5 ms·

> this is not how the human brain works To some extent, to succeed in the modern world, you have to override your instincts with a thought-out response to thin

> this is not how the human brain works To some extent, to succeed in the modern world, you have to override your instincts with a thought-out response to things. Whatever heuristics we've built up over thousands of years are interesting, but some of them don't apply anymore. (If you were a prehistoric human and you encountered a beehive, you'd eat all the honey. But now you can get hundreds of beehives worth of honey in one trip to the grocery store, and your body's not going to tell you to not eat it. It will tell you the opposite -- this tastes great! You just have to learn that it's bad for your health and choose not to do it, no matter how good it would taste.) Basically, the human brain is a neural network trained on very old data. Fortunately, it is also very adaptive and can ignore that training data with some conscious effort. You need to operate your brain in that mode more now than you did 10,000 years ago. "Strong opinions weakly held" means "argue your point but sometimes you're going to lose." Evolution might not have rewarded losing over the course of millions of years, but whatever meeting you're arguing in doesn't have that kind of staying power. You can lose the argument and the human race will survive another generation. Whatever argument you're having probably doesn't matter in any meaningful sense. It's not life or death.

austincheney 6y ago
I have found that in software that is wrong. If you want to succeed in software you need to do what is most popular which largely follows innate human behavior. Unlike many other professions there is no licensing, accredidation, or common ethic which means there is no standard baseline of competence. This is visually explained when thinking about the difference between success and capability and then viewing that difference on a graph, such as a bell curve. https://news.ycombinator.com/item?id=23768137 https://news.ycombinator.com/item?id=23768137
- lucideer 6y ago
  I think it's (as is often the case) a little more nuanced than this. There's roughly two ways to follow trends "successfully" in the software world: 1. Eager and somewhat blind. This fits with your comment and follows the kind of innate behaviour we're discussing. And it will bring you some success: you'll always be up on "cool new thing" as long as you're afforded the freedom to jump from project to project and don't fall into a rut where you're tasked with maintaining the last crappy "cool new thing" you championed that noone wants anymore. 2. Pragmatic skepticism. You look at network effect and community support as rational, technical pros when weighing tech choices. You're successful because, while you're usually up on "cool new thing", you approach tech tentatively enough not to fall into the vendor lock-in hole. If you're ever tasked with maintaining some old thing you made, you're less screwed as you never jumped with both feet. Of the above two, the first is a much more common success story, largely because the software community loves reinventing the wheel, so most engineers don't fall into that particular maintenance rut. There's no moral hazard with approach number #1, but I still wouldn't go as far as to say it's the right way.
- opan 6y ago
  Free Software involves both licensing and ethics.
  - atq2119 6y ago
    Parent poster clearly meant licensing in the sense of accreditation for the people involved. I.e. there is no formal barrier to entry, which is great on the one hand, but it also means that there are pretty unqualified developers out there.
    - fapjacks 6y ago
      Anybody who's done hiring will attest that "formal barriers to entry" (i.e. accreditation via certifications and degrees) say very little about a programmer's ability to produce valuable output. Hiring would not be "the big problem" in tech if one could simply outsource the hiring decision to such a formal barrier to entry. But, since formal barriers to entry are demonstrably useless signals specifically for hiring programmers, we resort to all sorts of other superstitions which, when put together, at least bestow a visceral sense of being able to divine a person's ability to produce valuable output. People and companies all seem to have their own preferred superstitions, almost none of which include your "formal barriers to entry".
      - austincheney 6y ago
        > But, since formal barriers to entry are demonstrably useless Could you provide an example of such a barrier? Bar exams and medical review are extremely valuable to other industries. They demonstrate both competence and ethics. Since software is utterly lacking anything resembling a formal qualification barrier it’s hard to say a thing isn’t valuable. It doesn’t make sense to disqualify something that was never there in the first place. Industry ethical standards are also hard to disqualify because they similarly don’t exist. To someone who has never met such a burden for any professional industry it’s an easy thing to deem as worthless. In other industries compliance to ethical standards are more important than perceptions of competence.
- zbentley 6y ago
  > If you want to succeed in software you need to do what is most popular which largely follows innate human behavior I totally disagree. Many of the most valuable, respected, and successful software engineers I've encountered have evinced deep distrust of what is currently popular in the field. Their avoidance of (and advocacy for continued avoidance of) new/trendy software fads, particularly on the frontend, enabled excellent engineering, and instilled a powerful attitude of pragmatism and productivity among their colleagues. It's anecdata so take it with a grain, but in organizations big and small I've observed that some of the most successful and capable engineers are distinguished by antipathy towards popularity-driven engineering.
  - austincheney 6y ago
    > Their avoidance of (and advocacy for continued avoidance of) new/trendy software fads, particularly on the frontend, enabled excellent engineering I absolutely agree with that. You are describing an expert, which is not likely somebody successful or respected, particularly with regard to their front-end web technology peers. Expertise is not frequently revered by other developers, particularly when compared with compatibility. That distinction is even more evident when applying for jobs at other locations.
    - zbentley 6y ago
      Maybe I've just been lucky with where I've worked, but I've often seen those people get significantly rewarded for their anti-fad efforts--both in terms of money/promotions and respect from their colleagues. I've seen the opposite too, but not as often.
      - austincheney 6y ago
        A complete disdain for jQuery is what allowed me, as a full time JavaScript developer, to become promoted to a senior and nearly double my salary. That is only because I was working, at the time, in a niche area where jQuery failed amazingly in production. I was deemed an expert and rewarded for it, but that is astonishingly rare. Now you cannot get hired, much less attain any kind of success, without something like Angular or React even you can write superior code without those in half the time. That is why I only look for jobs that primarily deal with writing for Node, but even still competence is not well appreciated when interviewing.
netcan 6y ago
I agree, but that doesn't make the "old data" problem go away. It's worth noticing that in the first part (Paul Saffo forecasting), the problem is all in one head. Saffo recognizes an unintuitive methodology: strong opinions weakly held. He's only really dealing with himself, his intuitions, tactics that work for him. The second part gets social: "you’ve decided, along with your boss, to build a particular type of product for a particular subsection of the self-service checkout market." This is a good example, imo. These decisions are made with imperfect information, and they incorporate a lot of interlocking opinions. Ex ante can only get you so far. You need to make a decision and start rowing to progress, change paths when necessary. Multiple people are involved. Credibility is on the line, among many subtle subroutines of group psychology that we aren't even aware of. The solution from Tetlock, express your opinion as a probability, is a solution to the problem in its group form. When people are in meetings making big decisions, a lot more than the big decisions is going on. Human power dynamics are taking place. Being right can play a very small role in these. I think the author has a blindspot. "This is not how the human brain works" is one kind of challenge. "This is not how human groups work" is another. It isn't a derivative of the first.
- marcinzm 6y ago
  >Multiple people are involved. Credibility is on the line, among many subtle subroutines of group psychology that we aren't even aware of. The solution from Tetlock, express your opinion as a probability, is a solution to the problem in its group form. When people are in meetings making big decisions, a lot more than the big decisions is going on. Human power dynamics are taking place. Being right can play a very small role in these. Except expressing it in a probability form doesn't remove the group dynamics problem but rather makes it worse I feel. Or rather you're back to square one and gain little. The probability is no longer your confidence but the number you need to give to satisfy group dynamics. Not too high or you may be punished if the view turns out wrong. Also not one that disagrees with the people in charge too much. The original approach implicitly argues that without power dynamics you make better decisions.
