a white male, but I have the brain scan of a Latino grandmother, does that count as diversity in the workplace?


It might hurt your chances because of the fear of Las Chanclas!


Your comment strongly suggests that won't be the case, so I wouldn't worry about it.


If this means I don't have to sit through annoying whiteboard interviews, or talk to HR bots, and all i have to do is just send them a link to my brainscan on linked in. Nice!


Hey so we checked your scan and noticed you only have experience with .NET 4.5, we’re really looking for someone with .NET 4.7.


HR will still ask for 8 years of .Net 4.9


"Actually our engineering team is looking more for CLR experience right now."


I hear that plastic brain surgery clinics are booming.


I'll take the "volkswagon special" for one please



As dystopian as "brain scan hiring" sounds, I'd gladly accept it as normal if it meant the chance to get rid of the recruiters that most companies these days use.


We believe that you have the potential to fill this position but we are sorry to tell you that you may have brain cancer. We have already filled the position with someone else whose less potentially tumorous brain was a better fit.

If you survive, we encourage you to pay attention for future openings. Thanks!


> If you survive, we encourage you to pay attention for future openings.

We wish you luck, but our HMO has given us a healthy workplace credit for pre-identifying your condition and we are not allowed to consider you for future positions.


I dont think it sounds dystopian. There is a taboo around intelligence as if it is the most precious thing on earth. History (and AI) prove that intelligence matters in some extremely high levels, but for most knowledge jobs intelligence is far from everything.


Not dystopian at all, if you've got the right sort of scans. Otherwise welcome to the unwashed masses.


Who mentioned intelligence? This is about skills, a far more valuable and meaningful thing to measure.


And I wouldn't mind if they were fair. I strongly believe Amazon, Google, Microsoft & co. interviews are 65% luck, 35% skill.


I think "personality" deserves its own category + percentage.


Once a metric becomes a target, it ceases to be a good metric. I think that people would find ways of targeting their learning towards these brain scans, without actually getting useful experience and learning, so their usefulness would go down. Similar to standardized tests I guess, that are probably a great measure of ability until teachers start targeting them.


I think it's important to be careful with the vocabulary here (if only because of the site we're on).

There can be metrics that are targets which still remain good metrics. For example, in many machine learning competitions, the submissions optimize a known, given metric; but the test data is not known. Therefore, it is still a good metric.

I agree with the sentiment in this case, though.


I don't doubt that this would eventually be the case, people always learn to game the system eventually.


Based on the article, the scan basically measures how much you are using muscle-memory.

I am not an expert. But I'm inclined to believe that a brain scan taking while performing programming tasks would mostly measure how good of a touch-typist you are.


Nice! also scan if you are likely to cause any trouble for the company or are high medical risk therefore elevating costs for the company.


TheBrainBook.com : Find the right match for your sulci!


Phrenology 2.0


Right, like Astronomy is Astrology 2.0 and Chemistry is Alchemy 2.0.


Well, that's not entirely fair. Identifying a difference between an experienced surgeon an inexperienced is not at all the same as assessing skill in some meaningful manner, but could end up being used as such.

While phrenology ended up really far into, or even outside the fringes of science, one could argue that it was "based" on a true fact. A zero volume brain is going to be less useful than a brain with non zero volume. Then some rather brutal extrapolations happened, and the results were both terrible, and disastrous.

I fully believe the risk for a similar repeat of history is very real. Not because science fails us, but because society still posses many of its old weaknesses as clearly shown during the last year or so. As the performance improves, it is becoming more and more likely that we'll have to outlaw functional brain scanning for any non-medical purposes, or it will be somehow co-opted by actors wanting money or power at any price.

I don't think I'm either malignant, or creative enough to figure out how, but someone will, because the tech will become endlessly abusable when it becomes - or seems to be - sofisticated enough.


No, this is phrenology 2.0. (Have PhD in Neuroscience)


[flagged]


We need to ask you to do a much better job at posting civilly. We've asked before, and we ban accounts that won't, so please do!

https://news.ycombinator.com/newsguidelines.html


You didn't called it out as being p-hacked.

You called it out as not being possible in principle, that we will never be able to predict skills/whatever from a brain scan.


> literally an idiot

literally, idiot is someone who minds exclusively their own business. I don't see how the parent exhibited that trait.


code name "fMRI Neuroscience"


Maybe someone can make this an SaaS for dating app profiles.

I've heard girls only want boyfriends who have great skills.


Hm. Wonder if they have a brainscan for aholes.


Please don't post unsubstantive comments here.




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: