Hacker News new | past | comments | ask | show | jobs | submit login
How can we develop transformative tools for thought? (2019) [pdf] (numinous.productions)
127 points by stopachka on Aug 20, 2022 | hide | past | favorite | 53 comments



I'm not sure why this currently links to the PDF of the article [0], rather than the webpage, which includes an embedded animation of an interaction, and embedded YouTube videos: https://numinous.productions/ttft/

[0]: https://numinous.productions/ttft/print/TTFT.pdf


> Novel hardware devices (e.g., for VR, or the Wii remote, or for new musical instruments) can be used as the basis for new tools for thought. While hardware can be duplicated, it’s often much more expensive than duplicating software. And, in any case, the advantage for such companies is often in distribution, marketing, and relationships with vendors who make products for the platform.

SimulaVR was heavily influenced by the early visions for computing devices as "Tools for Thought".[1]

"Computing as an intelligence augmenter" seemed to be a really popular viewpoint through the 70s and 80s (even in corporate advertisements for computing devices). Then it somehow got co-opted by "computing as a medium for passive entertainment and video games", and then eventually as "a medium for social networking". I'm certainly not against these things in all contexts, but to me the exciting thing about VR/AR is that it creates a completely open field to make new advancements in the way human beings interact with machines to get stuff done immersively & with greater creativity.

Despite this, the entire VR ecosystem seems to be focused almost exclusively on gaming/entertainment + social networking (facebook "Metaverse"). I personally find it really exciting to think about how the extra dimension could help us create more interactive tools in line with Alan Kay (& co's) early visions. Before that's possible, we first need to get the basic hardware good enough to replace PCs & Laptops (for everyday 2-dimensional applications). After that, we can these devices as platforms to build better thinking tools on.

[1] https://simulavr.com


> "Computing as an intelligence augmenter" seemed to be a really popular viewpoint through the 70s and 80s (even in corporate advertisements for computing devices). Then it somehow got co-opted by "computing as a medium for passive entertainment and video games", and then eventually as "a medium for social networking".

This is probably a demonstration that intelligence only gets you so far and isn't really that useful in life. Nerds don't like this because intelligence is what they value in themselves; that's why so many of them think that if "artificial general intelligence" was invented that is "superintelligent" it would somehow also have superpowers and be able to invent viruses, convince anyone of anything, etc.

But how's it going to do that? Just imagine them really hard? What part about being able to think really fast says any of those thoughts are accurate?

Basically, computing technology is good at creating virtual worlds of logic, but it's much harder to affect the real world with it.

See also https://en.wikipedia.org/wiki/Moravec%27s_paradox.


For me, emacs (with org-mode) and array languages (like J) do augment my intelligence, because the productivity boost is so profound that it affords the opportunity to try more things and be more creative.


Passenger jet engines, nitrogen fixation, silicon fabs -- these don't just magically appear. It takes careful reasoning and years of exploration in the vast ocean of ideas. Most places in this ocean are not useful, so it's helpful to sail faster.


I think it come down to intelligence augmenter is hard to separate from skill augmenter. Reading a chess book can make a beginner much better at the game, where following the instructions of a chess engine doesn't have any enduring impact on skill. Computer as orical such as with chess engines, or weather models isn’t actually improving intelligence any more than a truck is improving strength.

Word processor, email etc, can speed up how quickly an author gets stuff done, but they don't turn the average person into a good writer. You basically need to become an engineer before CAD tools let you design a state of the art car engine.


Why don't we see more powerful tools for thoughts?

>> Put another way, many tools for thought are public goods. They often cost a lot to develop initially, but it’s easy for others to duplicate and improve on them, free riding on the initial investment. While such duplication and improvement is good for our society as a whole, it’s bad for the companies that make that initial investment. And so such tools for thought suffer the fate of many public goods: our society collectively underinvests in them, relative to the benefits they provide.


Thoughts themselves used to be public goods, but thanks to aggressive copyright laws and customs, even the most basic implementation of Bush's Memex from 1945 is illegal to use.

We're no longer free to copy and improve the thoughts of others by law, custom or practice, except for some limited areas like HN.

In this environment, the tools can't be built that we most desperately need


> Thoughts themselves used to be public goods, but thanks to aggressive copyright laws and customs, even the most basic implementation of Bush's Memex from 1945 is illegal to use.

I don't believe this is true at all.