- solidasparagus 6y ago
  Probability really doesn't help in a group context. In most cases you are pulling a number out of your ass, people don't have intuitive understanding of probability* , and it implies a level of certainty in the likelihood of an outcome that doesn't exist (which leads to bad decisions because you are building on a faulty basis). *look at how many people consider Trump's election impossible when the odds clearly indicated he had a non-negligible chance of success.
jt2190 6y ago
> "Strong opinions weakly held" means "argue your point but sometimes you're going to lose. The definition is not about arguing with others (so there’s no “looser”), it’s about arguing with yourself to avoid being wedded to poorly thought out points of view. From the article (emphasis mine): > Allow your intuition to guide you to a conclusion, no matter how imperfect — this is the “strong opinion” part. Then – and this is the “weakly held” part – prove yourself wrong.
- justin66 6y ago
  > The definition is not about arguing with others (so there’s no “looser”), it’s about arguing with yourself to avoid being wedded to poorly thought out points of view. That's one definition, but here is an exercise. If you meet someone who identifies with the phrase "strong opinions loosely held," ask yourself whether they're an argumentative asshole.
scott_s 6y ago
The author of this post knows what "strong opinions, weakly held" means. They explain it in their own words, and take effort to explain the difference between the catchphrase understanding and application of it versus the genuine application of its principles. The author also presents arguments as to why even the genuine application fall short, which I don't believe you responded to. (If it seems I'm picking on you: yours is the top post in the comments, and I don't feel like you're engaging with the post beyond a surface level.)
- chillacy 6y ago
  To go further, the Author also suggests an alternate technique along the lines of Annie Duke's Thinking in Bets and Superforecasting where probabilities are assigned to estimates. I had previously encountered such practice in Scott Alexander's blog and thought it clever: have skin in the game and keeping track of wins and losses. This whole section of the author's blog is very interesting. It goes into the US Intelligence Community's interest in superforecasters and the Good Judgement Project, a kind of elite tournament of forecasters. It talks about how superforecasters estimate by overcoming cognitive biases and questioning assumptions. And then it ultimately points out though that there are limits to forecasting and low probability events do occur all the time. https://commoncog.com/blog/the-forecasting-series/ https://commoncog.com/blog/the-forecasting-series/
WhompingWindows 6y ago
Bee Facts: You don't get hundreds of beehives worth of honey from a trip to the store. Hives typically produce multiple liters in a single summer. Last summer, our Hive made around 3 liters and there was enough for all the family to have their own jar. Many apiarists have multiple hives and a group of 10 hives could fill the grocery's honey section in September.
- travisjungroth 6y ago
  Subscribe!
  - WhompingWindows 6y ago
    Bee Facts: Honeybees have a radius of collection of around 3 miles, which means one hive collects pollen and nectar from around 28 square miles of land.
ouid 6y ago
A lot of people have replied to this, but in what way is a brain a neural network trained on old data? That does not make any sense at all.
- aspenmayer 6y ago
  You inherited the DNA which built your brain from your parents. I believe they refer to the historical conditions which created selection pressures for your parents’ ancestors to survive rather than fail to live long enough to reproduce.
  - ouid 6y ago
    DNA is not a neural network, (and neither is a brain, for that matter). The comment is still decidedly nonsensical.
    - aspenmayer 6y ago
      DNA is not a neural network (that we know of), and I wasn’t meaning to imply it was. However, the actual neural networks (that is, meatspace neurons arranged in a live human brain) are tested simultaneously alongside the DNA and they succeed or fail to be propagated as a result of their fitness in the environment; the interaction between DNA, behavior, and acquired traits was once thought to be solely fiction, but it is actual science[1], though we don’t know much about how this works yet. The brain is as involved with your behavior and your DNA; I think that was the common thread in OP’s point. [1] https://www.cell.com/fulltext/S0092-8674(11)01341-9 https://www.cell.com/fulltext/S0092-8674(11)01341-9

