🦋 Glasswing #1 - Communication is not the same as meaningful communication.

A great advantage of the internet has been the ability for people to communicate globally. A downside of that however, has been the decontextualization of information and the way it impacts how we communicate to each other. The strongest form of communication is exhibited in the physical world. In the physical world, you are often able to achieve privacy and ensure that everyone is on the same page with what it is you are saying (preserving context). Digital communication has broken both these properties and much more. This has impacted our ability to coordinate, communicate, and have a strong sense of identity. With the advent of language models and generative AI, we are only going to further distrust the information we see digitally without other security measures in place (ie., cryptographic signatures).

---

It’s Time for Deep Social Technology

In the past, political institutions were used as a solution to address the complex and often initially invisible interdependencies between the interactions of different human beings, as revealed by social scientists.

In the past 20 years however, technology companies have attempted to play this role by “bringing the world closer together” (Meta), “giving everyone a voice” (YouTube) “empowering expression” (Snapchat),  or “communicating without barriers” (Twitter) .

One of the key lessons that social technological platforms have taught social scientists is that it is misguided to assume that once a social problem is understood, it can be solved by technology alone. Rather, technology should be designed with more input from the social sciences in order to better adapt to the constantly evolving interdependencies between people and technology. This approach is essential for ensuring that technological solutions are effective and sustainable in addressing complex social issues.

Investing in technology has shown promising success in addressing some of the most pressing global issues, such as climate change, healthcare, security, and governance. In recent years, there has been a growing focus on investing specifically in "deep technology" – technologies that often have a high knowledge and cost barrier to entry, are not easily venture-backed, and may not have an immediate customer base. Examples of deep technology include blockchain, artificial intelligence, advanced materials, biotechnology, quantum computing, and robotics.

The investments made in deep technology are starting to bear fruit in these domains, with significant advances being made in fields such as blockchain and artificial intelligence (AlphaFold, chatGPT, GPT-3, Ethereum, scaling zero-knowledge proofs, and much more). The rapid progress in deep technology has also raised concerns and sparked discussions about the ethics, privacy, security, responsibility, and safety of these technologies. As a result, many researchers, policymakers, and industry stakeholders are working to develop appropriate guardrails and guidelines for the types of experiments and applications being developed in this field. This is a critical area of study, as it is essential to ensure that the benefits of deep technology are realized in a responsible and ethical manner, without compromising the safety and privacy of individuals or society as a whole.

Social technology is fundamentally concerned with designing technology based on social problems, whereas other branches of ethics, security, responsibility, and safety focus on how to manage already existing technologies. As these technologies start to interact more with humans, the social consequences of these interactions become much more complex and hard to predict. This highlights the need for an interdisciplinary approach that incorporates insights from the social sciences, humanities, and other fields in order to understand and address the ethical, social, and policy implications of deep technology.

Similar to “deep tech”, deep social technology aims at attempting to solve very complex social problems through technology that will likely also initially have a high barrier to entry in both cost and knowledge, work over long time horizons and not be venture backable. Deep social technology is grounded at the intersection of cryptography, machine learning, and the social sciences. However, as the space matures and the output of deep social technology starts to demonstrate the financial returns it can generate, the willingness to invest up-front capital to make these barriers to entry harder will be reduced. 

Current Focuses of Deep Social Technology

The current focus of deep social technology is trying to address digital communication.

A great advantage of the internet has been the ability for people to communicate globally. A downside of that however, has been the decontextualization of information and the way it impacts how we communicate to each other. The strongest form of communication is exhibited in the physical world. In the physical world, you are often able to achieve privacy and ensure that everyone is on the same page with what it is you are saying (preserving context). Digital communication has broken both these properties and much more. With the advent of language models and generative AI, we are only going to further distrust the information we see digitally without other security measures in place (ie., cryptographic signatures).

The impacts of digital social platforms have impacted our ability to coordinate, preserve context, and have a strong sense of identity.