How so?


I think the claim that there are any restrictions on memex is somewhat unusual and is the thing that requires explaination!

Bush's Memex was explained in an essay[1]. There's no intellectual property restrictions around any of the ideas there, and some of the explanations were so rooted in the technology of the day that the idea of restrictions on them doesn't make any sense:

> Most of the memex contents are purchased on microfilm ready for insertion. Books of all sorts, pictures, current periodicals, newspapers, are thus obtained and dropped into place. Business correspondence takes the same path. And there is provision for direct entry. On the top of the memex is a transparent platen. On this are placed longhand notes, photographs, memoranda, all sorts of things. When one is in place, the depression of a lever causes it to be photographed onto the next blank space in a section of the memex film, dry photography being employed.

Even putting aside the microfilm technology, most (all?) of the concepts described in the essay seem freely available.

[1] https://www.theatlantic.com/magazine/archive/1945/07/as-we-m...


I'm not saying the technology would be prohibited... I'm saying that making copies of copyrighted works would be prohibited... which is the main way that a memex was meant to work.

Record companies freaked out when people started using Napster, a Memex for music. I'm sure the same would happen if I were to download an article from the New York Times, for example... and repost it with a copy of my notes added.

The copying of material is the thing that is just a no-go these days.


> tools for thoughts

Given his dedication to the concept through decades, arguably iPad — together with apps for it like LiquidText[1] — embodies Job’s “bicycle for the mind”.

1. https://www.liquidtext.net/

Speaking of, not just for iPad, and with emphasis on “tool”, Obsidian[2] is emerging as a remarkable tool for thoughts.

2. https://obsidian.md/

One key to popularity seems to be the plugin ecosystem which ratchets up a swiss army knife aspect of tool, suggesting its macro power may be any number of “scratches exactly that itch” micro tools for thought.

While other tools are more extensible or easier to use, this one pulled the plugin ecosystem into first class citizen status very early, with browser and manager built in. This let the “public” not just ride, but contribute early and often, cementing investment.


About 8 to 9 years ago I asked Michael Nielsen (one of the authors of this PDF) by email what he thinks about interactive theorem proving. He was kind enough to answer, and replied something along the lines of that he doesn't believe mechanised logic will change how mathematicians work (if I remember correctly).

I think interactive theorem proving (ITP) is the ultimate tool of thought, though.

Although as a field it exists for quite some time, I think it is only now starting to show its full potential, as everybody starts to realise what powerful AI can do for ITP automation. Things like Mathematica and AutoCad should really be just special apps running on top of an ITP operating system.


I agree. Learning how to use interactive theorem proving tools like Coq, LEAN etc. has blown my mind and given me a much better way to think about software development. The same way that learning basic maths teaches you a new way to think clearly about the messy real world.


How did you do approach learning it to make it “stick” to such an extent that it’s changed how you approach problems?


I completed the first volume of Software Foundations:

https://softwarefoundations.cis.upenn.edu/

and then started proving tiny toy examples correct from scratch. I also read a lot of papers on type theory and the history of proof assistants. It’s really interesting fun stuff.


ITP is a key part and a good start but I wouldn't call it the ultimate tool.

Thought is more than logic. Logic operates on a model, after things have been named and conceptualized. As the saying goes, there are two hard things in computer science, and ITP doesn't cover all of them.


I would say it covers naming things and cache invalidation. And it is really really good at avoiding off-by-one errors.


how does it cover naming things?


That depends on the ITP system. But in general, in an ITP system you can separate the name of something from how it is displayed. In the ITP system I am currently building, Practal, the displayed syntax does not need to be unique. So choosing a unique name becomes simpler, because you don't have to worry so much about how it looks, because that can be entirely different and doesn't need to be unique. Apart from that, every accessible element of your mathematical universe already has a name: it is the term you use to describe it.


I've never thought that "naming things" meant keeping track of the name or anything, but that's hard too. I've always thought it meant the descriptive aspect of naming.


Sure you can rename things easily, but NAMING something CORRECTLY is hard, in an ontological sense where you may not know what the thing's name is until you truly know what it is.


Mathematics is the most creative work there is, so if you can automate the tedious work, it helps, but no miracles. Theorem proving requires unnecessarily formal reasoning chain that is usually more work than worth.

The way most mathematicians work is: First they figure out what the lemma or theorem they want is. Then they try to prove it.

