3 ms·

It sounds like you're advocating a system of education, thinking, and work that is exemplified by 'Software Architecture Astronauts', people who have not recent

It sounds like you're advocating a system of education, thinking, and work that is exemplified by 'Software Architecture Astronauts', people who have not recently or ever contended with the details, the nuts and bolts, of actual software development and maintenance. To be honest, the idea that you can produce valuable intellectual work by outsourcing the details is absurd. This would have seen me absolutely fail my challenging philosophy and engineering coursework, and end up a fundamentally incompetent dilettante in work and life.

sitkack 4y ago
We can do both, ChatGPT can code. So the Astronauts should be able to produce working systems that pass functional tests. The majority of the detractors of these models don't see the forest for the trees. It only gives new capabilities, they remove nothing.
- thundergolfer 4y ago
  > It only gives new capabilities, they remove nothing. This is not how technology works in the world. See the arguments of for example McLuhan. Technologies are appendages for the human body or intellect, and at the same time they numb the part of that they amplify. Also the point of the Architecture Astronauts idea is that they don't produce working systems. It's a specific pejorative, it doesn't refer generally to people that work at a high-level without harming the systems they engage with.
  - sitkack 4y ago
    You clearly have a bone to pick with software architects, if anything it should reduce the amount of Space Travel for said architects. McLuhan, a media analysis astronaut. Average is the message. > Technologies are appendages for the human body or intellect We previously had fleets of people calculating tables of numbers for hours per day. Simple calculators removed that need and increased the speed of results by orders of magnitude. You seem to be arguing for handicap system, an intellectual conservatism. Is that correct? What does ChatGPT remove from ones ability to write low level code or for a software architect to both architect a system AND produce a working model?
- wpietri 4y ago
  > So the Astronauts should be able to produce working systems that pass functional tests. Nah. This is just a new version of an old mistake, one people have been repeating for decades. People keep trying to come up with ways for people to "code" without understanding what's going on. Code generation wizards. Visual programming tools. Model driven architecture. And a bunch more. The hard part about making software isn't writing a bit of starter code. What we saw with the "code wizard" approach is that somebody clueless could click some buttons and get something working. But then they couldn't maintain it. It just kicked the "understand what's going on" problem down the road from "I don't know how to start" to "I am now trapped in a hell of generated code and people are yelling at me". If some particular application is so standard that it can be produced with only a shallow understanding of how software works, then the right thing isn't using ChatGPT to produce source code. It's when people who have a deeper understanding produce an app that is configurable in the right ways.
  - sitkack 4y ago
    My 8 year old asks you to look at from a different angle and evaluate what it can actually do right now. I would add, just because someone made a prognostication decades ago does not invalidate the same prognostication now. Have you tried it?
    - wpietri 4y ago
      Those aren't decades-old "prognostications". They are things that happened. Could it be different this time? Maybe! But if you want to say that it will, then you have to make the argument. Or maybe you can get ChatGPT to do it for you?
      - sitkack 4y ago
        Have you tried coding with ChatGPT? It works now. You can use it now to architect a system AND produce working code.
        wpietri 4y ago
lordnacho 4y ago
Well I didn't say you should never concern yourself with the details. But at some point in your life you will have gone through the details so often you won't be getting them wrong, but you'll still be spending a lot of time on them. With a tool like cgpt you can save yourself a lot of time. Of course it's up to you where you take the trade. If your thing is a medical device that will kill someone if it isn't checked, maybe you carefully write the code yourself. If you're building some POC website, why not let the AI fill in a few API calls for you? And don't forget, plenty of details are already outsourced. Entire reference books are full of details that we prefer not to keep memorized. Plato: “If men learn this, it will implant forgetfulness in their souls; they will cease to exercise memory because they rely on that which is written, calling things to remembrance no longer from within themselves, but by means of external marks. What you have discovered is a recipe not for memory, but for reminder. And it is no true wisdom that you offer your disciples, but only its semblance, for by telling them of many things without teaching them you will make them seem to know much, while for the most part they know nothing, and as men filled, not with wisdom, but with the conceit of wisdom, they will be a burden to their fellows.” https://www.goodreads.com/quotes/259062-if-men-learn-this-it-will-implant-forgetfulness-in-their https://www.goodreads.com/quotes/259062-if-men-learn-this-it... I'm advocating cgpt as a tool that allows you to spend more time at the level of abstraction that is appropriate for the problem. If you're writing an essay about the fall of Rome, you don't want to be spending loads of time finding the details of a reference (year published, secondary author's names, exact title...). You want to be talking about economics of the time, incentives for different groups in Roman society, and so on.
- thundergolfer 4y ago
  Right, it seems like you're advocating ChatGPT as an information retreival tool, which seems perfectly acceptable. Looking up exact dates and names is tedious. By "filling in the details" I thought you were referring more generally to the essay writing process, where the structure of an essay is its skeleton and the "details" are the just as essential pieces of argumentation and evidence. If you write a sociology, history, or philosophy essay, for example, writing the high level structure (intro, topic-sentence 1, topic-sentence 2, ..., conclusion) and then outsourcing the rest would be intellectually bankrupt. You can't understand produce a synthesis argument of Hume's The Wealth of Nations and Theory of Moral Sentiments with ChatGPT.
wpietri 4y ago
Great points! But I want to quibble with this a bit. > To be honest, the idea that you can produce valuable intellectual work by outsourcing the details is absurd. It is absurd. But it's also very common! There are an awful lot of people who have done very well not by understanding anything, but by posturing and performing. E.g., take the ghostwritten book. The books of many politicians from all parties could generously be called mediocre. But they don't need to create valuable intellectual work; they just need something anodyne and book-shaped with their name on it. Surprisingly often, people get away with this kind of absurdity, especially when rich and/or powerful. Indeed, one of the fascinating things for me about Musk's ongoing clown show at Twitter is that we're getting a real-time look at what happens when somebody who refuses to contend with the details fools themself into thinking they know what's going on. That totally matches my experience where Architecture Astronauts have free reign. They are rewarded not for making this work, but talking impressively about big things. So what they produce is a lot of talk. When it's implemented by actual developers, does it work out well? Usually not! But that doesn't matter, because the people who empower them are also people who talk impressively about big things without understanding them.
- thundergolfer 4y ago
  Totally endorse the quibble. I think ChatGPT will be most often deployed in areas that are tied to a lot of economic value but are also full of bullshitters. Ghostwriting political memoirs is a good example, but its a niche, SEO spam is that big bucks at the moment.