Digital Cooperation Failures

To make a causal argument to explain why we have cooperation failures today would be ambitious, but one can critique the reasons as to why digital technology does not facilitate cooperation as well as it could.

What is cooperation?

Mutualistic cooperation, or positive sum cooperation, is when an individual confers a benefit to others while also conferring a benefit on themselves. This form of cooperation is not as much a problem of emotions related to reciprocation as much as it is an epistemological problem. 

Knowledge management platforms like Wikipedia, public discourse forums like Twitter and Reddit, and direct messaging tools like email and text, have been powerful at proving the impact that these social technologies can have on communication. However, not only do these platforms face challenges with the quality of information shared, they also are not sufficient in creating common knowledge amongst the individuals who are communicating. 

A proposition ɸ is said to be common knowledge if “everyone knows that ɸ is true, everyone knows that everyone knows that ɸ is true, and everyone knows that everyone knows that everyone knows that ɸ is true, ad infinitum”. The power of a system with common knowledge, as seen in supply chain and contract negotiation settings has demonstrated to be quite powerful but has still yet to propagate to mainstream communication tools for information sharing.

We need to better understand the parameters in a system degrade our quality of common knowledge, experiment in different settings for communication: everyday organization, sexual assault reporting, convening large groups of people towards a similar mission. Distributed ledger technologies (DLTs) and blockchains happen to be very good at achieving common knowledge which I will comment on in a future post.

Context Collapse

In her work, "It's Complicated: The Social Lives of Networked Teens," dana boyd discusses the phenomenon of "context collapse," which is when the different contexts in which a person operates (e.g. home, school, work, etc.) collide into one another. She argues that this is a result of the way social media is designed, and that it has negative consequences for users, who are often left feeling exposed, anxious, and confused. 

As generative models only become more powerful for text, image, and video generation, the amount of information that is going to become accessible online is going to increase dramatically. The road in which we are headed currently is an information tsunami that will be flowing in many different directions, with the inability for users to appropriately parse the information needed for their contexts. 

One axis of deep social technologies is to preserve context in communication. This means being able to mitigate the public revelation of private communication and being able to have persuasive guarantees on the authenticity of the information being displayed to a user. The tools and techniques to be leveraged here lie at the intersection of machine learning and applied cryptography with zero-knowledge proofs and designated verifier proofs.

Social Identities

Despite all of the work being done to define new digital identity primitives, the challenges with recovery and fake accounts (“bots”) still remain. The impacts of this result in the inability to do digital governance, mis/disinformation and distrust online.

As a crazy aside! People recently started using AIs for their digital dating life to optimize responses (one guy managed to be talking to > 5000 girls at one time with it). Imagine a world where you are in the “metaverse” falling in love to an AI because it is going to optimize on all of your recent data (Twitter posts, Strava data, GoodReads, Instagram posts, Facebook likes, LinkedIn shares, Swarm check-ins, Be-Real network, YouTube recently watched, Spotify recently listened to).

We need to fix this and find a way to get social identity right and start trusting people online. In fact, how do you know this entire article wasn’t written by an AI, or even parts of it aren’t? You don’t …. That feeling will only increase as our relationship with these generative tools increases in use.

Expanding Deep Social Technology

In the future posts of this newsletter, we will get a bit more tactical to begin to answer some the questions in this post. How do DLTs achieve common knowledge? How do we build a digital communication protocol that mimics the real world? Where does AI and cryptography fit into this picture and why?

Let’s build a digital world that closely mimics the physical world, while still benefiting from the ability to communicate globally.

My goal with my collaborators is to create an ecosystem of projects that work towards deep social technological problems. Similar to how there is an ecosystem of venture and culture around general purpose deep tech, we hope to develop similar strategies for deep social technologies. We believe that these projects can be experimented with under many vehicles: academic labs, startups, or large corporations. 

As OpenAI, DeepMind, and bio companies have done for deep technology, we believe we can manifest the same for deep social technologies, and hope to better engage the public along the way in doing so to make these tools as powerful for society as possible.

Thanks,

Shrey Jain