I think theorem proofing will be beneficial for programmers and inside a compiler. Messy program with assertions -> proof -> verified program.


I am pretty sure that in 20 years every mathematician will happily use an ITP system. That is because formal reasoning is not unnecessary, but just too burdensome to be done on paper. Ideally a future ITP system will give you the formalisation almost for free, and lets you concentrate on your creative insights. It will be empowering you, instead of restricting you. It will be a tool that will let you explore your ideas more freely and creatively than it was possible before due to your limitations as a human. That is also why it is so important to get the basic foundations of such an ITP system right, as the limitations of the ITP system will become your limitations.


hey no offense but right now you're giving off big crank energy. no one is going to take your "abstraction logic" seriously because it's disconnected from the literature and completely unverified.

your characterization of things is strange, there is no Rasiowa’s approach that missed a generalization that a modern presentation of Frege's work shows. it's hard to understand from a glance what you are calling "abstraction logic" is but it's near certain it's already been classified as something in the literature. if not then write up a basic result properly so someone in the field can tell you have something at a glance and publish your result with them, you can be first author so it's not like you're giving up any glory and they will ensure you are peer reviewed through the right channels.

but you most likely have no result to speak off. which is fine. just build something. we need more ITPs in the wild, and currently have more of a UI and OS problem than a lack of proper theory.


Shoutout to Freeplane, an open source mind mapping software available on Linux, Mac, and Windows.

It has a bit of a learning curve to use smoothly, but I believe it’s best in class once configured to your liking. I have had to avoid yakshaving, but I have the same “issue” (hobby) with software like Vim and Emacs.

I’m using Freeplane to outline a narrative non-fiction book, and find it invaluable for this purpose. Would love to hear other opinions on Freeplane or other mind-mapping or concept-mapping software.


The mind mappers I've tried all have a hierarchical tree structure. I want a network structure, like the internet itself. I'm building a local website to get what I want, something like the original HyperCard by Apple. I haven't tried Freeplane so thanks for the suggestion.


Infinitymaps.io has network structure and very, very deep zooming abilities


Nice, but I doubt I'd want to store my private thoughts on a server in Germany.



This looks good and is closer to what I want. The low price of $18 is also entirely reasonable.


Thanks for asking this question, I actually have been looking for the same thing. I’m going to check out the recs in these sibling comments.


I'm thinking of a hyperlinked local site like HyperCard but with the ability to view interconnect pages merged onto one large page. In other words, adding more views than a traditional hyperlinked system offers.

Heptabase may have what I am looking for:

https://heptabase.com/

There's a lot of interesting ideas floating around out there:

Xanadu Hypertext Documents

https://sentido-labs.com/en/library/201904240732/Xanadu%20Hy...


"that computers have not yet been nearly as transformative as far older tools for thought, such as language and writing"

Language and writing are the platform upon which our symbolic computing systems are based, so I'm not surprised.


There's more of Andy's (and Michael's) work at https://andymatuschak.org/ . Looks very interesting!


I remember quantum country from a couple of years ago and like most things online, I spent a bit of time on it, then it sat as an open tab for probably months before I went through and purged it because it wasn't important enough to revisit. It was very compelling, except it wasn't a tool. There was a view of a controversy in what education was (I associate it with Cardinal Newman), where the question was whether it was more effective to inculcate and find more efficient ways to encode and transfer information into minds, or whether it was best to "draw out," capabilities in students. It was framed as drawing out vs. putting in.

The idea of drawing out implies that there are more efficient ways to leverage innate human capabilities by providing students with tools to solve problems they already have. The complaint that "math isn't practical," is a failure of imagination on the part of teachers to present compelling problems, and teaching the abstractions without reasoning your way into them rewards only a very limited type of mind. "Math is social," is another one I hear a lot, which means you need to be engaged in an active conversation about it to do it, which is peculiar because I can hack my way through playing Bach and Chopin pieces on my own with just enough sight reading to get them under my fingers, and routinely memorize archaic and complex 1k+ word monologues for performance, but for some reason one still needs to be an initiate to do math because it's still modelled on the "putting in" mode of education.

Spaced repetition is amazing and necessary, and memory work is a very old practice, but they're limited by the "putting in" ideas behind it. When I'm doing those other activities, there is an urgency and a reward in them, where we haven't quite articulated a path for drawing it out in other fields.

