Collective Intelligence at the Knowledge Graph Conference
Notes from a fascinating Lounge Table discussion with Conor White Sullivan, founder of Roam Research
“I care about building a language for thinking in, good luck holding down a job, I’m glad Roam worked”
I’ve been to many conferences over the years, with lots of interesting conversations, however, the lounge table conversation at the recent Knowledge Graph Conference with Conor White Sullivan of Roam Research was a highlight.
Although I admit I am biased, as a co-founder of the conference and a subscriber of Roam, I have a personal interest in this emerging domain.
The conversation occurred at Knowledge Graph Conference Breakout Sessions organized by Larry Swanson and along with Larry, Matei Canavra, Paco Nathan, Ellie Young, Katariina Karii , Steve Herskovitz, and others, we enjoyed a 45 minute discussion which covered a wide range of topics from the current state of Roam, collective intelligence, blockchain, change propagation, product roadmaps, the importance of semantics and numerous interesting other references.
I have documented here a summary of the conversation, mostly quotes from Conor, and I’ve also added some references to the mix and a link to a recording of the call at the end.
Blockchain and Consensus Mechanisms
Conor is interested in the space, however, acknowledges it’s early days…
“How do we represent asynchronous convergent and divergent thinking with a group of people?”
“We need to first figure out intuitive user interfaces that allow us to asynchronously actually generate consensus among humans”
“We haven’t figured out how to have many to many asynchronous conversations representing convergent and divergent thinking…and create a record of what caused consensus to conform…allow a question to be reframed…and also allow for assumptions to change”
We are the DVD mailing stage of this…”
“Roam is still an R&D company, however, we believe we’ve created the primitives for creating this multiplayer thinking”
What is Roam focussed on now?
Matei Canavra, a power user of Roam, was interested to find out what stage Roam is at. Below are some quotes from Conor and this for me was an interesting learning from the conversation on Doug Engelbart and Bootstrapping…
I take inspiration from Doug Engelbart and we want a “Bootstrapping paradigm”… a team or organization whose mission is improving team collaboration or knowledge management technologies, innovation disciplines, or management practices is a candidate for bootstrapping if, by harnessing its own products and/or services, it is increasing its own collective IQ, becoming increasingly effective.
Doug Engelbert proposed a strategy for making the collective smarter. He initially called it “augmenting human intellect,” and later “He called it Bootstrapping Collective IQ”.
More about bootstrapping.
“We shape our tools thereafter our tools shape us”
Another interesting reference, though attribution of this quote is disputed.
“We are building for ourselves…how do we allow our team to do knowledge transfer and enable people to solve hard problems without having to meet in person…”
“Design is breaking things into atomic reusable pieces…Roam is world class at breaking a belief into its atomic pieces and we have good line of attack on other primitives…”
Larry asked about the Roam product roadmap, Conor is a skeptic on roadmaps and here we had another interesting reference to Hoftstadter’s Law.
“Planning is easier when copying something…for example, it’s not too difficult to estimate how long it takes to building a Twitter clone…it’s much harder when building on the frontier…”
Hofstadter’s law is a self-referential adage, coined by Douglas Hofstadter in his book Gödel, Escher, Bach: An Eternal Golden Braid (1979) to describe the widely experienced difficulty of accurately estimating the time it will take to complete tasks of substantial complexity
Hofstadter’s Law: It always takes longer than you expect, even when you take into account Hofstadter’s Law.
“We treat software development or technical research today like we are farmers…based on calendars…”
“A proper roadmap will allow for scenario modelling…..we are trying to build an interface that represents that variability…”
Paco Nathan, Chair of KGC Workshops and Tutorials, likes the free form nature of Roam. However, free form has its limitations, Conor noted that in the enterprise, senior level execs often use Roam and tools such as Notion which offer more structure are often used in the broader organization
“Free form can be an obstacle when working on a larger scale…we are working on templates…there is a bunch of things we can do around semantics”
Roam for Problem Solving
Ellie Young, Head of Community at The Knowledge Graph Conference had a brief discussion with Conor about the driving force behind Roam, using Roam for mapping climate information, and thinking beyond RDF
“We are trying to figure out this language on how to think collectively…we also wanted roam to be useful as a tool for building other tools…right now, we are mostly building for ourselves…”
“We don’t have a good system for thinking together”
“Mass communication is how society updates its beliefs and it is not fast enough”
COVID 19 and Governance, Mass Media
Conor referenced the Covid-19 crisis as an example of the issue with mass media today.
This brought to mind my favorite quote from Konstantin Todorov at the 2020 Knowledge Graph Conference. From Konstantin…
“Not only are we living through a pandemic, we are also living in an infodemic”
“Covid demonstrated how bad the situation is…we do not have governance that can update quickly on what is true…and we are going to have much worse crises in the future… mainstream media cannot deal with non linear change…”
“We need governance than can update quickly on what’s true…”
“We need to be able to do efficient allocation of expert attention”
“When COVID was emerging, I could not convince my educated friends to quarantine until the government actually said something…”
Validating Claims and Change Propagation
“It would be good to get a push notification if knowledge you have based a belief on has changed…
“We need a mechanism for investigating claims…”
Checkout this talk from Konstantin Todorov on Claims at the 2020 Knowledge Graph Conference.
