+2

Artificial Intelligence Models.md

reviewed

··· 5 5 - Classic ML system where humans are designing how the information is organized (feature engineering, linking, graph building) scale poorly ([the bitter lesson](http://www.incompleteideas.net/IncIdeas/BitterLesson.html)). LLMs are able to learn how to organize the information from the data itself. 6 6 - [LLMs may not yet have human-level depth, but they already have vastly superhuman breadth](https://news.ycombinator.com/item?id=42625851). 7 7 - Learning to prompt is similar to learning to search in a search engine (you have to develop a sense of how and what to search for). 8 8 + - LLMs have encyclopedic knowledge but suffer from hallucinations, jagged intelligence, and "amnesia" (no persistent memory). 8 9 - AI tools amplify existing expertise. The more skills and experience you have on a topic, the faster and better the results you can get from working with LLMs on that topic. 9 10 - [LLMs are useful when exploiting the asymmetry between coming up with an answer and verifying the answer](https://vitalik.eth.limo/general/2025/02/28/aihumans.html) (similar to how a sudoku is difficult to solve, but it's easy to verify that a solution is correct). 10 11 - [Software 2.0 automates what we can verify](https://x.com/karpathy/status/1990116666194456651). If a task/job is verifiable, then it is optimizable directly or via reinforcement learning, and a neural net can be trained to work extremely well. ··· 92 93 - The ability to ship to a preview environment. 93 94 - An instinct for what can be outsourced (to AI vs what needs human attention). 94 95 - An good (updated) sense of estimation. 96 96 + - Build "partial autonomy" products where humans stay in the loop to verify output, rather than fully autonomous agents. 95 97 96 98 ## Agents 97 99

+12

Public Goods Funding.md

reviewed

··· 22 22 - **Anti-Collusion Infrastructure**. Like secure voting systems, there is a threat of buying votes in a funding mechanism. Collusion can be discouraged by making it impossible for users to prove how they reported their preferences. This infrastructure must be extended to prevent collusion between the 3rd party and the users. 23 23 - **Predictable Schedules**. Participants need to know when are they getting funded. 24 24 25 25 + ## Methods 26 26 + 27 27 + ### Simulation Process (S-process) 28 28 + 29 29 + The [S-Process (Simulation Process)](https://www.youtube.com/watch?v=jWivz6KidkI) is a collaborative funding algorithm designed to optimize the distribution of resources to public goods. It allows multiple funders to delegate the complexity of grant-making to overlapping groups of trusted "recommenders". 30 30 + 31 31 + - **Aggregate Information, Not Just Money**. Instead of funders voting with dollars, participants input **Marginal Value Functions (MVFs)**. This creates a "How valuable is the next dollar given to this organization?" curve for each organization. 32 32 + - **Delegation to Trusted Recommenders:** Funders often lack the time to evaluate every opportunity, the system allows them to delegate the creation of these value curves to trusted experts. Funders still have the "Final Say". This relieves stress on the advisors (recommenders), allowing them to express honest opinions without the anxiety of being the sole decision-maker. 33 33 + - **Discourse is Essential:** Numbers cannot replace conversation. The process requires real-time debate where advisors explain *why* they value an organization differently. 34 34 + 1. **Iterated Simulations:** The allocation algorithm is run repeatedly *during* the discussion. Advisors see where the money would go based on their current inputs, discuss the outcome, and adjust their inputs. This turns a "one-shot" game into an iterated cooperative game. 35 35 + 2. **The "Disagreement Matrix":** The system highlights where advisors disagree most (e.g., Advisor A loves Org X, Advisor B hates Org X). Discussion is focused specifically on these disagreements to surface new information. 36 36 + 25 37 ## Resources 26 38 27 39 - [Public Goods Funding Landscape](https://splittinginfinity.substack.com/p/the-public-goods-funding-landscape)