TL;DR: Should I work on modeling rapid AI progress impact on non-AI fields. I am trying to decide if I should work on a project I created called Accelerated Horizons. It would be a podcast + blog exploring how rapid AI progress could reshape specific industries and social systems by 2030.
AI-generated religions, synthetic art movements, novel material creation, defense automation, and state-sponsored hyper-targeted propaganda. I want more smart people, outside of EA-adjacent circles, to take superintelligent AI seriously — and to seriously consider what society could look like in a world of abundant, cheap cognition.
Broad Concept
- Interviews with scientists, technologists, academics, religious clergy, industry experts — basically anyone with an expert view on a domain where a sudden abundance of intelligence could drive rapid transformation.
- Core question: “What happens to your field if AI advances 10–1000x faster than expected?” I’d need to explain why this is plausible and help guests think through implications. One framing might be: “If you had 10,000 fully-funded MIT PhDs working remotely under your direction, what could you do?”
- The goal is to ground AI discourse in real-world systems, constraints, and leverage points — and then convene thoughtful people around how we might prepare.
Why This?
- Most AI discussion is either very abstract or focused on technical alignment. That creates hand-waving when it comes to real-world implications.
- We need more domain-specific imagination and forecasting.
- This could help:
- Surface blind spots and failure modes in key systems
- Attract broader talent and insight into x-risk and AI governance conversations
- Seed more grounded policy conversations across domains
- Support prioritization and resourcing in neglected areas
Sample Topics
There’s already a deep bench of high-leverage domains to explore. Just interviewing contributors to the Gap Map project (https://www.gap-map.org/?sort=rank&fields=materials-science) could fill a year of content.
Some example themes:
- AI-generated theology and decentralized spiritual movements
- State use of AI for narrative warfare and information control
- Pharma and biotech acceleration via closed-loop autonomous R&D
- Fragility in legal systems, financial markets, and global supply chains
- How sectors might react to the emergence of superhuman agency
What I Need
- Feedback: Is this useful? What would make it more so? Is this worth my time?
- Any advice on starting a podcast and doing high-quality outreach
- Guest suggestions — especially from surprising or underexplored fields
- Interview questions you’d want answered
- Pointers to similar or adjacent work I should study or collaborate with
- Do I need to work with someone on a project like this?
I like it! Seems somewhat neglected. Econ researchers have worked on modeling broad economic impacts, but I haven't seen as much done that tries to model specific impacts on specific jobs. It would be especially helpful if you're able to extrapolate benchmark results to job automation. Like finding some data on SWE work that's been automated in the past few years and correlate it to coding benchmark performance, then see if that approach can generalize to other professions.
Thats a good thought. I also have been playing with using aggregated research bottlenecks like gap-map.com and automated research to identify progress towards or completion of those tasks.