The Genesis and Evolution of ComplexChaos

Introduction

In an era where complex problems require innovative solutions, the traditional siloed approach to problem-solving is no longer sufficient. ComplexChaos was born out of a need to break down these silos and foster interdisciplinary collaboration that leverages diverse expertise to address the world's most pressing challenges.

This article explores the journey of ComplexChaos from its inception, the ideas that fueled its development, and the technological innovations that continue to shape its future.

What follows is a comprehensive but not exhaustive description of our journey for the last 12 months.

1. The Birth of an Idea

The Initial Spark (2012)

In 2012, while leading a marketing company that served global brands, the limitations of the industry became increasingly apparent. The focus was often on surface-level insights and repetitive creative strategies, with a primary emphasis on delivering campaigns for well-known global brands.

These campaigns, while successful on the surface, in my humble opinion lacked depth and the potential for true innovation. Strategic Planners were employed to delve into customer insights, but their findings were often funneled into narrow, pre-existing frameworks, leaving little room for groundbreaking ideas.

The frustration with this repetitive and shallow approach led to the birth of a new idea: what if the creative process could be reimagined by forming multidisciplinary teams? These teams would not be limited to marketing professionals but would include experts from a variety of fields, such as social sciences, game design, architecture, biology, and computer science, among others.

The hypothesis was that such a diverse group could bring fresh perspectives and innovative solutions that a traditional team might overlook. This was not just a simple rearrangement of roles; it was an experiment to see if the convergence of different domains could lead to novel ideas and approaches.

The concept was bold and ahead of its time. However, despite the excitement around it, the idea was shelved due to the complexities of implementation and the inertia of the existing business model.

The marketing industry was not yet ready for such a radical shift, and the resources required to bring such a multidisciplinary team together were significant. However, the seed was planted, and the idea lingered, waiting for the right moment to resurface.

Revisiting the Idea (2023)

Fast forward to 2023, a decade after the initial idea was conceived, the landscape of technology and innovation had dramatically shifted. The rise of artificial intelligence, particularly the success of ChatGPT, reignited the spark of multidisciplinary collaboration.

The world had changed, and so had the tools available to bring this vision to life. The time was ripe to revisit the idea, not just as a marketing experiment but as a transformative approach to problem-solving across industries.

Initially, the concept of ComplexChaos was envisioned as an art project. The idea was to use AI to facilitate creative collaboration, allowing artists, scientists, and technologists to work together in ways that were previously unimaginable. However, as the exploration into AI deepened, it became clear that the potential applications of this idea extended far beyond the realm of art.

The discovery of startups like TeamGPT and CoPrompt, which were beginning to explore AI-driven collaborative environments, further fueled the vision. These platforms were designed to allow multiple users to prompt and interact with AI simultaneously, creating a new way of working together that was both efficient and innovative.

However, while these tools were impressive, they still did not fully align with the vision of ComplexChaos. The idea was to go beyond just using AI as a tool for collaboration; the goal was to integrate AI into the very fabric of the collaborative process, allowing it to act as a facilitator, translator, and innovator in its own right.

The Unveiling of ComplexChaos

The reimagined concept of ComplexChaos began to take shape. This time, the focus was not just on marketing but on solving complex global challenges through interdisciplinary collaboration. The idea was to create a platform where clients could define unique challenges and build custom teams from a vast array of disciplines, facilitated by cutting-edge machine learning systems. This was not just about bringing people together; it was about creating a space where diverse ideas could collide, merge, and evolve into innovative solutions.

The platform was envisioned to be flexible and accessible, allowing clients to tailor their teams based on their specific needs and budget. Whether they needed undergraduate students, PhDs, postdocs, or tenured professors, ComplexChaos would provide the tools to bring the right minds together. The project durations could range from a quick 3-hour workshop to an in-depth 2-month program, ensuring that the output matched the project's needs.

This vision was ambitious, but it was also grounded in the reality of the changing technological landscape. The advancements in AI and machine learning had reached a point where such a platform was not only feasible but also necessary. The world was facing increasingly complex problems that required innovative solutions, and ComplexChaos was poised to provide the framework for these solutions to emerge.

As I began sharing the seed idea with some of the people that have influenced my thinking, I realized I was onto something. Above you can see the reaction of Don Norman , the world famous designer. In 2004 I trained with him and Jakob Nielsen at the UX Week in Amsterdam where I volunteer during the week-long conference.

2. The Santa Fe Institute Inspiration

Breaking Down Academic Walls

The development of ComplexChaos was deeply inspired by the Santa Fe Institute , a research center known for its interdisciplinary approach to science. The Institute's philosophy of removing academic walls to foster collaboration among scientists from different fields resonated strongly with the vision of ComplexChaos. The idea was not just to bring experts together but to create an environment where they could freely exchange ideas and build on each other's knowledge.

The Santa Fe Institute, in New Mexico, is an independent research and education centre where complex systems research is undertaken and the science of complex adaptive systems was defined. Founded in 1984, the institute is a private establishment where scientists deal with the most significant complex phenomena.?

Inter-disciplinary research is undertaken in a collaborative manner, with scientists coming to the Santa Fe Institute from various universities, research institutes and government agencies.? The institute came into existence through the continuous collaboration and knowledge sharing between George Cowan (a senior fellow at Los Alamos Laboratory) and other colleagues at the laboratory.? The founding members were George Cowan, David Pines, Stirling Colgate, Murray-Gell-Mann, Nick Metropolis, Herb Anderson, Peter A. Carruthers and Richard Slansky.?

Only Pines and Gell-Mann were not scientists at Los Alamos at the time.? The aspiration was to develop and institute where scientists could focus on problem-driven scientific research, rather than paradigm-orientated science.? The original focus of the institute was to propagate the concept of the new research field of complexity theory and related systems.

Entropically, complexity is the halfway point between the singular zero information of utter sameness and the thermodynamic maximum of uncoordinated possibilities. ?Likewise, chaos is the halfway point between laminar flow and some near-triple-point plasma-like state. Self-organizing criticality the same. So they're already quite similar, in fact the zones of life, and of maximum information transfer.

The breakthroughs achieved by the Santa Fe Institute demonstrated the power of interdisciplinary collaboration. By bringing together scientists from diverse fields such as physics, biology, economics, and computer science, the Institute was able to tackle complex problems in ways that traditional, single-discipline approaches could not.

This approach was particularly relevant to the vision of ComplexChaos, where the goal was to leverage the collective intelligence of diverse teams to solve complex challenges.

The Power of Metaphors

One of the most intriguing aspects of the Santa Fe Institute's approach was the use of metaphors as a tool for sharing ideas across scientific fields. Metaphors, by their nature, bridge the gap between different domains, allowing concepts to be understood and applied in new contexts. This approach to communication and idea-sharing was particularly relevant to the development of ComplexChaos.