austincheney 6y ago

I have found that in software that is wrong. If you want to succeed in software you need to do what is most popular which largely follows innate human behavior. Unlike many other professions there is no licensing, accredidation, or common ethic which means there is no standard baseline of competence. This is visually explained when thinking about the difference between success and capability and then viewing that difference on a graph, such as a bell curve. https://news.ycombinator.com/item?id=23768137 https://news.ycombinator.com/item?id=23768137

lucideer 6y ago
I think it's (as is often the case) a little more nuanced than this. There's roughly two ways to follow trends "successfully" in the software world: 1. Eager and somewhat blind. This fits with your comment and follows the kind of innate behaviour we're discussing. And it will bring you some success: you'll always be up on "cool new thing" as long as you're afforded the freedom to jump from project to project and don't fall into a rut where you're tasked with maintaining the last crappy "cool new thing" you championed that noone wants anymore. 2. Pragmatic skepticism. You look at network effect and community support as rational, technical pros when weighing tech choices. You're successful because, while you're usually up on "cool new thing", you approach tech tentatively enough not to fall into the vendor lock-in hole. If you're ever tasked with maintaining some old thing you made, you're less screwed as you never jumped with both feet. Of the above two, the first is a much more common success story, largely because the software community loves reinventing the wheel, so most engineers don't fall into that particular maintenance rut. There's no moral hazard with approach number #1, but I still wouldn't go as far as to say it's the right way.
opan 6y ago
Free Software involves both licensing and ethics.
- atq2119 6y ago
  Parent poster clearly meant licensing in the sense of accreditation for the people involved. I.e. there is no formal barrier to entry, which is great on the one hand, but it also means that there are pretty unqualified developers out there.
  - fapjacks 6y ago
    Anybody who's done hiring will attest that "formal barriers to entry" (i.e. accreditation via certifications and degrees) say very little about a programmer's ability to produce valuable output. Hiring would not be "the big problem" in tech if one could simply outsource the hiring decision to such a formal barrier to entry. But, since formal barriers to entry are demonstrably useless signals specifically for hiring programmers, we resort to all sorts of other superstitions which, when put together, at least bestow a visceral sense of being able to divine a person's ability to produce valuable output. People and companies all seem to have their own preferred superstitions, almost none of which include your "formal barriers to entry".
    - austincheney 6y ago
      > But, since formal barriers to entry are demonstrably useless Could you provide an example of such a barrier? Bar exams and medical review are extremely valuable to other industries. They demonstrate both competence and ethics. Since software is utterly lacking anything resembling a formal qualification barrier it’s hard to say a thing isn’t valuable. It doesn’t make sense to disqualify something that was never there in the first place. Industry ethical standards are also hard to disqualify because they similarly don’t exist. To someone who has never met such a burden for any professional industry it’s an easy thing to deem as worthless. In other industries compliance to ethical standards are more important than perceptions of competence.
zbentley 6y ago
> If you want to succeed in software you need to do what is most popular which largely follows innate human behavior I totally disagree. Many of the most valuable, respected, and successful software engineers I've encountered have evinced deep distrust of what is currently popular in the field. Their avoidance of (and advocacy for continued avoidance of) new/trendy software fads, particularly on the frontend, enabled excellent engineering, and instilled a powerful attitude of pragmatism and productivity among their colleagues. It's anecdata so take it with a grain, but in organizations big and small I've observed that some of the most successful and capable engineers are distinguished by antipathy towards popularity-driven engineering.
- austincheney 6y ago
  > Their avoidance of (and advocacy for continued avoidance of) new/trendy software fads, particularly on the frontend, enabled excellent engineering I absolutely agree with that. You are describing an expert, which is not likely somebody successful or respected, particularly with regard to their front-end web technology peers. Expertise is not frequently revered by other developers, particularly when compared with compatibility. That distinction is even more evident when applying for jobs at other locations.
  - zbentley 6y ago
    Maybe I've just been lucky with where I've worked, but I've often seen those people get significantly rewarded for their anti-fad efforts--both in terms of money/promotions and respect from their colleagues. I've seen the opposite too, but not as often.
    - austincheney 6y ago
      A complete disdain for jQuery is what allowed me, as a full time JavaScript developer, to become promoted to a senior and nearly double my salary. That is only because I was working, at the time, in a niche area where jQuery failed amazingly in production. I was deemed an expert and rewarded for it, but that is astonishingly rare. Now you cannot get hired, much less attain any kind of success, without something like Angular or React even you can write superior code without those in half the time. That is why I only look for jobs that primarily deal with writing for Node, but even still competence is not well appreciated when interviewing.

netcan 6y ago

I agree, but that doesn't make the "old data" problem go away. It's worth noticing that in the first part (Paul Saffo forecasting), the problem is all in one head. Saffo recognizes an unintuitive methodology: strong opinions weakly held. He's only really dealing with himself, his intuitions, tactics that work for him. The second part gets social: "you’ve decided, along with your boss, to build a particular type of product for a particular subsection of the self-service checkout market." This is a good example, imo. These decisions are made with imperfect information, and they incorporate a lot of interlocking opinions. Ex ante can only get you so far. You need to make a decision and start rowing to progress, change paths when necessary. Multiple people are involved. Credibility is on the line, among many subtle subroutines of group psychology that we aren't even aware of. The solution from Tetlock, express your opinion as a probability, is a solution to the problem in its group form. When people are in meetings making big decisions, a lot more than the big decisions is going on. Human power dynamics are taking place. Being right can play a very small role in these. I think the author has a blindspot. "This is not how the human brain works" is one kind of challenge. "This is not how human groups work" is another. It isn't a derivative of the first.

