Thought as a Technology | Hacker News Reader

Thought as a Technology

cognitivemedium.com

148 pointschewxy22 comments

I think there's a subtle danger to "cognitive outsourcing", or at least an important distinction to be aware of. "Cognitive outsourcing" works like an oracle, but many of our best tools work more like bicycles: very direct, very fluid, very analog, with continuous feedback that helps us build an constructive intuition. Bicycles extend the range of what a human can do; oracles do something that a human tells them to do.

Sometimes this difference is a function of how information is displayed and how it reacts to the user. Terminals and REPLs are inherently oracles; spreadsheets connect our fingertips directly to raw information.

Other times it’s just a difference in the conceptual model that users construct. Some people think of Google as an oracle ("what is the weather like?"); others as a bicycle ("weather"). Those who bicycle around Google aren’t just “better at Googling”; they have a fundamentally different view of what they're doing.

This isn’t to say that oracles are inherently wrong or that bicycles are always better, but there’s a huge difference between truly augmenting a human and merely interfacing with one. It’s important to know which idea is appropriate for any given problem space.

(most of this comment was ripped from here: http://joelgustafson.com/ideas/2016/08/25/oracles-and-bicycl...)

(see also http://mental.bike, which I don't know why I own)

I think your bicycle/oracle distinction is interesting. It seems like the tradeoffs in preferring a bicycle to an oracle are similar to considerations leading up to decisions not to abstract. If we could assume the oracle would fulfill its function perfectly, there would be no real reason to prefer the bicycle; unfortunately, however, experience shows us that the internals of complex black boxes we construct are rarely perfect, and that opening the box may be necessary, and unpleasant. On the other hand, with a bicycle, a feeling of control remains throughout: it's always understood how some action on the bicycle will evolve the state of your situation. The downside for the bicycle is that its definition requires it to remain relatively simple.

My first thought for getting around this was something like formal verification that a black box does what it says—but once 'what it says' is too complex for a person to really understand (e.g. a black box containing the behavior of a human brain), you've hit a wall with this approach.

The other thing I can imagine is building up 'bicycles' which have parts of proven-correct medium complexity, arranged together in a single higher-level system also of proven-correct medium complexity (where medium complexity is: more complex than bike parts, but still with readily human-comprehensible semantics). I'd characterize these systems as sort of like Iron Man suits. Maybe that doesn't really add anything of practical benefit beyond the original bicycle analogy though ;)

>If we could assume the oracle would fulfill its function perfectly, there would be no real reason to prefer the bicycle

The bicycle might be more fun to ride

Not only might it be more fun, but you'd build a better intuition about what to expect from it (or even how it works).

It's like how people often make fewer mistakes with slide-rules or abacuses than pocket calculators, since the former two require you to check your answer against your intuition, while the latter just tells you "the answer" to your (potentially ill-formed) query. Trusting a black box can be quite dangerous.

I threw 'real' in there to cover that. Maybe that's not fair though ;)

More seriously though: if the oracle were perfect, you should have to do almost nothing (i.e. the relative time savings would be near 100%; e.g. creating pizza: the oracle, you just tell: "get me pizza", whereas with the bike, you have to bike there—which I admit may be fun)—so as long as you have things you enjoy doing more in your non-productive free time, then it shouldn't be an issue. (Not that I think that's necessarily likely...)

Part of this reminds me of a project idea I've had in the past: figuring out how to teach people an 'abstract visual relationship' language. I'll quote from the article:

> ... he immediately responded that when he taught algebra courses, if he was discussing cyclic subgroups of a group, he had a mental image of group elements breaking into a formation organized into circular groups.

Jacque Hadamard once conducted a study of how mathematicians approach their work, which can still be found in a book called "The Psychology of Invention in the Mathematical Field". Here's an excerpt:

> Indeed, every mathematical research compels me to build such a schema, which is always and must be of a vague character, so as not to be deceptive. I shall give a less elementary example from my first researches (my thesis). I had to consider a sum of an infinite number of terms, intending to valuate its order of magnitude. In that case, there is a group of terms which chances to be predominant, all others having a negligible influence. Now, when I think of that question, I see not the formula itself, but the place it would take if written: a kind of ribbon, which is thicker or darker at the place corresponding to the possibly important terms; or (at other moments), I see something like a formula, but by no means a legible one, as I should see it (being strongly long-sighted) if I had no eye-glasses on, which letters seeming rather more apparent (though still not legible) at the place which is supposed to be the important one.

It seems to me that this is a surprisingly common approach to dealing with highly abstract subject matter. I first noticed myself doing it while reading Structure of Scientific Revolutions and later used it intentionally in working on math and CS stuff.

My experience with it so far leads me to believe it can be taught/strengthened. I've written a much more in depth essay on this in the past, but it's fairly unfinished. I still wonder about it though...

Please release your essay! You don't even have to finish it. You could even email it to me without commentary, and I would ask you questions about it.

This is interesting, but combines about five ideas. Expanding any one of those would be useful. The notion of a "transformative interface" seems to combine two concepts - "Wow factor" and "representation that yields insight". Those are different. The latter is more useful (but the former is more profitable.) Feynman diagrams come to mind.

Sussman's talk on how to think about circuits is here.[1] Here are the slides.[2] The video shows Sussman's talking head and him pointing at an off-screen display of the slides, which is not too helpful.

[1] https://www.infoq.com/presentations/We-Really-Dont-Know-How-... [2] http://web.mit.edu/xtalks/Sussman-xTalk-3-2-16.pdf

often though, representations that yield insight have some wow factor to them, especially if the world hasn't seen it before.

Between Nielsen and Bret Vector, I'm kinda jealous about what my future kids can learn with much more ease.

One things that I think would enable the kind of thinking necessary to come up with new interfaces for exposing concepts is to have a library of existing patterns.

I have seen various books and resources that come close (Edward Tufte books for example) but I think there could be much more work done.

There seem to be quite a few books on mechanical motion that are basically an encyclopedia of patterns that can be used to create new devices but I haven't really seen much type of this work done for computer interfaces.

If anyone has any good references please share.

You should have a look at books about the TRIZ system, such as this one

And Suddenly the Inventor Appeared: TRIZ, the Theory of Inventive Problem Solving https://www.amazon.com/dp/0964074028/ref=cm_sw_r_cp_api_

It is rather hard to describe quite simply, but I would try like this: TRIZ is about all of the patterns and heuristics found in the invention of mechanical engineering elements. It purports to provide an algorithm for innovation, but you don't have to accept that grand of a claim in order to find the content fascinating.

I'll check it out. Thanks.

Interesting insights. There's a lot to unpack here, but ultimately it boils down to working with, and creating, a "bicycle for the mind." The computer is one such example of such a technology, but so are the software programs it runs, as are art forms and language.

I understand what you mean, but I think it goes even further than a bicycle. At the risk of abusing the comparison, a bicycle will let you do what you already can do, move about, just a little bit faster. By contrast, what I think the article is talking about is doing something fundamentally new that you couldn't do before.

It's clear that overtime human beings can think about things that they didn't used to be able to think about. We can bootstrap our brains in this way, because it is all cumulative. Explaining the Internet to someone from the Middle Ages would be almost impossible, but it wouldn't be that difficult for someone from the 1950s who only understood telephones.

So this phenomenon of bootstrapping into new kinds of knowledge is clearly real and powerful. It's very useful and quite meta to think about how we could do that faster or better.

> Explaining the Internet to someone from the Middle Ages would be almost impossible

I think it's quite possible. "In the future, by pressing fingers on these letters printed on a specially built rectangular object in a pattern to invoke certain incantations, people will be able to speak directly to other people from around the world, as well as control machines in factories to produce stuff with little or no manual labor." Do make it clear you have no part in any of it, to avoid getting hanged or burned, but I'm sure they'll understand you.

I'm in the middle of the second "Off to be the Wizard" books and this is exactly how they deal with it. Really it just comes down to abstracting away everything under hood and relating to existing concepts and ideologies.

Even today, most people wouldn't get how a car functions, but they would understand that you push this and turn that and this moves you.

Imagining is the easy part if you ask me. What would you do, if you could program your own mind? My first app would probably be some more manual control over the body. Sure would be nice to have a 75 bpm in a stressful situation.

More consistency.

If I like something, let's like it more consistently. Let's really eat pizza for 3 days for every meal. So I can order the large one and pop slices into the microwave and enjoy the fuck out of them.

More motivation for progress.

If I decide I want to build a house let's not lose enjoyment after 1 hour of searching the net for floorplans, materials, warehouses, plots of land, and so on.

Also, better [internal] metrics. And naturally better interrupt management. Both internal and external. With better metrics we could see whether what we were doing (and possible what we can do) was (would be) effective/good/useful and how much. And with more efficiency we could reschedule more frequently (not just after we spent a day fiddling with something) so we'd be more effective overall.

And so on.

And probably a lot of these are pretty hackable, as in sprinkle cocaine on that pizza, alaways a bit more the longer it was in the fridge. And probably physiologically it's possible to make a pill that'd give you focus for one hour, we already have a dozen or so ADD/ADHD pills used as attention enhancer nootropics. And we have drugs that alter our view of our own life. Bam, instant feedback on your life. Are we going the right way? Yes, good, keep doing.

Is that too mechanical? Sure, yet sometimes that'd be pretty useful.

And in general, what if you could take a pill to feel in love with someone for 500 days? You get together things are okay, but emotionally something is bothering you a bit, both of you take a pill, and done. Instant good relationship.

The same could go for not getting dragged down by half-hearted relationships.

And that's basic limbic system efficiency, nowhere near mind-hacking. We'd benefit a lot from better pattern matching, better memory (faster, more accurate, etc.), better associations, better theory of mind. A bit of raw working memory, maybe a numeric submodule, something for proper alarms, alerts, better introspection, and so on.

All of these assume that the conscious "you" knows better about optimizing some meta-objective function than does the limbic "you." Except the latter has been proved out to be species-useful for tens of millions of years, and even more if cross-species evolutionary conservation is considered.

So the interesting question from my PoV is: beyond the immediate reward you hope to enjoy (e.g., loving the pizza as much on day 3 as on the first bite; easy relationships for 500 days) what would the consequences be of these overrides? I'm betting they're all catastrophic, but it's fun to think about the nature of the catastrophe.

> Except the latter has been proved out to be species-useful for tens of millions of years, and even more if cross-species evolutionary conservation is considered.

One of the things that makes humans dominate the planet so thoroughly is that we are programmable, so we can evolve culturally on a scale of decades, instead of genetically on a scale of centuries.

The side effect of this is that that parts of us that aren't programmable are not keeping up with rapidly changing environment eg being both cognitive misers and obese (https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4547044/).

It does lead to a neat heuristic though - if some easy change improves intelligent/memory/whatever, it should be expected to have side effects that would have been a poor tradeoff in our ancestral environment, otherwise it would have been selected for already - https://www.gwern.net/Drug%20heuristics

My life would be exponentially better if I could fall asleep at will. Especially if I could wake up no later than some given time interval.

It's the bottleneck to my productivity.

Reply on news.ycombinator.com