[HN Gopher] The Commission for Stopping Further Improvements
       ___________________________________________________________________
        
       The Commission for Stopping Further Improvements
        
       Author : feross
       Score  : 71 points
       Date   : 2023-04-21 17:47 UTC (5 hours ago)
        
 (HTM) web link (rootsofprogress.org)
 (TXT) w3m dump (rootsofprogress.org)
        
       | carls wrote:
       | It seems to me that there's a natural tension in emerging fields
       | of science and engineering between establishing clear guidelines
       | and regulations early on to minimize harms, or instead allowing
       | practitioners to experiment, tinker, build and create outcomes
       | that may be potentially harmful.
       | 
       | What are some frameworks for how to think about navigating this
       | tension in emerging scientific or engineering fields?
       | 
       | Some ones I'm mulling over:
       | 
       | 1. Rate of innovation: In rapidly evolving fields, imposing
       | strict regulations too early can hinder innovation and progress.
       | In such cases, it might be better to minimize restrictions early
       | on to allow practitioners to explore new ideas. Then, as the
       | field matures, regulations and standards can be gradually
       | introduced.
       | 
       | 2. Adaptive regulation: Implement a flexible regulatory framework
       | that can be updated as new information becomes available.
       | 
       | 3. Self-regulation: In some cases, maybe we should expect and
       | encourage the industry to use self-regulation via developing
       | guidelines and codes of conduct. This may be one way to try and
       | strike a balance between responsible innovation while minimizing
       | bureaucratic obstacles.
       | 
       | What do others think?
        
         | TechBro8615 wrote:
         | I think if I want to write some software, and I have the
         | knowledge and compute power to do what I want, then I really
         | don't care whether you or some authoritarian committee tells me
         | I'm allowed to run my software. I'm going to do it anyway.
        
           | nine_k wrote:
           | Now think of synthesizing and distributing novel chemical
           | compounds, or using novel medical devices.
        
       | zeteo wrote:
       | Brunel was probably right that the work of the commission would
       | be detrimental to the rate of innovation. But is that the only
       | thing that matters? We would probably get a faster rate of
       | innovation in personal transportation if we allowed any
       | unlicensed vehicles to use the public highways at any speed. But
       | there would be a terrible cost to pay in terms of lives lost or
       | ruined. It doesn't sound like the commission was planning to
       | forbid any and all usage of cast iron, just not in places where
       | public safety was at risk. An innovative engineer could always
       | build a novel cast iron bridge in a test location and produce
       | some dramatic demonstration of its strength and reliability that
       | didn't involve running fully loaded passenger trains over it. The
       | trade-off between the rate of innovation and the casualty rate
       | does not need to be chosen at either of the extremes.
        
       | mannykannot wrote:
       | While some counter-examples may be found, subsequent events have
       | generally repudiated Brunel's concerns.
        
       | mnd999 wrote:
       | > The interpretation of Brunel's opinions, and applicability to
       | today, are left to the reader
       | 
       | I don't think Brunel would have much liked SOC2.
        
       | IshKebab wrote:
       | Interesting, and I guess this applies to plenty of modern
       | standards (e.g. my work won't use wireguard because it isn't
       | approved by some nonsense security standards.
       | 
       | But on the other hand I don't think the sensible solution is to
       | leave everything unregulated! I am very glad that building
       | regulations exist!
        
         | jasoncrawford wrote:
         | I think regulations can work when they enforce very well-
         | supported, long-established best practices, such that if you
         | don't do them it amounts to negligence.
         | 
         | They might also work better if they say "you can't do it _that_
         | way, which is known to be unsafe," as opposed to "you must do
         | it _this_ way, which is the only safe thing."
         | 
         | Note also that regulatory standards are not the only mechanism
         | in the law to create safety. Liability law can be very
         | effective at creating safety, by giving the right incentives to
         | the right parties, but liability law doesn't tell anyone what
         | to do--only what will happen to you if you cause harm.
        
         | nathan_compton wrote:
         | Yeah, this "we can't build anymore" shit is _often_ just people
         | complaining about not being able to externalize costs that they
         | don't want to pay for.
        
         | zwieback wrote:
         | Agreed, it's a fine anecdote but easy to draw the wrong
         | conclusion of "rules hamper progress and counterproductive". I
         | think the main challenge is to attract the right talent to the
         | rule-making body.
        