“We are at risk of an existential crisis, Hyper media could allow us to do a much better communication than face to face…especially if there are many opinions”
“If we cannot solve this we will end up in a world that is way worse than it could be…”
“Figuring out how to do many to many communication….is the the problem we need to solve”
“There has been no justice in how bad the mass media got covid so wrong, and how we as a society get this wrong, they did know how to model exponential functions…”
“Can we model these hard decisions and communicate to people in a way they can trust it?
Ellie asked Conor about Climate and how much that drives Roam.
“We try not prioritize particular use cases or maps…”
“Your paper or notebook should never judge your beliefs…”
We want to create a medium where you can share your beliefs with people…”
“The first public endorsement of roam was from a philosophy grad student on Twitter who needed a software tool to map conspiracy theories because that’s how the student read their philosophy papers…”
“Roam is great for thinking about a complex network or complex system”
“You need to create a tool that allows people to be wrong in a way that is easy for them to update…”
“Map of your mind”
“The model for Roam as a tool is it should be a map of your mind….and when trying to build consensus across people, we should be collaborating across everybody’s graph…”
“When I interact with someone, I am reconstructing parts of their graph in my head…”
Thoughts on RDF
How will “anointing” happen in the future
Larry asked how Conor sees the process of anointing experts or spokespeople happening in the future.
Conor is a fan “Liquid democracy” the ability to defer your vote…
“What are the different kinds of trusts? There is transitive trust and non transitive trust.
Transitivity and Direction
All trusts are either one-way or two-way, and either transitive or non-transitive.
In a two-way trust, permissions extend mutually from both objects. One-way trusts are only given from one domain to another.
A transitive trust is a trust that is extended not only to a child object, but also to each object that the child trusts. (In contrast, a non-transitive trust extends only to one object.)
“We trust the people who were hired or anointed by mainstream media and this is a terrible metric…”
“I want a paper trail, a record of who I trusted and who they trust…”
The Underlay is a global, distributed graph of public knowledge. Initial hosts will include universities and individuals, such that no single group controls the content. This is an attempt to replicate the richness of private knowledge graphs in a public, decentralized manner.
“How people change their narrative, cognitive dissonance/surprise occurs when bottom up senses conflict with top down perspective and beliefs…this is what happens when people to lose faith in institutions…”
Conor recommends reading: “Surfacing uncertainty”
How is it that thoroughly physical material beings such as ourselves can think, dream, feel, create and understand ideas, theories and concepts? How does mere matter give rise to all these non-material mental states, including consciousness itself? An answer to this central question of our existence is emerging at the busy intersection of neuroscience, psychology, artificial intelligence, and robotics.
In this groundbreaking work, philosopher and cognitive scientist Andy Clark explores exciting new theories from these fields that reveal minds like ours to be prediction machines — devices that have evolved to anticipate the incoming streams of sensory stimulation before they arrive. These
predictions then initiate actions that structure our worlds and alter the very things we need to engage and predict. Clark takes us on a journey in discovering the circular causal flows and the self-structuring of the environment that define “the predictive brain.” What emerges is a bold, new,
cutting-edge vision that reveals the brain as our driving force in the daily surf through the waves of sensory stimulation.
Technology and Human Behavior
Conor spoke about thresholds require for something to become a behavior.
“There are certain things that will not happen until you make them easy enough to do…so it doesn’t interrupt the workflow… and we are working to make things easy enough so that the behaviors can happen…so we can build the behaviors, for example, change propagation through a graph is technically possible but too slow right now…”
“Making something fast enough so it doesn’t disrupt workflow, is the difference between that thing happening and it not…”
See this reference: Experimental results of tying a brick to a pencil
Collective Intelligence vs Google Page Rank
“What is order of magnitude between collective intelligence and Google’s page ranking algorithm?” asked Larry, Chair of the KGC Networking track.
“Google organizes the worlds information into to serve the perfect advertisement”
“Everyone will have a different page rank…reflecting their own social graphs and their knowledge graphs…at the time”
“In Google page rank algorithm the entities are not exposed to the end user, lots of stuff is behind the scenes, not configurable, not an empowering tool….lack of choice…”
“Intention or context is not factored into google”
“Most hard questions need to be broken down into other questions…”
The statements from Conor highlight the importance of contextual reasoning and semantics, a topic the Knowledge Graph Conference covers extensively.
“Should we have open standards?”
“Roam implemented JSON export once people started using Roam at least four times a week. Open standards are great…want to have the ability for people to define schemas…for apis….being able to define object types…you can fake it right now…”
Another interesting reference: Algorithms of thought
Algorithms of Thought then are procedures for thinking through a problem or situation to solve it.
Parting Thoughts from Conor…
“Checkout hypergraphs and hyperarcs…we have the schemas for it now…important for convergent thinking and how to track provenance”
However, we were not sure he would make it…
Conor made it clear the 6am PST start time was not ideal, perhaps next year, we will get Conor on our formal agenda at a more amenable time and perhaps in person?
We were glad Conor joined in what was our most extensive program ever, summarized here.
Checkout the full video of the conversation with Conor here. The KGC 2021 full video library is accessible from our website…if you are interested, please also join our slack Community at the Knowledge Graph Conference where we are discussing this and many other topics.
It will be exciting to see how Roam develops over the next year and how the overall space of Collective Intelligence and Tools for Thought evolves…