In the context of AI and interdisciplinary collaboration, metaphors could play a crucial role in translating complex ideas and facilitating communication between team members from different fields. For example, AI could be used to generate metaphors that simplify scientific concepts, making them accessible to non-experts. This ability to translate and communicate ideas effectively was seen as a key component of the ComplexChaos platform.

The inspiration drawn from the Santa Fe Institute was not just about adopting their methods but also about adapting and enhancing them with the latest technological advancements. The goal was to create a platform that could emulate the Institute's success in fostering interdisciplinary collaboration but on a much larger scale, leveraging AI to remove barriers and facilitate communication across diverse fields.

I called this "Google Translate for Human Cooperation".

3. Researching the space

Cognitive Synergy and Decision-Making

"Unleashing Cognitive Synergy in Large Language Models: A Task-Solving Agent Through Multi-Persona Self-Collaboration" We discovered this paper which explores how large language models can collaborate internally using multiple personas to enhance problem-solving abilities. The idea of multi-persona self-collaboration aligns closely with our vision for ComplexChaos, where diverse perspectives are brought together to tackle complex challenges. This research supports our belief in the power of cognitive synergy and has influenced our approach to using AI as an active participant in interdisciplinary collaboration.

In this work, they propose Solo Performance Prompting (SPP), which transforms a single LLM into a cognitive synergist by engaging in multi-turn self-collaboration with multiple personas. A cognitive synergist refers to an intelligent agent that collaborates with multiple minds, combining their individual strengths and knowledge, to enhance problem-solving and overall performance in complex tasks. By dynamically identifying and simulating different personas based on task inputs, SPP unleashes the potential of cognitive synergy in LLMs.

"From DDMs to DNNs: Using Process Data and Models of Decision-Making to Improve Human-AI Interactions" In this paper, the authors say that over the past decades, cognitive neuroscientists and behavioral economists have recognized the value of describing the process of decision making in detail and modeling the emergence of decisions over time. For example, the time it takes to decide can reveal more about an agents true hidden preferences than only the decision itself.

Similarly, data that track the ongoing decision process such as eye movements or neural recordings contain critical information that can be exploited, even if no decision is made. Here, we argue that artificial intelligence (AI) research would benefit from a stronger focus on insights about how decisions emerge over time and incorporate related process data to improve AI predictions in general and human-AI interactions in particular.

They argue for a stronger focus on decision-making processes in AI research. They introduce a computational framework that assumes decisions emerge from the noisy accumulation of evidence. This idea has inspired us to think about how we can integrate decision-making models into ComplexChaos, particularly in how AI can facilitate group decision-making by considering the process data, such as eye movements or neural recordings, to improve interactions between humans and AI.

Enhancing LLMs with emotions

"Large Language Models Understand and Can Be Enhanced by Emotional Stimuli" We found this paper particularly fascinating as it discusses how emotional stimuli can impact the performance of large language models. This research underlines the importance of incorporating emotional intelligence into AI, which is a concept we are exploring in the context of ComplexChaos. By understanding and reacting to emotional cues, AI can become a more effective facilitator in collaborative environments, particularly in sensitive areas like conflict resolution or negotiation.

Group Decision-Making and Consensus Building

"Group Decision-Making Based on Artificial Intelligence: A Bibliometric Analysis" This paper provided us with a comprehensive overview of how AI is being applied to group decision-making (GDM). It highlights that research in this area is rapidly increasing, particularly in domains like engineering and economics. The paper also identifies current trends, such as the use of deep learning to estimate the importance of experts in decision-making processes. This analysis has been instrumental in shaping our understanding of how ComplexChaos can leverage AI to support group decisions, especially in large, diverse groups where consensus is challenging to achieve.

In this paper, a systematic and highly automated bibliometric workflow has been followed to analyze the literature on group decision-making based on artificial intelligence. Their longitudinal analysis shows that:

• Research on AI-GDM is increasing as the number of papers and citations to those papers is growing substantially.
• Most research has been carried out by Chinese universities. Nevertheless, a few Spanish investigators lead research in terms of productivity and collaboration network centrality.
• Two basic bibliometric laws hold to a great extent, Lotka’s law and Bradford’s law, which model authors’ and journal productivity concentrations, respectively.
• AI-GDM is being applied to a variety of domains, including engineering, operations research management science, automation control systems, robotics, economic, telecommunications, imaging science, etc.
• Currently, themes such as terms-sets, analytical-hierarchical-process, Vikor-method, similarity-measure, consensus, consensus-reaching-process, and multi-attribute- group-decision-making are motor in AI-GDM research.
• In summary, the conceptual evolution of the AI-GDM research fields delved into seven thematic areas: multi-attribute/criteria in GDM, analytical network process, decision-making and uncertainty, fuzzy sets, recommender systems, consensus and majority, and agent systems.

Finally, recent literature on AI-GDM reveals the following trends and challenges:

• There is an increased need to support the consensus of huge groups of decision-makers. This need arises in several contexts, such as social networks, e-democracy platforms, crowd-funding systems, group recommender systems, etc. Those large groups are typically decomposed into smaller ones by applying different clustering algorithms, such as hierarchical clustering [58], discriminant analysis [59], etc.
• In classical GDM, a reduced group of experts needs to make a consensual decision. Presently, the experts’ group is often replaced by internet users’ opinions. As a result, natural language processing techniques have started to be applied for mining linguistic information that is subsequently processed by GDM systems [60].
• As AI-GDM problems become more complex, advanced models and simulations are required to support the experts’ group dynamics [61], e.g., for identifying the most influential experts, detecting manipulative and non-cooperative behaviors, etc.
• Deep learning has started to be used [62] for (i) estimating the importance (or weight) of the experts, their preferences, and their relationships, and (ii) learning the optimal settings of parameterized aggregation operators.

"AI-Powered Group Decision-Making" We explored this paper to understand better how AI can be used to facilitate fair, efficient, and trustworthy collective decisions for groups of agents. The research delves into various scenarios, such as elections and public policy-making, where AI can play a crucial role. This aligns with our mission to use ComplexChaos as a tool for solving global problems by making group decision-making more accessible and effective across different contexts.

Loomio, for example, is a flexible decision-making tool that helps you create a more engaged and collaborative culture, build trust and coordinate action. Loomio explains the Consensus Process, to build consensus for a decision you need to make together, to reach an agreement that satisfies the needs and concerns of all participants.

Check out these common decision making processes:

Advice process - Seek advice on a decision you need to make, with the advice of people impacted or who have expertise, so you can make a better decision for your organization.

Consent process - Seek consent on a decision you need to make, where there are no meaningful objections to your proposal, so you can make a fast decision that is 'safe to try' now.