I have a pet theory that the way to leverage peoples minds is to set the correct initial conditions that yield internal consistency from their iteration. Look at how encyclopedic people become when they believe in something. They have an explanation for everything that reinforces the initial idea, and some of it is actually not always terrible. Purely synthetic ideologies are just over a century old, but they are unreasonably effective as knowledge graphs. There is a formulaic recipe for ideologies I've been noodling, and maybe a good test run of it would be whether we could use it to create an Ancient Aliens tv show but for math, where we can travel around interviewing slightly bemused mathematicians about how really complex ideas are just homologies of a few elegant ones. I'm not saying it's category theory, but it's category theory. The memes write themselves.


It seems a bit over the top to call a memorization hack based on well understood (and commonly employed) memory methods a “tool for thought”.


Drugs and/or meditation are pretty potent "thought tools". (If only as a side effect). Popular with scientists, artists and engineers the world over.

Proven useful but maybe still under the shadow of taboo.

Much serious development could be done there

I imagine a class of specialized "thought workers", dosing up like Mentats. Or meditating in government-issue sensory-deprivation tanks.


Never seems to define what a "tool for thought" is supposed to be/do. Seems like a pretty big problem before moving on building them, let alone building "transformative" ones.


Honestly it seems to me like that quality of "thought" has decreased since technology arrived on the scene. People don't learn Latin anymore, and they can't write letters like the soldiers in WW1, let alone the Civil War. But we do watch a lot more porn and we do drink a lot more sugary drinks. So my advice if you want to improve your thought is to get rid of your computer and iPhone and eat only red meat.


Do the acts of learning Latin and writing letters truly represent the peak quality of thought?

I'd think that technological ability and thought quality would improve in unison. Better thought leads to technological leaps, which potentially lead to better thought.

Being a luddite hardly sounds like a solution. Thought not unlike meditation do have their place, but I believe they help enhance "normal" thought, rather than completely substitute it.


Nope -- try reading a collection of essays written 75 - 150 years ago and it will soon become clear how stupid people are today. And why do you think technological leaps lead to better thought? The most profound thoughts that have ever crossed the consciousness of Man are contained in the Bible, which was written by technologically inferior people -- men who walked with God.


I do think there's a degree of selection bias, here. Only quality writing from long ago would survive; the C19 equivalent of tweets (or HN comments) will have long since vanished, so not be around for us to evaluate.


> try reading a collection of essays written 75 - 150 years ago and it will soon become clear how stupid people are today.

I read a lot of essays, novels and newspapers from that era. Plenty of dumb, and completely wrong essays were written then.

I invite you to read - for example - "The Case for Spirit Photography" (1920)[1] by Arthur Conan Doyle (yes, the author of Sherlock Holmes) in which he argues photos of fairies are real.

It's well written, and completely stupid.

> The most profound thoughts that have ever crossed the consciousness of Man are contained in the Bible, which was written by technologically inferior people -- men who walked with God.

Au contraire, the most profound thoughts were clearly those of Aleister Crowley[2]. He wrote right in the period you suggest the best writing occurred (75-150 years ago) and because he was in direct communication with the devil via his sex magick they are obviously the most profound. Or did I mean profane?

Take your religious nonsense somewhere else.

[1] https://www.gutenberg.org/files/59809/59809-h/59809-h.htm

[2] https://en.wikipedia.org/wiki/Aleister_Crowley

[3] https://en.wikipedia.org/wiki/Sex_magic


…that’s one rollercoaster of a comment.

But why haven’t you gotten rid of your iPhone and computer? Or did you submit this by post?


internet café


But we do watch a lot more porn

Surely porn is preferable to the prostitutes of the past?


STDs are better than computer viruses you get when downloading porn torrents. STDs are part of nature.

/s


In what way is slapping the salami to provocative pixels with zero feedback to the end user, better than real live human being? How did we get to this point where this is even a question?


Better for the women, not the men lol


It’s likely too early to tell what the instantly and freely available and limitless variety of cartoonishly depraved imagery is doing to human brains. But I have doubts about its benefits.


"Cartoonishly depraved"? The vast majority of porn is just pretty normal sex with more attractive than usual participants. Porn has demonstrable benefits for reducing sexual violence, in fact, that's the strongest non-free speech argument for legalizing distribution of illegal porn (obv. not an argument for the legalization of production!).




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

Search: