Tools for thought as cultural practices, not computational objects
maggieappleton.com
maggieappleton.com
I find the only truly excellent tool for thought is a long walk.
From analysing many recent articles on Elizabeth Holmes' Theranos caper and Bankman-Fried FTX snafu, I see a much deeper criticism of "silicon valley values" is afoot. Focus is on the tendency of digital system zealots to optimise naively, prematurely and (paradoxically) inefficiently for the human value problems they claim to solve.
Both placed the "ethical" dimensions of their projects at the fore, but created loss and misery because the lens through which they see (and interact with) the world, their tools, mono-cultural environment and communication bubbles led to optimising on investment, risk, presentation (appearance over reality) and so on. Both were enchanted by their own story, and "because technology good" gave them the excuse.
Nobody, even Ken Thompson or Steve Jobs (Unix, Apple) can create a transparent "tool for thought" that doesn't impose hidden values on the operator. Stare into the digital abyss and it stares back at you.
The best tool for thought is a human brain that can operate alone, treat technology as a tool that has its place, and use it as the need arises.
That is why we need to experiment with better tools for information manipulation and discovery (I did not say tool for thought).
I'm, like you, quite familiar with the explosion in digital music. Don't you think we need better designed "tools for music creation" to explore this infinite ocean of sound possibilities that has opened before us? I believe we do.
It is very similar for Knowledge. But given the more fundamental, "personal" and malleable nature of "thought", this will take a lot of guess work (and stupid ideas). In the long term though, I believe we need these better tools to adapt ourselves to this almost unhuman ocean of (mostly outer) information.
To me, knowledge and music do seem like infinite landscapes, and I'm excited by the idea of ways to extend the range and speed of exploring that infinity.
But to many people - I dare say a majority - that's not how they use technology. They want right answers. Correct opinions. Curated lists. Approved facts. Selected content channels... And our industry has shaped itself to "deliver" that, in cheap convenient forms.
Or rather, that's not what technology allows them. An additional layer has grown over the top; People who have opinions on what others should think, say, do and listen to, and those critics, censors, pundits, marketers, influencers and companies who fence-off the roads and place their signposts decide what may or may not be explored.
To call that exploration is a stretch. The tools they end up with, perhaps from choice, are not a pith hat, map and compass for exploration. Not a hitchhikers guide. Not even a taxi service. They are strapped into a roller-coaster and spun around a track by someone who has decided their "user experience".
Exploration is risky. Explorers freeze to death, or are eaten by lions. I don't see a lot of call for that, sadly. Given what we've built in entertainment I am a little pessimistic about extending that to knowledge engineering. The idea that computers might bring that same culture of constraint and approval to the very act of thought, a kind of thinking-on-rails, is depressing.
> That computers might bring that same culture of constraint and approval to the very act of thought, a kind of thinking-on-rails, is depressing.
Well, these are exactly the kind of problems the "tft community" need to engage in more.
For others, the naming (tool for thought) should not obscure the real implication of this domain : there will be 300 million college educated people by 2030, and a lot of them will be staring right at these kind of tools for hours every day.
If there's a way I can help them please tell me more (PM me) via Digital Vegan.
One thing I stress a great deal in my lectures is the difference between AI and IA.
Intelligence Amplification, to me, is the original purpose of computing, and while AI can aid IA, most current thinking goes the opposite way, toward usurping or minimising human intelligence. I often point to the existence of a UIX culture in which book titles like "Don't make me think" have become a tagline for an entire industry.
This HN article is a nice summary. If you want a deep hyperdive into their "ways of thinking and linking" (IA is actually not that far), you can follow some rabbit holes on are.ena like this one https://www.are.na/krish/tools-for-thought-lnujvshtshi
Tough luck. Scientists and those who want to go off-road still need tools for their thought. I'm happy to let the majority hold onto their 'curated exprience', but we need to have our needs met too.
Alternatively, researchers can try to use the same digital tools that entertainment consumers or corporate knowledge workers use. They might be less suited, but can potentially make up for it with extra polish.
