OPEN INQUIRIES
A breathing list of open inquiries
If you’re interested in digging deeper into any of the below topics, feel free to shoot me a note.
I’m almost always hiring or looking for collaborators. The below are the topics that I’m actively exploring at work.
Limits of my language. Limits of my world.
In Tractatus, Wittgenstein stated: “what can be said at all can be said clearly, and what we cannot talk about we must pass over in silence.” However, our experiences extend beyond mere words. The concept of a multimodal vector space illustrates how language can be an abstraction of multiple sensations. The boundary of my world is likely a blend of all sensory data and the intricate workings of our physiological mechanisms, which can be somewhat mapped into the same “language” space. This fundamentally alters the extent to which we can understand ourselves and the world around us, where symbols and meanings become more interconnected and interoperable. There has never been a better time to cultivate greater self-awareness.
The freedom quotient: free will in the modern world
Stephen Cave formalized “Freedom Quotient” as an additional dimension to evaluate human competencies beyond IQs (intelligent quotient) and EQ (emotional quotient). Cave defined FQ as a measure of natural free will, and he argues that smoothing out inequalities in FQ – and to raise the FQ of everyone – should be a goal of schools and of social policy. The call for redistributing FQ may just be the far cry of the century. To what extent do our increasingly omnipotent machines enhance and deprive us of our agency? More importantly, what are the principles to consider when building agency-enhancing machines?
The politics of centralized and decentralized technologies
All political struggles will become technological struggles. At the center of the technological struggle lies the fight between decentralization and centralization, the crowd and the state, the tangible and the intangible assets, and individualism and collectivism. These struggles will take place at an accelerating pace. Understanding and engaging with the technological struggle is essential to shape a more equitable, just, and democratic political future. Decentralization needs to be protected to preserve the randomness, diversity, and vivacity of human creativity and development, both of which are essential to geopolitical competitiveness.
Self-sovereign digital commodities
At the intersection of AI and crypto technologies is a highly sought-after resource: access to GPUs. Bitcoin and Ethereum display sophisticated incentive design that aligns the interests of validators, miners, and consumers without needing a centralized third party. Moreover, self-custody wallets in crypto have taught tens of millions of Internet users about consuming resources by spending "gas" and signing transactions with a self-managed private key. But what if consumers had more control over their digital assets, allowing them to access scarce computing resources to enhance their personal finances, businesses, and lives?
New GUIs for LLM-centric experiences
The last decade of software design has focused on minimizing friction in data entry and orchestration. The goal of deterministic software architecture is to narrow the gap between "what you see" and "what you get". ChatGPT has opened up new possibilities for a chat-first interface, offering a more fluid way to collect, interpret, and process input. With the advancement of multimodal learning and the prevalence of computing devices, we are closer than ever to digital synesthesia, a fertile ground for exploring new Graphical User Interfaces (GUIs), including Nodes and Canvases.
Remystifying the world: meaning-finding at the end of rationality
Religion operates in an enchanted world, featuring spirits, myths, magical or providential events, and external interventions. Conversely, science aims to disenchant this world, replacing these elements with objective factors like logic, reason, and autonomy. So, do these processes occur in cycles? What triggers the ascent and descent of enchantment and science? Where do we stand in this cycle currently? And how can these two forces collaborate more effectively?
Cryptoeconomics as playground for political economy
The speed with which markets get created and collapsed makes crypto an interesting case study for understanding market mechanism at its rawest. Rather than criticizing the “ponzieconomics” as pure scam, are there lessons we can learned when we expose the interior of capitalism? Bottoms-up NFTs communities and decentralized finance protocol DAOs are a few places where we see glimpses of possibilities of highly incentivized local anarchy facilitated by shared containers of value encoded with smart contracts.
Fiction machines: tools for world-building
Just like Borges imagined the Library of Babel with all kinds of books, we can think of LLMs as a collection of interesting texts. When we use an LLM, every new piece of text we put in changes the story and opens up new possibilities. But just like Borges' Librarians, who walk around the endless Library of Babel, we should remember that LLMs aren't always going to give us the right answers. They can sometimes say things we don't want them to, and that can be hard to control. But, just like a good story can teach us something real, LLMs can help us think about things in new ways and learn from them. They can help us be more creative and understand things better if we use them the right way. We believe it's important to remember that what LLMs make up isn't the same as what more factual AI systems can tell us.
New financing models for software companies
Venture capital, the primary funding for software companies over the past decade, is now rivaled by Micro SaaS acquirers. Although venture capital can lead to great outcomes, it often mismatches with the nature of certain types of businesses, leading to founder burnout. Building software can be intrinsically rewarding, but the venture-scale lifestyle isn't for everyone. Companies like Constellation Software and Tiny’s propose alternative funding models, recognizing that most software companies are mispriced due to private market involution. This suggests unexplored funding models.
Attention-is-all-you-need
Attention is becoming more important in a few context. Attention is becoming one of the most scarce intangible asset in the digital economy. In the context of ML, attention is a mechanism that allows models to dynamically focus on specific parts of the input data, assigning varying levels of importance to different elements. Although attention carry significantly different meanings in the two context, meditating on how the two context can inform one another may be an edifying exercise. How can machines gage the significance of a particular context and “price” it as an asset accurately?
Side quests
These are long-term personal interests that I explore privately. I’m not working on these problems on a professional level, but I’m always seeking collaborations and learnings with experts in these fields.
︎︎︎ Continuation work on Jung and Myers Briggs for self-understanding and development
︎︎︎ Mediated and moderated psychedelics for mind expansion and healing
︎︎︎Alternative self-directed education models enabled by AI