sitkack 4y ago

We can do both, ChatGPT can code. So the Astronauts should be able to produce working systems that pass functional tests. The majority of the detractors of these models don't see the forest for the trees. It only gives new capabilities, they remove nothing.

thundergolfer 4y ago
> It only gives new capabilities, they remove nothing. This is not how technology works in the world. See the arguments of for example McLuhan. Technologies are appendages for the human body or intellect, and at the same time they numb the part of that they amplify. Also the point of the Architecture Astronauts idea is that they don't produce working systems. It's a specific pejorative, it doesn't refer generally to people that work at a high-level without harming the systems they engage with.
- sitkack 4y ago
  You clearly have a bone to pick with software architects, if anything it should reduce the amount of Space Travel for said architects. McLuhan, a media analysis astronaut. Average is the message. > Technologies are appendages for the human body or intellect We previously had fleets of people calculating tables of numbers for hours per day. Simple calculators removed that need and increased the speed of results by orders of magnitude. You seem to be arguing for handicap system, an intellectual conservatism. Is that correct? What does ChatGPT remove from ones ability to write low level code or for a software architect to both architect a system AND produce a working model?
wpietri 4y ago
> So the Astronauts should be able to produce working systems that pass functional tests. Nah. This is just a new version of an old mistake, one people have been repeating for decades. People keep trying to come up with ways for people to "code" without understanding what's going on. Code generation wizards. Visual programming tools. Model driven architecture. And a bunch more. The hard part about making software isn't writing a bit of starter code. What we saw with the "code wizard" approach is that somebody clueless could click some buttons and get something working. But then they couldn't maintain it. It just kicked the "understand what's going on" problem down the road from "I don't know how to start" to "I am now trapped in a hell of generated code and people are yelling at me". If some particular application is so standard that it can be produced with only a shallow understanding of how software works, then the right thing isn't using ChatGPT to produce source code. It's when people who have a deeper understanding produce an app that is configurable in the right ways.
- sitkack 4y ago
  My 8 year old asks you to look at from a different angle and evaluate what it can actually do right now. I would add, just because someone made a prognostication decades ago does not invalidate the same prognostication now. Have you tried it?
  - wpietri 4y ago
    Those aren't decades-old "prognostications". They are things that happened. Could it be different this time? Maybe! But if you want to say that it will, then you have to make the argument. Or maybe you can get ChatGPT to do it for you?
    - sitkack 4y ago
      Have you tried coding with ChatGPT? It works now. You can use it now to architect a system AND produce working code.
      - wpietri 4y ago

