Augmenting Human Intellect (1962)
dougengelbart.org
dougengelbart.org
Shameless self plug: I'm part of a couple of crazy scientists and enthusiasts following work from Doug Engelbart, Wiener, Stelarc etc. Our yearly conference is coming up next week: https://augmented-humans.org/
Here's a preview of this years work ... happy to see some crazy works there as well. https://www.youtube.com/watch?v=YYmu2bFN6kE
Last year we had a keynote from Stelarc (linked from my blog): https://kaikunze.de/post/2020-02-18-augmented-humans/
edited: anecdote fixed + spelling
Me, about an hour ago[1], in the "How can you not be romantic about software?" discussion. Amusing to see Doug echoed back on HN so quickly. Small world!
It's in reply to one of the most critical pieces of the thread, to me. The overarching article is somewhat fetishistic, just in to the wizardry of it all, the power, the skill. I was replying specifically to someone promoting the use of personal knowledge systems, to capture everything you are learning, to build a log & indexes. This, really, is the heart of what I find so romantic: this augmentative, compounding nature of software, of informatics. To gain heightened view points, and to make ourselves steersmen[2].
That thread included a shout out to Karli Coss, who has done more than anyone I've seen to draw together all the streams, to gather & create views of his own existence as it forms across the digital[2]. He's brought a wide range of software- from ebook readers to music players to web browsers- into his own domain, to build his extended brain. Karli has augmented his intellect far further than anyone else.
There's a lot of compelling things about software, a lot of ways to have fun. But of the ways to be romantic about programming, Augmenting Human Intellect is the most profound & resounding romantic notions that I have ever encountered. I see so many people engaged in software, but this seems like the core of where almost all real true meaning lives, and we should reshape our focus, abandon so much software, to chase an integrative, augmentative vision of computing in pursuit of this romance. We should be inspired, as Doug was.
"A" work is where the concept flourishes: he spoke of networked imrovement communities (NICs) where the "A" team is an improvement community, and a NIC is those "A" teams in a network. Think IEEE as a NIC.
But, he didn't stop there. What is an improvement community? My interpretation is that he uses the term "Dynamic Knowledge Repository" (DKR) to explain the A team; "repository" does not refer to a bank account; rather, a DKR is a combination of human systems, tool systems, and the knowledge brought to the table by the humans. Draw a circle around that. That's how I interpret a DKR.
Key point to take home: all of the elements of a DKR are in a state of continuous coevolution. As humans improve, the tools they use must rise to the occasion.
What held back that concept, historically, has been the tool systems. When he did MOAD, I was working at IBM on OS 360 mainframes. That's what we had in the 60's. Today, my cell phone has more compute power and memory than one of the 360s I serviced back then.
It's for that reason that I am optimistic that we are on a path to realizing Doug's grand vision; our tool systems are now way ahead of we humans.
Which would sound like someone whining about how the bigger kids are running too fast, but when I finally get fed up and fix the code, half the time I’m the one who spots the underlying problem that makes a big difference. Accidental complexity kills invention. Every time we make a better tool, we reduce the accidental complexity.
If we ever figure out how to draw or write directly from the motor cortex, no hands involved, that will be a new Enlightenment.
If we did figure out drawing or writing without hands, I think that would be a noticeable, but still small improvement.
Much like with coding, the insight/information /s rate is very rarely keyboard limited.
A good programmer switching to Vim may help them a bit, and it can be a strong enough effect that lots of people will want to switch to better tools like Vim
But I doubt there's room enough in there to extract a whole new enlightenment from :)
Once I know what code needs to be written, I still need hours to actually type the damn thing out (not to mention the constant typos, deletions, reformatting, and so on). One of my loftier startup ideas is a Neuralink-PC interface that projects the code in your head to code on a screen.
I really hope I see it in my lifetime.
I hope so too, please don't let me stop you! :)
But I know that other people do not work this way, so 'draw myself a picture' is still a gating condition on writing the code.
I find this an ironic bit of User Interface failure.
Sounds like math.
BTW, I wonder how many of the people who upvoted this have actually read it.
I'm appalled by intellectual laziness and codependency these days. We have a mostly illiterate, ignorant, distracted, atomized society that doesn't read, doesn't dig into primary or even secondary sources, doesn't have critical thinking skills, is easily-manipulated by advertising, mainstream news, and social media, doesn't communicate in-person that much, and more and more people feel entitled to be spoon-fed answers and information like helpless chicks tweeting for a worm (the definition of codependency). Only by more people reading, learning, researching, thinking, and doing for themselves more often can these trends be reversed.
PSA: Turn off the screens and read a respectable nonfiction book more often, please.
"Give a man a fish and you feed him for a day; teach a man to fish and you feed him for a lifetime." - English proverb
Writers should aim for concision where possible, but not skimp on evidence, rationale, or interesting anecdotes or analysis. Einstein said something about "irreducible basic elements" but never said "Everything should be made as simple as possible, but not simpler."
Time and attention are valuable, but the essence shouldn't be sacrificed on the altar of expediency.
I actually spend a lot of time reading. But if I have, say 4 hours per day allocated to reading, I want to choose material that gives me the highest possible value per minute of reading.
You're basically saying that the optimal learning and information consumption strategy is to always either read (or skim) the whole thing or ignore it. I disagree, the optimal strategy includes reading TL;DRs if available because they give you good value per unit of time or energy and they give you information whether it's worth to read the whole thing.
If I was to add my own spin to the problem outlined, I think it would be one of information and organisation. If we can significantly improve the bandwidth of information flow between individuals and groups, and then more efficiently organised larger groups towards a singular task we could achieve many things.
Some groups do achieve great things but it's not clear to me that we've codified how this can be achieved and/or identified what is required for such success.
It is good to have well-constructed readers to summarize and discuss implications of subject matter, but not as primary material, only to ensure mastery of it.
I mean, why not spend <1 minute reading a TL;DR when it can sometimes give you most of the value and if not, provide you with information whether it's worth reading the whole thing. It seems like a good deal, why avoid it?
I'd rather spend my time reading stuff thay has a higher expected value (e.g. a book about macroeconomics with great reviews if I'm interested in that topic) rather than random long texts on HN that can turn out to be complete garbage.
For example, plenty of people watch the demo I mentioned above and spend all their time marveling at the invention of the mouse, or some other detail of the hardware, rather than taking in what was really being explored by the users of the hardware in terms of human/computer and human/human interaction.
If you insist upon having a tl;dr, it has its own: https://www.dougengelbart.org/content/view/138/#5a2
(I didn't remember it being that long from both times I've read it, but I guess I found the thesis interesting enough that the time passed quickly?)
(Something else that I haven't seen pointed in this thread is that people with whose english is an L2 may legitimately be slower than in their L1.)
"I'm sorry, Dave, but I'm afraid you need to purchase the Universal Homework Cheating app from the App Store on your mobile device or move to Finland where there is no homework and better educational outcomes."
If AI/ML can be made general enough and enough computing horsepower is available, then a UHC is possible. Soon afterwards though, self-reprogramming will be possible, rendering most programmers and software engineers redundant. You then have to wonder what jobs cannot be automated until emotion AI and truly general AI arrive, and displace people into perhaps an economically-driven technological singularity / organic transcendence.