Consensus process - Build consensus for a decision you need to make together, to reach an agreement that satisfies the needs and concerns of all participants.

Simple decision process - A simple process to introduce and discuss the decision to be made, listen and sense how people think about it, propose and decide. Start here if you are new to collaborative decision making.

Now, it turns out, that consensus presents a paradox. If you ask 100 people in the US what their definition of consensus is... you will get 100 different answers. I tried this at events like TED AI. Who would imagine that there is no consensus on what consensus means?

In a highly individualistic culture like we have in the United States, it is not surprising that consensus almost sounds like a bad word. Management prefers "disagree and commit" or "move fast and break things". But don't get me started on the unintended consequences. It is not surprising hence that in Europe, where I lived for 15 years (Germany and Spain), Europeans have a much different view on consensus.

ConsensUs: Supporting Multi-Criteria Group Decisions by Visualizing Points of Disagreement

In this research paper they claim that groups often face difficulty reaching consensus. For complex decisions with multiple latent criteria, discourse alone may impede groups from pinpointing fundamental disagreements.

To help support a consensus building process, they introduce ConsensUs, a novel visualization tool that highlights disagreements in comparative decisions. The tool facilitates groups to specify comparison criteria and to quantify their subjective opinions across these criteria.

ConsensUs then highlights salient differences between members. An evaluation with 87 participants shows that ConsensUs helps individuals identify points of disagreement within groups and leads people to align their scores more with the group opinion. They discuss the larger design space for supporting the group consensus process, and their future directions to extend this approach to large-scale decision making platforms.

"Decision-Oriented Dialogue for Human-AI Collaboration"

This paper introduced us to the concept of decision-oriented dialogues, where AI assistants collaborate with humans via natural language to help them make complex decisions. The emphasis on AI's role as a collaborator rather than just a tool is something we resonate with deeply at ComplexChaos. We are inspired by this research to design our platform to support decision-making dialogues, where AI actively engages with human users to navigate complex scenarios.

Non-Cooperative and Cooperative Social Choice Theories

In this Bachelor Thesis, Beāte Hermansone aims to study the relationship between Artificial Intelligence and collective decision-making models in order to understand how Artificial Intelligence could contribute to non-cooperative and cooperative social choice theories.

The relationship between AI and two cooperative decision-making theories namely, Nash bargaining solution and vote trading is examined as well as AI’s relatedness to three non-cooperative group decision-making theories, namely, Condorcet’s paradox, Arrow’s impossibility theorem and Black’s Median Voter theorem, is studied.

As the literature relating these concepts is scarce, a systematic literature review following PRISMA guidelines is performed to aim to fill in the existing gaps. The initial dataset incorporated 1,619 documents, which eventually led to a final corpus of 51 relevant articles after applying the exclusion criteria.

Results show that AI can be applied two one of the selected cooperative decision-making models, Nash bargaining solution, and the non-cooperative decision-making model Condorcet’s paradox. However, there are no studies indicating how Artificial Intelligence algorithms could facilitate the other examined social choice theories

From Consensus to Multi-Party negotiations

Multi-party negotiation is a complex, iterative process involving the exchange of views, ideas and perspectives among a number of parties that might include organizations, groups, regions, countries or individuals within larger entities. Complexity may appear chaotic, especially in the absence of structure and leadership.

In this workshop published by the U.S. Institute for Environmental Conflict Resolution and John Godec with Brian Manwaring, we learn that traditional bargaining is often about relative power and willingness to use it against each other, often at the expense of a better agreement or relationship; however interest-based negotiation (IBN) has proven its effectiveness in multi-party settings.

"Multi-party negotiation is not merely two-party negotiation with “more people.” Multi-party dynamics generate complexity across all of the dimensions of the Frames of Reference triangle: group dynamics increase exponentially because of the multiplicity of people, interests, and differing Best Alternatives to a Negotiated Agreement (BATNAs); relational dynamics become more complex, such as role definition and issues of unequal power and control; and substantive issues also increase in complexity, such as increased potential for misinformation, differing viewpoints and interpretation of data and different views on the mission and mandates of the group."

Innovation and Creative Problem-Solving

In "The Crowdless Future? How Generative AI Is Shaping the Future of Human Crowdsourcing" we read how generative AI is changing the landscape of human crowdsourcing. The paper written by Léonard Boussioux , Miaomiao Zhang , Vladimir Jacimovic and Karim Lakhani from Harvard Business School, argue that AI can replace traditional crowdsourcing methods by automating the generation and evaluation of ideas.

Their results indicate that while human crowd solutions exhibited higher novelty—both on average and for highly novel outcomes—human-AI solutions demonstrated superior strategic viability, financial and environmental value, and overall quality.

"Notably, human-AI solutions co-created through differentiated search, where human-guided prompts instructed the large language model (LLM) to sequentially generate outputs distinct from previous iterations, outperformed solutions generated through independent search." they say.

Chain of Density Prompting

"From Sparse to Dense: GPT-4 Summarization with Chain of Density Prompting" This paper explores techniques for generating detailed yet concise summaries using GPT-4. The idea of selecting the "right" amount of information to include in a summary resonates with our approach to using AI in ComplexChaos. We see potential in applying these summarization techniques to distill complex discussions into actionable insights, which can then be used to drive decision-making and innovation within teams.

"Decision-Oriented Dialogue for Human-AI Collaboration" In this paper, the researchers describe tasks that involve decision-oriented dialogues, where AI assists humans in making complex decisions through natural language collaboration. This concept of AI facilitating human decision-making through dialogue aligns with our goals for ComplexChaos. We are particularly interested in how AI can help synthesize and prioritize information in real-time, guiding teams toward consensus and innovative solutions.

How Might We Craft an Interactive Artificial Society that Reflects Believable Human Behavior?

In Generative Agents: Interactive Simulacra of Human Behavior, the authors Joon Sung Park from Stanford University, Carrie Cai from Google Research, Meredith Morris and others from Google DeepMind, they sat that "believable proxies of human behavior can empower interactive applications ranging from immersive environments to rehearsal spaces for interpersonal communication to prototyping tools. In this paper, we introduce generative agents: computational software agents that simulate believable human behavior. Generative agents wake up, cook breakfast, and head to work; artists paint, while authors write; they form opinions, notice each other, and initiate conversations; they remember and reflect on days past as they plan the next day."

"To enable generative agents" they describe "an architecture that extends a large language model to store a complete record of the agent’s experiences using natural language, synthesize those memories over time into higher-level reflections, and retrieve them dynamically to plan behavior.

Along the lines of synthetic research, here's a demo of how Synthetic Users works:

AI Ethics and Collective Decision-Making

"Collective Constitutional AI: Aligning a Language Model with Public Input" We explored this paper to understand how collective input from society can shape AI development. The research discusses training AI models to behave according to a collectively-designed constitution, which is a concept that intrigues us at ComplexChaos. We believe that aligning AI behavior with public input can lead to more ethical and democratic decision-making processes.