When this happens, the researchers (who were the initial user base) basically lose control of the technology. This process can benefit the researchers or not, but is basically indifferent to their needs.
I think the omission of technology helps obscure a key part of the most recent rise in "tools for thought". "Tools for thought" are not so much about thinking as they are about substituting thinking with the exploration of technologies and tools (something I am certainly guilty of). "Tools for thought" give you the feeling that you are doing thinking much in the same way that TikTok videos might give you the feeling that you are learning something. That is why so many tools can proliferate in this space: they are not satisfying the desire to think, they are satisfying the desire to use (new) tools. Hence people keep switching and finding new and 'better' "tools for thought".
You would think that the people most popular in the "tools for thought" space would be those producing the most interesting or novel ... thoughts. Instead, they are people who explain how to use "tools for thought", introducing new taxonomical systems and technologies. To what end?
I'd honestly have thought that after 50 years we'd have overcome this, settled into a plateau of productive, reflective progress that uses digital tools instead of indulging in them.
But the churn of form and fashion, relentless progress of the tools and their markets needed to fuel growth has actually kept us mentally paralysed, enchanted, in an era where deep human thought is more necessary than ever.
Perhaps long-lived alien civilisations know this as the real answer to the Fermi Paradox. It isn't war and nuclear weapons that kills species. They just get PlayStations and amuse themselves to death.
IMO The problem is that, for all these years, "digital tools" progress was highjacked in service of corporation’s goals. the tft spirit is kind of different in that respect.
> has actually kept us mentally paralysed, enchanted, in an era where deep human thought is more necessary than ever.
Every time I take a deep dive into "quality" journalism/academia/books, I 'm reminded that all that necessary "deep human thought" we crave is already out there for the taking. We don't need more of it or even better. We need more people do open their minds to it.
But if billions of those people amuse themselves to death with media, and millions fashionably consume their digital tools as they do for everything else in their life, I think you should solely blame the influence of crass capitalism for that.
Tft kind of ideas, are a solution-in-waiting, not the cause of it.
> Instead, they are people who explain how to use "tools for thought", introducing new taxonomical systems and technologies. To what end?
At a higher level, I see this as natural evolution feeding off all that "desire for new tools" to (in the longer term) create better design for our "Information machines".
It is like this is the Cambrian explosion for the "information level" of life (Whatever that is). I say "let it be".
But (to return to the personal level) We should indeed explicitly police the boundaries between the 2 activities: experimenting (or even playing) with new stuff on one hand, and producing/thinking on a stable healthy (but still evolving) system on the other.
It is confounding the two that is the real problem, not the "tools of thought" in themselves. I'm not saying it is easy either. (I'm guilty as charged)
It seems we are at the cusp of something in the "Tools for Thought" domain - the question is what comes next because something big is likely about to happen. The article doesn't answer that AFAICS, but it does lay some solid groundwork IMO.
Mental arithmetic -> Paper spreadsheets -> computer spreadsheets -> spreadsheets with neural interface?
i have my drafts public too but i separate them for this reason.
PS: 1977 Tools for Thought was by C.H. Waddington, a UK biologist with an interest in complex systems. Club of Rome era about systems thinking and all.
I'm not sure about the practical utility of following in the footsteps of some of these suggested 'giants in the field'... we're supposed to learn Kenneth Iverson's APL?
https://dev.to/awwsmm/20-intriguing-unusual-and-goofy-progra...
I fear I am not worthy, nay, not even to wash the laboratory glassware in the facilities of such a genius. Donkeys must pull carts, while wizards fly through the sky above.
Generally my problem with digital paper right now is I need many of them and they need to interact with each other seamlessly.
The advantage of paper is you can spread it out over a large area and see twenty or so pieces of paper when necessary. And you can easily sort and shuffle them and process stackd. This sort of thing has been extremely difficult to reproduce with digital artifacts. But it's not insurmountable. If your entire desk surface were color e-ink, say and you could trivially stuff papers into a portable contraption. Anyway to me it seens more that the tech isn't there to reproduce "space". Something like today's reMarkable is ultimately tiny as hell and can't actually be used like paper.