           | 364758483 wrote:
           | This is the exact direction I think we should reshape
           | political systems. Perhaps even make empowering experts the
           | main point. Too much is legislated and enforced by people who
           | are overly self-interested or out of touch. It's a huge
           | missed opportunity that experts are forced to watch
           | representatives fumble around or make wrong decisions.
           | 
           | A long those lines, governance power should not require being
           | a celebrity either.
           | 
           | Applying this post to the topic of AI alignment, I'd like a
           | democratic option for large entitlements of tax funds to be
           | applied as Yudkowsky sees fit, if enough people vote to
           | appoint him as AI Minister, and vote to give him governance
           | teeth against big tech and the thousands of startups driving
           | Moloch via the standard economic paradigm.
        
         | m463 wrote:
         | There must be ways to write regulations in meaningful ways.
         | Like a train bridge must meet specification a, b and c and pass
         | tests d, e and f, instead of saying it must be made of wrought
         | iron.
        
       | eximius wrote:
       | So, 'update regulations when new information comes to light'?
       | 
       | It is frustrating that common action often precludes such common
       | sense sentiments due to the friction of interests or bad actors.
        
         | dgs_sgd wrote:
         | Ah, but regulatory updates come at a rate much much slower than
         | new information and innovation. For example, the main
         | securities law in the US was written 90 years ago (Securities
         | Act of 1933).
        
           | JackFr wrote:
           | The Securities Act of 1933 is a tremendously successful law.
           | 
           | (And it's been amended a half dozen times so it's not really
           | trapped in amber.)
        
       | mortenjorck wrote:
       | Much like a Commission for Stopping Further Improvements, a
       | Coalition for Stopping Further Regulation is equally short-
       | sighted.
       | 
       | Brunel's contention that structural engineers should be qualified
       | to make their own decisions unencumbered by regulation makes
       | sense on the face of it. But then, what qualifies a structural
       | engineer? Licensing, a product of regulation.
       | 
       | Like many things, regulation is, in isolation, neutral. It can
       | serve the common good, or it can serve the interests of an
       | unrepresentative cabal; it can be forward-looking and adaptable,
       | or it can be hastily drafted by unqualified functionaries. Why
       | couldn't the regulators in Brunel's day have carved out an open-
       | ended "unless otherwise demonstrated" around the use of iron
       | structural members?
        
         | its_ethan wrote:
         | It feels like you answered your own question in a way..
         | 
         | "Why couldn't the regulators in Brunel's day have carved out an
         | open-ended "unless otherwise demonstrated" around the use of
         | iron structural members?"
         | 
         | The answer is partly in your statement: "regulation is, in
         | isolation, neutral." Looking around at regulatory/ bureaucratic
         | bodies, "in isolation" is rarely (never?) achieved. Satisfying
         | ambiguous "unless otherwise demonstrated"-esque clauses that
         | are enforced by unelected paper pushers is very much part of
         | Brunel's concern.
        
       | flybrand wrote:
       | I feel like we're living in the age of an unidentified,
       | "Committee to Stop Maintenance of Civil Infrastructue."
        
       | krisoft wrote:
       | But what is Brunel's solution to the underlying problem? A bridge
       | has fallen down. Presumably the engineers who designed it had no
       | murderous intent. They tried their best to make a solid bridge.
       | Do we just let engineers build bridges however they please and
       | then be angry if they fall down? How is that going to help
       | anybody?
        
       | netbioserror wrote:
       | In economics, the closest analogy might be economic substitutes.
       | Who was to predict that electricity would displace oil lamps but
       | the electric pioneers? The car to replace the horse-drawn
       | carriage? Internal combustion to replace steam? The standard
       | shipping container to replace irregular loading? Integrated
       | circuits to replace mazes of wires and vacuum tubes? Spreadsheets
       | to replace calculation workers? The Internet to quite nearly
       | replace all uses of technical books? Government planners are not
       | these pioneers and can never make universally applicable rules
       | which hold their fairness into an unknowable future.
        
         | narrator wrote:
         | The worst outcome is when government promotes and protects bad
         | science for ideological reasons. The classic examples being
         | Lysenkoism in the Soviet Union and Mao's terrible ideas during
         | The Great Leap Forward and subsequent famine.
        
       ___________________________________________________________________
       (page generated 2023-04-21 23:00 UTC)