"A Proposal for Importing Society’s Values: Building Towards Coherent Extrapolated Volition with Language Models" This paper presents a proposal for aligning AI with societal values by incorporating language models into the decision-making process. The idea of using AI to reflect and amplify societal values is something we are deeply interested in at ComplexChaos. This research has inspired us to think about how we can design our platform to not only facilitate collaboration but also ensure that the outcomes align with the broader values and ethics of the communities involved.

Cooperative AI

In this as this point that I discovered the Cooperative AI Foundation where Research Director Lewis Hammond leads studies as a charitable entity, backed by a $15 million philanthropic commitment from Polaris Ventures. CAIF's mission is to support research that will improve the cooperative intelligence of advanced AI for the benefit of all. Some of the trustees include Dario Amodei founder of Anthropic along with Allan Dafoe and Eric Horvitz , Chief Scientific Officer at Microsoft.

Some of the talks that the foundation hosted includes this one above by Wolfram Barfuss , a research scientist at the Tübingen AI Center at the University of Tübingen in Germany and a guest researcher at the Potsdam Institute for Climate Impact Research and at Princeton University. Wolfram writes:

Cooperation is a crucial organizing principle across all scales, from cells to societies. The question, however, of how collective, cooperative behavior emerges from interacting, self-regarding individuals and feeds back to influence those individuals remains open. Complex systems theory focuses on how cooperation emerges from deliberately simple individuals. Fields such as machine learning and cognitive science emphasize individual capabilities in changing environments without considering the collective level much. To date, however, little work went into modeling emergent collective, cooperative behavior from individually self-learning agents. As a result, cooperative intelligence misses out on a holistic complex systems approach, yet-to-be-discovered basic principles on collective cooperative AI cannot find their way into advanced AI systems, and multiagent (human-) machine learning applications may lose out on fulfilling their full cooperative potential. I am proposing a research agenda on collective cooperative AI that will unify classic approaches of multiagent learning, such as reinforcement learning and multi-agent systems, with complex systems theory, i.e., non-linear dynamics and evolutionary game theory.

The foundation also organized the AI 4 Institutions program providing grants for super interesting research, for example, "Conversational AI for Non-Human Representation in Decision-Making" by Shu Yang Lin .

In their project "Say Hi to the River" Shu says "some legal systems have already begun to recognize the rights of rivers as legal entities, such as the Whanganui River in New Zealand and the Ganges and Yamuna Rivers in India. This progressive shift in legal and philosophical thinking provides a strong foundation for the project's goals and highlights the growing importance of recognizing rivers as stakeholders in our society."

Imagine "if AI can facilitate dialogue between humans and non-human entities. Such discussions can serve as an interface of care, allowing non-human entities to express views and concerns... conversational AI can collect data on the well-being of non-human entities through voice or text-based conversations. This information can inform decision-making processes and deepen our understanding of the impact of human activities on non-human entities."

"A conversational AI can enable non-human entities to express perspectives, state interests, co-create plausible solutions with other entities (e.g., in a context of deliberative workshop), and/or communicate with decision-makers directly."

Another project by Lewis Hammond proposes "Natural Language Negotiation" because "In the near future, individuals may have their own personal AI assistants – in essence, a more sophisticated version of Apple’s Siri or Amazon’s Alexa – which may need to engage in many lower stakes negotiations on behalf of their users, such as finding mutually convenient meeting times."

Fine-tuning Language Models to Find Agreement Among Humans with Diverse Preferences

In this research study by Michiel Bakker from Google DeepMind , they asked how might a machine help people with diverse views find agreement?

They fine-tuned a 70 billion parameter LLM to generate statements that maximize the expected approval for a group of people with potentially diverse opinions. Human participants provide written opinions on thousands of questions touching on moral and political issues (e.g., “should we raise taxes on the rich?”), and rate the LLM’s generated candidate consensus statements for agreement and quality.

A reward model was then trained to predict individual preferences, enabling it to quantify and rank consensus statements in terms of their appeal to the overall group, defined according to different aggregation (social welfare) functions. The model produced consensus statements that are preferred by human users over those from prompted LLMs (> 70%) and significantly outperforms a tight fine-tuned baseline that lacks the final ranking step.

Further, their best model’s consensus statements are preferred over the best human-generated opinions (> 65%). We find that when we silently constructed consensus statements from only a subset of group members, those who were excluded were more likely to dissent, revealing the sensitivity of the consensus to individual contributions. These results highlight the potential to use LLMs to help groups of humans align their values with one another.

In my conversation with Michiel, he said:

• Project Value: what we’re building is incredibly valuable, particularly in its potential to drive impactful change.
• Meta-Reviews and Aggregation: When aggregating individual reviews, it's crucial to focus on how aggregated preferences are summarized. This is a key challenge in ensuring the accuracy and fairness of the summarized outputs.
• Challenges: One of the major hurdles identified is determining the right software infrastructure. For instance, integrating into platforms like Slack is promising, but the real challenge lies in whether such integrations will significantly change user behavior.
• Recommended Reading: He suggested reading the paper Generative Social Choice by Ariel Procaccia and Manuel Butti, which could offer valuable insights into our project.
• Support for Tech Application: He expressed 100% support for the application of this technology, highlighting its broad potential.
• Custom Solutions: He mentioned a pitch he received about budget planning, noting it as highly valuable but requiring very custom solutions.
• Personal Context in Automation: Automation should consider the context. For example, when booking a meeting, understanding whether it’s urgent or can be rescheduled based on calendar and email data is crucial.
• Autonomy and Error Tolerance: The level of autonomy granted to agents should depend on the task's importance. For instance, minor errors in booking meetings might be acceptable, but mistakes in organizing a group holiday would be more problematic.
• Current Capabilities: He believes that the current state of the art in AI is sufficient for our agent to communicate with other stakeholders' agents. This would involve a combination of prompt engineering and fine-tuning LLMs.
• Scoring User Authority/Expertise: In theory, models can do almost anything if they have access to user data. The real question is how effectively they can execute these tasks.
• Personalization and Translation: Personalization could be achieved by conditioning on previous interactions or using system instructions to translate and bridge gaps in communication.

Better Forecasters than Humans?

Interesting papers show how groups of LLMs, especially when prompted correctly and retrieving relevant content, can be comparable and, in some cases, better forecasters than humans.“

[...] Current LLMs are able to provide a human-crowd-competitive level of accurate forecasting about future real-world events. In order to do so, it is sufficient to apply the simple, practically applicable method of forecast aggregation, manifesting as the LLM ensemble approach in the so-called silicon setting.

