Presenter
Amanda Ngo, Ought
Hi there, I’m Amanda! I’m a Kiwi living in the United States. I work at Ought, a startup building tools to automate complex reasoning. Check out our forecasting product, Elicit! I love co-opting friends into social experiments, understanding how brains work, and figuring out what it means to live a good life.
Shaun Conway, iXO
Inventor of the ixo protocol for identifying, verifying and valuing impact. Shaun is a clinically experienced physician, international development specialist and visionary leader at the forefront of solving complex social challenges, through systems-thinking and technological innovation. He has a passion for designing and building ethical, data-driven technologies and systems that can positively transform human experience and improve well-being for millions of people....
Isabella Duan, University of Chicago
Ethics & Policy Intern @DeepMind; Computational Social Science @ UChicago
Kevin Compher, IN-Q-TEL
I am a seasoned manager, engineer and data scientist with over twenty years of strategic and tactical analytic experience in the applied research, development and scaling of streaming, platforms and data hungry applications. I have demonstrated achievements in the areas of: ...
Summary:
What are you trying to do?
We want to help decision makers to make effective choices that are transparent and easy to understand.
How is it done today? What are the limitations of the current system?
Opaque (black box group-decisions) context around allocation of resources and program design to achieve measurable outcomes. Limited transparency create difficult to troubleshoot steps, thus hindering discovery of optimal paths.
What is new in your approach and why do you think it will be successful?
Our approach uses large language models semantically to recommend process to optimize outputs.
If successful, what difference will it make?
Find, evaluate, optimize best processes across nearly any outcome, for instance creating better jobs/increase income. The public or other stakeholders that can hold decision-makers accountable.
How much will it cost?
The estimated cost is roughly $10k.
How long will it take?
It could be very soon. We may need GPT4 or 5 to get it to the point of excellent proposals.
What are the mid-term and final exams to check for completeness?
Mid-term exams: Create a generalized/abstracted system that works across a diverse set of outcomes.
Final exams: Evaluation and feedback of the model against tested proposals.
Who is interested in exploring the working project further?
Amanda: I’m interested in helping people build a prototype ([email protected])
Kevin: Follow-up with team, gauge interest in collaborating on prototype. ([email protected])