marcinzm 6y ago
>Multiple people are involved. Credibility is on the line, among many subtle subroutines of group psychology that we aren't even aware of. The solution from Tetlock, express your opinion as a probability, is a solution to the problem in its group form. When people are in meetings making big decisions, a lot more than the big decisions is going on. Human power dynamics are taking place. Being right can play a very small role in these. Except expressing it in a probability form doesn't remove the group dynamics problem but rather makes it worse I feel. Or rather you're back to square one and gain little. The probability is no longer your confidence but the number you need to give to satisfy group dynamics. Not too high or you may be punished if the view turns out wrong. Also not one that disagrees with the people in charge too much. The original approach implicitly argues that without power dynamics you make better decisions.
solidasparagus 6y ago
Probability really doesn't help in a group context. In most cases you are pulling a number out of your ass, people don't have intuitive understanding of probability* , and it implies a level of certainty in the likelihood of an outcome that doesn't exist (which leads to bad decisions because you are building on a faulty basis). *look at how many people consider Trump's election impossible when the odds clearly indicated he had a non-negligible chance of success.

jt2190 6y ago

> "Strong opinions weakly held" means "argue your point but sometimes you're going to lose. The definition is not about arguing with others (so there’s no “looser”), it’s about arguing with yourself to avoid being wedded to poorly thought out points of view. From the article (emphasis mine): > Allow your intuition to guide you to a conclusion, no matter how imperfect — this is the “strong opinion” part. Then – and this is the “weakly held” part – prove yourself wrong.

justin66 6y ago
> The definition is not about arguing with others (so there’s no “looser”), it’s about arguing with yourself to avoid being wedded to poorly thought out points of view. That's one definition, but here is an exercise. If you meet someone who identifies with the phrase "strong opinions loosely held," ask yourself whether they're an argumentative asshole.

scott_s 6y ago

The author of this post knows what "strong opinions, weakly held" means. They explain it in their own words, and take effort to explain the difference between the catchphrase understanding and application of it versus the genuine application of its principles. The author also presents arguments as to why even the genuine application fall short, which I don't believe you responded to. (If it seems I'm picking on you: yours is the top post in the comments, and I don't feel like you're engaging with the post beyond a surface level.)

chillacy 6y ago
To go further, the Author also suggests an alternate technique along the lines of Annie Duke's Thinking in Bets and Superforecasting where probabilities are assigned to estimates. I had previously encountered such practice in Scott Alexander's blog and thought it clever: have skin in the game and keeping track of wins and losses. This whole section of the author's blog is very interesting. It goes into the US Intelligence Community's interest in superforecasters and the Good Judgement Project, a kind of elite tournament of forecasters. It talks about how superforecasters estimate by overcoming cognitive biases and questioning assumptions. And then it ultimately points out though that there are limits to forecasting and low probability events do occur all the time. https://commoncog.com/blog/the-forecasting-series/ https://commoncog.com/blog/the-forecasting-series/

WhompingWindows 6y ago

Bee Facts: You don't get hundreds of beehives worth of honey from a trip to the store. Hives typically produce multiple liters in a single summer. Last summer, our Hive made around 3 liters and there was enough for all the family to have their own jar. Many apiarists have multiple hives and a group of 10 hives could fill the grocery's honey section in September.

travisjungroth 6y ago
Subscribe!
- WhompingWindows 6y ago
  Bee Facts: Honeybees have a radius of collection of around 3 miles, which means one hive collects pollen and nectar from around 28 square miles of land.

ouid 6y ago

A lot of people have replied to this, but in what way is a brain a neural network trained on old data? That does not make any sense at all.

aspenmayer 6y ago
You inherited the DNA which built your brain from your parents. I believe they refer to the historical conditions which created selection pressures for your parents’ ancestors to survive rather than fail to live long enough to reproduce.
- ouid 6y ago
  DNA is not a neural network, (and neither is a brain, for that matter). The comment is still decidedly nonsensical.
  - aspenmayer 6y ago
    DNA is not a neural network (that we know of), and I wasn’t meaning to imply it was. However, the actual neural networks (that is, meatspace neurons arranged in a live human brain) are tested simultaneously alongside the DNA and they succeed or fail to be propagated as a result of their fitness in the environment; the interaction between DNA, behavior, and acquired traits was once thought to be solely fiction, but it is actual science[1], though we don’t know much about how this works yet. The brain is as involved with your behavior and your DNA; I think that was the common thread in OP’s point. [1] https://www.cell.com/fulltext/S0092-8674(11)01341-9 https://www.cell.com/fulltext/S0092-8674(11)01341-9