lordnacho 4y ago

Well I didn't say you should never concern yourself with the details. But at some point in your life you will have gone through the details so often you won't be getting them wrong, but you'll still be spending a lot of time on them. With a tool like cgpt you can save yourself a lot of time. Of course it's up to you where you take the trade. If your thing is a medical device that will kill someone if it isn't checked, maybe you carefully write the code yourself. If you're building some POC website, why not let the AI fill in a few API calls for you? And don't forget, plenty of details are already outsourced. Entire reference books are full of details that we prefer not to keep memorized. Plato: “If men learn this, it will implant forgetfulness in their souls; they will cease to exercise memory because they rely on that which is written, calling things to remembrance no longer from within themselves, but by means of external marks. What you have discovered is a recipe not for memory, but for reminder. And it is no true wisdom that you offer your disciples, but only its semblance, for by telling them of many things without teaching them you will make them seem to know much, while for the most part they know nothing, and as men filled, not with wisdom, but with the conceit of wisdom, they will be a burden to their fellows.” https://www.goodreads.com/quotes/259062-if-men-learn-this-it-will-implant-forgetfulness-in-their https://www.goodreads.com/quotes/259062-if-men-learn-this-it... I'm advocating cgpt as a tool that allows you to spend more time at the level of abstraction that is appropriate for the problem. If you're writing an essay about the fall of Rome, you don't want to be spending loads of time finding the details of a reference (year published, secondary author's names, exact title...). You want to be talking about economics of the time, incentives for different groups in Roman society, and so on.

thundergolfer 4y ago
Right, it seems like you're advocating ChatGPT as an information retreival tool, which seems perfectly acceptable. Looking up exact dates and names is tedious. By "filling in the details" I thought you were referring more generally to the essay writing process, where the structure of an essay is its skeleton and the "details" are the just as essential pieces of argumentation and evidence. If you write a sociology, history, or philosophy essay, for example, writing the high level structure (intro, topic-sentence 1, topic-sentence 2, ..., conclusion) and then outsourcing the rest would be intellectually bankrupt. You can't understand produce a synthesis argument of Hume's The Wealth of Nations and Theory of Moral Sentiments with ChatGPT.

wpietri 4y ago

Great points! But I want to quibble with this a bit. > To be honest, the idea that you can produce valuable intellectual work by outsourcing the details is absurd. It is absurd. But it's also very common! There are an awful lot of people who have done very well not by understanding anything, but by posturing and performing. E.g., take the ghostwritten book. The books of many politicians from all parties could generously be called mediocre. But they don't need to create valuable intellectual work; they just need something anodyne and book-shaped with their name on it. Surprisingly often, people get away with this kind of absurdity, especially when rich and/or powerful. Indeed, one of the fascinating things for me about Musk's ongoing clown show at Twitter is that we're getting a real-time look at what happens when somebody who refuses to contend with the details fools themself into thinking they know what's going on. That totally matches my experience where Architecture Astronauts have free reign. They are rewarded not for making this work, but talking impressively about big things. So what they produce is a lot of talk. When it's implemented by actual developers, does it work out well? Usually not! But that doesn't matter, because the people who empower them are also people who talk impressively about big things without understanding them.

thundergolfer 4y ago
Totally endorse the quibble. I think ChatGPT will be most often deployed in areas that are tied to a lot of economic value but are also full of bullshitters. Ghostwriting political memoirs is a good example, but its a niche, SEO spam is that big bucks at the moment.