This replicates the human forecasting tournament’s ‘wisdom of the crowd’ effect for LLMs: a phenomenon we call the ‘wisdom of the silicon crowd.’“[...]

A retrieval-augmented LM system designed to automatically search for relevant information, generate forecasts, and aggregate predictions. [...] On average, the system nears the crowd aggregate of competitive forecasters, and in some settings surpasses it”

A Bottom-Up Institutional Approach to Cooperative Governance of Risky Commons

This article in Nature.com explains how avoiding the effects of climate change may be framed as a public goods dilemma in which the risk of future losses is non-negligible while realizing that the public good may be far in the future.

On the contrary, global institutions provide, at best, marginal improvements regarding overall cooperation. Our results clearly suggest that a polycentric approach involving multiple institutions is more effective than that associated with a single, global one, indicating that such a bottom-up, self-organization approach, set up at a local scale, provides a better ground on which to attempt a solution for such a complex and global dilemma.

Mindstorms in Natural Language-Based Societies of Mind

Both Minsky's "society of mind" and Schmidhuber's "learning to think" inspire diverse societies of large multimodal neural networks (NNs) that solve problems by interviewing each other in a "mindstorm."

Recent implementations of NN-based societies of minds consist of large language models (LLMs) and other NN-based experts communicating through a natural language interface. In doing so, they overcome the limitations of single LLMs, improving multimodal zero-shot reasoning. In these natural language-based societies of mind (NLSOMs), new agents -- all communicating through the same universal symbolic language -- are easily added in a modular fashion.

To demonstrate the power of NLSOMs, we assemble and experiment with several of them (having up to 129 members), leveraging mindstorms in them to solve some practical AI tasks: visual question answering, image captioning, text-to-image synthesis, 3D generation, egocentric retrieval, embodied AI, and general language-based task solving. THey view this as a starting point towards much larger NLSOMs with billions of agents-some of which may be humans. And with this emergence of great societies of heterogeneous minds, many new research questions have suddenly become paramount to the future of artificial intelligence.

What should be the social structure of an NLSOM? What would be the (dis)advantages of having a monarchical rather than a democratic structure? How can principles of NN economies be used to maximize the total reward of a reinforcement learning NLSOM? In this work, we identify, discuss, and try to answer some of these questions.

The Less You Know, The More You Know?

This was exactly my thesis in 2012. PEople who are not subject matter experts can elicit and ask thought-provoking questions.

In this paper "Putting the Pieces Back Together Again: Contest Webs for Large-Scale Problem Solving" they explain onr interesting finding from Climate CoLab was that the ideas selected as most promising were 3 to 5 percent more likely to come from contributors who had less experience in the field of climate change and had not attended graduate school. And although fewer women submitted ideas overall, those who did were 5 percent more likely to have them judged promising. For more, see the CCI conference paper on Broad Participation in Online Problem Solving.

4. From Vision to Prototype: ComplexChaos in Action

The First Draft: ComplexChaos.ai

With the inspiration and ideas in place, the next step was to turn the vision into a reality. The first draft of ComplexChaos.ai was developed as a collaborative platform designed to unleash the power of interdisciplinary collaboration. The platform was envisioned as a tool that would allow clients, whether they were non-profits, corporations, or government agencies, to define unique challenges and assemble custom teams from a wide range of disciplines.

The idea that AI could mediate and facilitate such a complex process seemed a monumental endeavor, and at the same time, the motivation to solve the problem.

The process was designed to be user-friendly and flexible. Clients could start by defining their challenge or problem, and then explore the limitless possibilities by inviting experts from various fields to join their project. The platform would offer a range of expertise levels, from undergraduate students to tenured professors, allowing clients to tailor their team based on their needs and budget.

Once the team was assembled, the platform's machine learning system would spring into action, identifying the perfect candidates from an extensive database. The goal was to create a seamless and efficient process that would allow clients to build their dream team quickly and easily.

Visual Collaboration Tools

As the platform began to take shape, it became clear that visual collaboration tools would be a crucial component of the ComplexChaos experience. These tools would allow teams to map out their ideas, explore different perspectives, and collaborate more effectively.

The research into tools like Plectica , which offers a visual collaboration platform for systems thinkers, and InfraNodus, which creates real-time knowledge graphs during ChatGPT conversations, provided valuable insights into how these tools could be integrated into ComplexChaos.

InfraNodus uses a combination of text mining, network analysis, data visualization, NLP, and artificial intelligence to provide insights about any discourse and enhance a research process. Network analysis is the secret sauce that sets InfraNodus apart from other text analysis and AI-based tools. Thanks to this approach, we can design highly relevant prompts for AI systems and use the most advanced language models to derive meaning from text and see how it can be developed further.

The idea was to create an environment where team members could visualize complex ideas and relationships, making it easier to understand and navigate the challenges they were working on. This visual approach to collaboration was seen as a way to enhance creativity and innovation, allowing teams to explore new possibilities and solutions.

The First Sketches: Bringing Ideas to Life

With the first draft in place and the visual collaboration tools identified, the next step was to develop a prototype of ComplexChaos. The prototype was designed to showcase the platform's potential and demonstrate how it could facilitate interdisciplinary collaboration on complex problems.

The initial prototype allowed users to have private ChatGPT conversations about their projects while viewing the group's output in a separate window. The idea was to create a space where clients could interact with a group of experts as if they were a single entity, mediated by an AI facilitator. This approach allowed for a more streamlined and efficient collaboration process, where individual experts could contribute their knowledge and insights without the need for constant back-and-forth communication.

As I was exploring use cases synchronistically I cam across this idea in @Reid Hoffman's book:

"A teacher can use large language models to facilitate interactive discussions or debates among students on various topics or issues. The AI can act as a moderator or a participant, providing prompts, questions, facts, opinions, or counterarguments that stimulate critical thinking and dialogue. The teacher can observe the interactions and intervene when necessary, or join the conversation and provide guidance or feedback."

5. The Idea Maze

AI-Powered Collaboration

One of the key ideas was the integration of AI into the collaborative process. The platform would be designed to use AI not just as a tool but as an active participant in the collaboration. AI would facilitate interactions, translate complex ideas, and provide insights that would help teams navigate their challenges.

For example, tools like MyMap.ai were explored for their ability to visually fork ChatGPT threads, allowing users to map out different scenarios and explore various possibilities. This visual approach to collaboration was complemented by the use of AI-driven community summaries, which could distill the collective wisdom of a group into a concise and actionable format.

The goal was to create a platform where AI could act as a bridge between different disciplines, translating and synthesizing ideas in real-time. This would allow teams to move quickly and efficiently from problem definition to solution development, leveraging the power of AI to enhance their creativity and productivity.

Exploring One-to-Many and Many-to-One Use Cases

As the platform evolved, it became clear that the traditional multiplayer collaboration model had its limitations. To maximize the platform's potential, the focus shifted towards one-to-many and many-to-one use cases. This approach would allow for greater scalability and broader applications of the platform.