I got a Boox a while ago and the fiest thing I realized (unexpectedly) was that I need at least two. One to read and annotate and another to jot working notes. And they would need to be in sync with each other.
The developments in this space are very exciting, and remind me of the early Wordpress plugin days. For example, I use a Logseq plugin called "random note" that pulls up a random note from my library. It's a great way to serendipitously re-surface ideas that you haven't thought about in a while.
For now, tools like Logseq and Obsidian can be used completely offline, and synced with anything that supports local file versioning, like Syncthing or Git. I wish I could pay $25 flat and have a lifetime license for the software exactly as it works now. As these are flat Markdown files sitting on my local drive, it doesn't seem like something that would need ongoing maintenance.
But it is a useful distinction, and Logseq/Obsidian seem indeed on course to become the WordPresses of PKB with their large plugin’s ecosystems.
However, you never know: It is a very fast moving (AI is barely in) and low entry (and exit, like @AthensResearch) domain. So, my personal wish is for better "interoperability" rather than any "lifetime licence" type of stability.
> the range of expressive thoughts possible in such a medium is an emergent property of the elementary objects and actions in that medium. If those are well chosen, the medium expands the possible range of human thought. [1]
So, my question is, when designing a software that helps users design things, what are the design principles that help expand the possible range of human thought?
The thing that computers do well is detailed simulation of possible configurations and visualization of aggregate results. The single best, most inspiring thing a computer can present is an entirely unthought of arrangement, i.e. a surprise.
Humans are routinely surprised by computers, but what software is principally designed to surprise its user?
Unfortunately this is insignificant compared to the explicitly anti-speculative reductionist mode most software developers operate it.
When designing a talk, is that a good thing to go for, or a bad thing?
Such ideas are hardly novel, but in subtle ways largely incompatible with a great deal of philosophy. Continue down that path and language stops having abstract meaning and logical supposition becomes irrelevant. Simply using the same words to describe something doesn’t mean both expressions have the same meaning.
In other words, create AI with neuronal networks. However, if we figure out which ideals make people get obsessed with large breasts, wouldn't we have a framework to build AI with logic, abstractions and ontologies?
Keep digging into this stuff and you don’t find a convergence of preferences to some ideal. Restaurants give people salt shakers because people don’t all want the same amount of it on their food. Which causes issues for companies selling prepackaged food etc.
At the same time the ideal of perfection doesn’t go away, we assume restaurants can be ranked on the qualify of their food while acknowledging people have different food preferences.
What makes you think she hasn't done this? The point of the article is that "tools for thought" as presented by Ken Iverson and others are a narrow focus in a much broader subject. Since her aim is to zoom out it makes sense to defer mention of Iverson. It's not indicative that she hasn't engaged with or understood his work.
Writing, drawing, and on-paper math are representational media. The process of writing, drawing, or calculating/proving is a form of manual processing because it requires organisation and transformation of representations.
Automated - computed - processing is a completely new category. It's a metaprocess which allows processes themselves to be represented and processed mechanically.
A lot of computing only does this is in a very basic way. The web is essentially a document distribution medium. Note organisers are essentially just a paper substitute that let you draw arrows and move things around, possibly with some hyperlinks.
They're both representations that support manual processing but don't improve on it. So they don't offer metaprocessing for most users. And with current tools metaprocessing is barely an option for most developers.
A genuine automated tool for thought would do the organising and summarising for you, perhaps teaching you if you needed that. But more - it would also improvise its own processes for achieving those ends.
Or at least make it easier for you to improvise your own at an abstract level that would be much higher than conventional code representations.
[1] Using an unaided human brain. Not hands, except incidentally.
I don't cite by postman for delivering the journal paper, now do I ? Another thing to put into my 'the West hates India and its 'heathens' so deeply that is now considered a blasphemy to question the zeitgeist' bin.