The prototype also included features for summarizing and visualizing social media replies, creating landing pages for posts, and generating community-driven content. These features were designed to enhance the collaborative experience and provide users with valuable tools for managing and sharing their ideas.

Dear Mr. President,

While some of us celebrate the reported job gains under your administration, many remain skeptical of the broader economic picture. A few of us offer congratulations and thanks, with some expressing love through heart emojis.

But for most Americans, the reality is more difficult. Though jobs may be increasing, many of us remain unemployed or underemployed. Those who have found new jobs often struggle as salaries fail to cover today's high inflation and costs. Gas prices are soaring along with grocery bills, making it hard for working families to get by.

Some of us view the reported numbers with distrust, accusing your administration of lies or manipulation. A minority speculate about conspiracies and worsening trends. Others resort to mocking memes or joking criticisms of your "Bidenomics."

Numbers can be claimed, but our lived experiences cannot be denied. There remains a long road ahead to rebuild an economy that works for all Americans, not just those at the top. Our voices remind you that headlines and statistics must be grounded in the struggles and hopes of real people.

With gratitude for your service, we urge you to hear the nuances within these responses: the difficulties, the distrust, and the hope for tangible improvement in our lives and livelihoods.

Sincerely, The American people

One idea was to enable social media influencers to engage with their audiences at scale, using AI to summarize replies and generate community-driven content. This approach not only made social media content more accessible but also provided valuable insights into the collective opinions and preferences of a large group of people. Similarly, here’s an example parsing Twitter’s replies to this user:

Collective Answer:

1. Emphasis on Law and Enforcement: We strongly advocate for stricter enforcement of laws and regulations. Many of us, like @FLCons and @usmcvet97, think the absence of real consequences is enabling disruptive behavior. In line with @JAX2866, we insist on more law enforcement, more arrests, and greater penalties for disruptive actions.
2. Accountability and Responsibility: We express concerns about a lack of accountability and responsibility among individuals involved in disruptive incidents. As @Rick Donkel and @KymSredinski point out, this starts with families and parents. We support holding parents accountable for the actions of their children as part of the solution.
3. Societal and Cultural Factors: We notice underlying societal issues contributing to these behaviors. As @RealStarrBurst, @FabioSi57695354, and @Will G. suggest, various cultural dynamics and historical issues, including irresponsible parenting and community problems, are exacerbating the issue. A shift in these areas is deemed necessary.
4. Political and Institutional Criticisms: We, like @avibebert and @thewriterme, question the role of current leadership and political figures in resolving these challenges. There is a sentiment among us that significant change cannot occur without a shift in political strategy and leadership.
5. Conspiracy Theories and Speculation: Some among us, including @RebelPrinces29 and @mehrensp7, harbor suspicions about these incidents being either staged or intentionally propagated for various reasons. We acknowledge these viewpoints as indicative of a broader distrust in established narratives.
6. Call for Behavioral Change and Self-Reflection: We, like @DaleMeech and @Grace Vasquez, think that there's a need for cultural and internal community changes. We believe the people involved in disruptive behaviors must acknowledge their actions and work toward self-improvement.
7. More Radical Approaches: A few of us, like @LeastIDidThat and @species_x, advocate for more extreme solutions like conscription or compulsory military service. Though not universally accepted among us, these opinions represent a desire for immediate, impactful change.
8. Divine and Spiritual Intervention: There is a section among us, such as @LloydFromUSA, who think that the re-introduction of spirituality and faith into public life might provide a solution to these problems.

In summary, we collectively see the issue as multi-faceted, requiring an overhaul in law enforcement, political leadership, societal values, and individual accountability. While our solutions may differ, our ultimate goal remains the same: a safer, more harmonious society for all.

Another idea was to use AI to facilitate interviews at scale, summarizing and visualizing the responses in a way that made it easier for clients to understand and act on the information. This approach had the potential to revolutionize how organizations conducted research and gathered insights, providing a more efficient and effective way to engage with large groups of people. We designed this idea with Frank Scott Krueger and his team at Good Knife .

Mediation with Bots

Making the "right decision" is not just a matter of judgement, but of clear intentions. In an experiment that we did with synthetic agents, two agents play the prisoners dilemma. When we added a third agent to mediate the negotiation, chances of success increased by 70%.

In a small experiment with a?Slack-chatbot called Harmony, a team released a beta version of a chatbot powered by GPT, designed for facilitating Double Crux dialogues between two users on Slack or Discord.

Similarly there was a pre-LLM research experiment in Finland using?Chatbots Facilitating Consensus-Building in Asynchronous Co-Design?that suggests mediation with chatbots works:

Finally, Remesh published a paper on "Faster Peace via Inclusivity: An Efficient Paradigm to Understand Populations in Conflict Zones".

6. Challenges and Iterations

Pitch Decks and Prototypes

The development of ComplexChaos was not without its challenges. One of the key hurdles was creating a compelling pitch that would attract investors and talent. The first pitch deck focused on the platform's ability to facilitate interdisciplinary collaboration and its potential to solve complex global challenges.

The pitch was well-received, and it helped secure the initial angel funding needed to develop the platform further. However, as the idea evolved, it became clear that the original vision needed to be refined and expanded. The prototypes developed during this phase allowed for testing and iteration, providing valuable feedback that helped shape path forward.

The Atomic Network

Another significant challenge for scaling the platform and ensuring its adoption by a broad user base was the need for network effects. The traditional multiplayer collaboration model required multiple users to see the value of the platform, which could be a barrier to entry. To address this, the focus shifted towards one-to-many and many-to-one use cases, which allowed for greater scalability and broader applications.

The development of AI-driven features, such as community summaries and interview facilitation, provided new ways for users to engage with the platform. These features not only enhanced the collaborative experience but also made the platform more accessible to a wider audience.

Iterating on the User Experience

User experience was a key focus during the development of ComplexChaos. The goal was to create a platform that was not only powerful but also intuitive and easy to use. This required multiple iterations of the platform's design and features, with feedback from users playing a crucial role in shaping the MVP.

As we started to research the ideal customer it became obvious that membership-based organizations would need our solution the most.

• Give every member the power to become a change maker
• Take people’s passionate point of views and make them actionable
• Passion without aggression
• Turn non-productive discussions into constructive dialogue.
• Turn passion into cooperative action.

The we saw how OpenAI launched the ability to bring ChatGPT into your meetings.

A day later at Google I/O Chip was introduced:

Someone I talked to asked me to consider this issues eventually:

1. How are you and your team thinking about the risks (e.g. The algorithms that they use to train their models/). I suspect it will have a significant influence on how information is presented to people and could lead to bias in decision making.
2. How will you ensure the data informs decision making rather than influencing the decision making?
3. Is there a potential that presenting the data in a certain way will lead to groupthink rather than creativity?
4. How is it that the algorithm decides which voices to amplify and by how much and how does it weight them and if everyone is being presented with a consistent view but there is less time spent asking questions and clarifying

7. Related and Relevant Startups I Found Along the Way:

Collective Intelligence

• Although a non-profit, the Collective Intelligence Project has a very similar roadmap to ComplexChaos. In this blog post they describe the four tools that they are working on under the leadership of Divya Siddarth , namely Open-Source Collective Constitutional AI (OSCCAI), Scenario Construction & Exploration Tool for Policymaking, Collective Intelligence with Agents (CIwA) and Voice of Nature
• The AI Objectives Institute, another non-profit, has been working on Talk to the City, an open-source LLM interface for improving collective deliberation and decision-making by analyzing detailed, qualitative data. It aggregates responses and arranges similar arguments into clusters.
• Unanimous.ai under the leadership of Louis Rosenberg has been working on swarm intelligence for a long time. Their project ThinkScape, enables groups of any size (up to 400 participants) to hold productive real-time conversations that optimize?group insights and amplify collective intelligence.
• Janars is a real-time platform for leaders of change, mediators and facilitators to enable deep collaborative AI-supported conversations about complex challenges and/or learning opportunities.
• CrowdSmart: takes the complexity and friction out of the collaboration process, so people can think creatively, align strategically and work together to get things done

Innovation and idea management

• Viima is a powerful innovation platform designed to democratize innovation.
• Ideanote is a fully customizable idea management flow from start to finish
• Hives collects, prioritizes, and keep track of ideas & feedback from employees and customers.
• Qmarkets is a software to engage employees, stakeholders and customers and drive revolutionary business innovation.
• Bright Idea is to source better ideas, ensure data-driven decisions, and enforce disciplined execution on your innovation projects.

AI-interviews for research

• In a very undifferentiated red ocean of customer intelligence, there’s obvious demand for the research use case with startups like?Remesh?($39M Series A),?Reveal?($2M seed) or?Outset?($4M seed) and?Voicepanel?(last two funded by YC). Because this is such a low hanging fruit use case we expect every player who’s in the survey/research space will incorporate 1:1 AI-interviews, with leaders like Qualtrics and SurveyMonkey. Some other AI-interview startups include Trove, Perspective, and unSurvey.
• The field of research using synthetic data is also super interesting (see?Synthetic Users?and?CulturePulse). Culture Pulse allows you to create digital twins of your organizations and was hired by the UN to untangle the Israeli-Palestinian crisis.
• ThoughtExchange is a collective Intelligence Platform powered by patented anti-bias technology. Leaders use ThoughtExchange to leverage their people’s collective intelligence and quickly gain critical insights for effective decision-making. Engage ten stakeholders or a community of 10,000 people to drive strategic discussions at scale. Unlike traditional surveys, respondents feel more comfortable answering questions candidly, they say.

Consensus-building

• D-Agree?(Japan) streamlines all discussions and consensus building.
• ConsiderIt: forum to visually summarize what your community thinks and why
• Pol.is?(open source) which has been widely used in Taiwan for digital democracy and powers the "community notes" of Twitter/X.
• RadicalXchange?(powered by?Pol.is)
• and Harmonica.chat, an early stage tool for consensus across digital channels.
• Deepmind has put out research on?Fine-tuning language models to find agreement among humans with diverse preferences. Lastly, the Collective Intelligence Project has a?very similar vision and roadmap as ours.
• Welphi: Build web surveys with real feedback among participants by implementing the Delphi method.
• Dembrane: Common Ground can be conceptualised as a multi-player variant of Pol.is. Instead of voting on Statements in isolation, we match participants into small groups of three people where they are encouraged to deliberate over the Statements they vote on, and where an AI moderator powered by GPT4 synthesises new Statements from the content of their discussion.
• DeliberAIde: to enhance dialogue-based decision-making with deliberative AI technologies that match the public interest and follow the highest ethical standards.
• Stanford Online Deliberation Platform: a video discussion platform for groups of 8-15 people. The platform is designed to facilitate a structured and equitable conversation with better opportunity for participants to speak up. It is developed by the Stanford Crowdsourced Democracy Team in collaboration with the Deliberative Democracy Lab.

Enterprise Listening tools

It’s interesting that very few companies offer tools for listening to enterprise Slack conversations and when do it is mostly for compliance/risk/legal issues or HR sentiment analysis, like?AwareHQ?($87M in funding).

Group decision-making and intelligence

• Balloon?($5M) with focus on DEI and decisions without groupthink, which helps reduce meetings.
• Cloverpop?($9M) allows you to frame your decision, clearly define decision rights, connect relevant data and set your team up for success in every big decision you make. You can tap into Cloverpop’s Decision Playbook library or create your own playbooks.
• 1000minds which is used for allocating budgets or other scarce resources across competing business projects or investments, e.g. for capital or operations expenditure (CAPEX or OPEX) decision-making, procurement, etc. as well as conjoint analysis, sometimes also known as a “discrete choice experiment” (DCE) or “choice modeling”, is a survey-based methodology for revealing consumers’ feelings about the attributes or characteristics of a business’ products, e.g. a smartphone or an airline’s services.

Strategy alignment

• The founders of DotWork, who sold their previous company to Atlassian, have been working on a new strategy platform along with their Uncertainty Project where Chris Butler writes about autonomous agents in decision making. After raising their Series A they discuss in this blog post the Patterns of Decision Architecture.
• Alignment: Create, align, and execute your vision with Alignment's AI-assisted strategy platform

Agent-guided platforms

We see an opportunity to let decision-making methodologies (see examples?or any template from?Miro/Mural) to become “alive” by developing both visual UI and AI-agents that guide groups async or even single-player for decision rehearsal.

Each user could have its own personal agent (that learns preferences, background, skills, knowledge profile and how they think about decisions, aka their personal context) which will become smarter over time to represent the users in decision-making processes connected with the enterprise context (whether indexing ourselves or connecting via API with enterprise AI search companies like?Akooda,?Glean, or meeting transcription apps, etc).

• Frame: Manage your business end-to-end with a unified set of collaboration apps & connected Al employees.
• Lindy.ai: The world’s easiest way to build AI automations to save you time and grow your business.
• Wand.ai: redefining how humans and AI collaborate to accomplish the toughest tasks

Policy-making:

• Decidim: is an open-source participatory democracy platform that allows organizations and governments to engage with stakeholders through consultations, debates, and collaborative decision-making processes.
• Delib: provides digital democracy tools like Citizen Space and Dialogue, which help in managing consultations, crowdsourcing ideas, and engaging with stakeholders
• Participedia: is a global community and platform that collects and shares knowledge about participatory democratic innovations, including tools and methodologies for stakeholder engagement in policy-making
• Consul: is an open-source platform for citizen participation that supports collaborative decision-making processes, enabling stakeholders to propose, discuss, and vote on policies.
• Civis Analytics: uses data science and AI to help organizations understand stakeholder opinions, segment audiences, and engage effectively in policy discussions
• Zencity: is an AI-driven platform that aggregates and analyzes feedback from various sources (social media, surveys, etc.) to provide insights into public opinion and stakeholder sentiment.
• MindMixer: facilitates community engagement by allowing stakeholders to share ideas, provide feedback, and collaborate on policy proposals
• Enthelo: uses algorithms to find the most broadly supported policy options by engaging stakeholders in collaborative decision-making processes.
• EngageSuite: is a suite of tools designed to enhance stakeholder engagement through surveys, consultations, and collaborative platforms
• Systemic: a game that simulates how policy-making systems function and possible shifts that can be made to improve policy outcomes. Systemic has been inspired by the world of board games.

8. The Future of Innovation: ComplexChaos and Beyond

Exploring New Horizons

As ComplexChaos continues to evolve, the focus remains on leveraging AI to facilitate cognitive synergy, drive innovation, and solve global challenges. The platform needs to be designed to be flexible and adaptable, allowing it to grow and evolve with the needs of its users.

One of the key areas of exploration is the integration of decision-making processes into the platform. By leveraging AI to facilitate group decision-making, ComplexChaos has the potential to revolutionize how organizations approach complex problems. This includes the development of tools that can help teams navigate the decision-making process, from problem definition to solution implementation.

Another area of exploration is the use of AI to facilitate consensus-building. The ability to bring together diverse perspectives and arrive at a consensus is crucial for solving complex problems, and AI has the potential to play a key role in this process. By leveraging the power of AI, ComplexChaos can help teams navigate the complexities of group decision-making and arrive at solutions that are both innovative and effective.

The Role of AI in Shaping the Future

The role of AI in shaping the future of innovation cannot be overstated. As AI continues to evolve, its potential to facilitate interdisciplinary collaboration and drive innovation will only increase. ComplexChaos is at the forefront of this evolution, providing a platform that leverages AI to enhance creativity, collaboration, and problem-solving.

The future of ComplexChaos lies in its ability to adapt and evolve with the changing technological landscape. As new tools and technologies emerge, the platform needs continue to integrate them, providing users with the most advanced and effective tools for collaboration and innovation.

Below you can read a write up by our investor and Engineering Fellow, G?khan Bakir :

9. Understanding My Big Why

On August 20th, 2020, the CZU complex lighting wildfires raged through the Bay Area. My wife and I left home a day or two before the mandatory evacuations in Boulder Creek, Santa Cruz Mountains.

When we came back to save more belongings, we drove to the house of our friend Diego Saez Gil . The sheriff was blocking the road on highway 236 but somehow I managed to convince them to let us drive through. "At your own risk" she said, "we are not coming to rescue you!".

As we approached the house it was clear his property was still safe. Desperately we looked everywhere to find a key. But after a few minutes it seemed like the only option was to break a window.

My wife, Tina, reaching a point of anxiety, yelled "we need to go, this is too dangerous!". There was smoke everywhere but I wanted to save something, a piece of art, a book. But at the same time it felt extraordinarily violent to break-in.

It was just impossible to imagine that a few hours later the house would be gone.

Diego, who happens to be the founder of Pachama, a climate-tech startup, later wrote a post "On Losing Everything to the Climate Crisis, except for Hope".

It is meaningful that now my house is taken by the consequences of climate change, and that those forests will need restoring soon. You can’t make our mission more personal to me now. The consequences of climate change will not be an abstract idea in a distant future, but the biggest disaster of my life. This, of course, renews my determination to give my all, for the rest of my life, to work on this purpose. And I feel very hopeful that we can make a dent, that we are still on time to reverse a big part of this planetary catastrophe and leave a safe and beautiful planet to our future generations. Let’s do it.

Experiencing first hand the devastating consequences of extreme weather, I asked myself "why did it take 10 years to sign the Paris Agreement?". Of course, I am not na?f, and I know that capitalism and global geopolitics are the primary reason, but what if AI could help humans understand each other better, and hence support collective decision-making efforts?"

Reimagining Human Cooperation at Scale

The co-founder of DeepMind, Mustafa Suleyman, is also known for co-founding the conflict resolution firm Reos Partners. This firm specializes in addressing complex, systemic challenges, bringing together diverse stakeholders to find common ground and resolve conflicts.

Mustafa wrote "One project I facilitated in 2009 at the Copenhagen climate negotiations involved convening hundreds of NGOs and scientific experts to align their negotiating positions. [...] I was there trying to help facilitate the process, perhaps the most complex, high-stakes multiparty negotiation in human history, but from the start it looked almost impossible".

As ComplexChaos continues to grow, the focus will be on expanding its reach and impact. This includes exploring new markets and applications for the platform, as well as developing partnerships with organizations that share the vision of interdisciplinary collaboration.

While we are experiencing the peak of expectations with GenAI, there are compounding effects of converging technology that is still on the rise of innovation. We shall be prepared to ride the wave if we focus on our mission to help humanity cooperate at scale rather than automating people out of their jobs.

“If you find yourself deciding between only two choices, maybe you need to widen the aperture,” Botha said on Sequoia’s “Crucible Moments” podcast, in a recent episode featuring billionaire entrepreneur Jack Dorsey. “Is there a third or fourth option that you haven’t even considered that you need to throw into the mix to really test whether that is the right path?”

The goal is to make ComplexChaos the go-to platform for solving complex problems, whether they are in the fields of science, technology, business, or beyond. By providing a flexible and powerful platform for cooperation, ComplexChaos has the potential to drive innovation and create solutions that we hope can change the world.

Ricardo Baeza-Yates (long time VP of Research at Yahoo! Labs and now Director of AI Experimental Research at Northwestern University) told me that everything we are doing is covered by the ACM Conference on Computer-Supported Cooperative Work and Social Computing. This means we are definitely not alone.

10. Conclusion

A Call to Innovators

ComplexChaos is more than just a platform; it's a movement towards unlocking the potential of interdisciplinary collaboration. In a world where complex problems require innovative solutions, the ability to bring together diverse perspectives and expertise is more important than ever.

Here's the pitch deck that we used to raise our $2.5M pre-seed.

As we look to the future, we invite innovators, thinkers, and creators to join us in shaping the next frontier of collaborative problem-solving. Whether you are a scientist, artist, technologist, or entrepreneur, there is a place for you in the ComplexChaos community.

Together, we can leverage the power of interdisciplinary cooperation to tackle the world's most pressing challenges and create a better future for all.