Student
Plans
Progress
Problems
Nine Inch Иails
- Prepare slides for an intro talk about compute verification.
- Test grasp of content with Anki flashcards.
- Deck created.
- Each session has a document tab for its glossary– cards could be created from these.
- Talk delivery still pending, further iteration needed.
- Exercises for concept handles (e.g Dresher, Fearon, Putnam, Gintis, Binmore, Skyrms, Farell & Newman) can be confusing.
Weezer[4]
- Have a broader view of field's demands.
- Start project on Cryptographic AI verification.
- Discussed costs, timelines, and implementation of verification mechanisms.
- They are exploring projects related to cryptographic verification of AI computation, but started two days before the intensive!
- There are ~60 engineers working on verification, "the field" is pre-paradigmatic.
- This wasn't a focus in the intensive, but Lens will be running a Part 2 which will go further into technical components.
Portishead
- Make tool or product to help policymakers.
- Learned about geographical supply-chain bottlenecks.
- Underspecified.
DJ Shadow
- Finish intensive.
- Connect finance or cybersecurity work to material.
- Done
- Limited evidence of progress.
- DJ Shadow may have done all the readings and DJ Shadow may be able to correctly answer questions about them, but completing the week isn't strong evidence of that.
- The example DJ Shadow cites as being applicable to DJ Shadow's work does not seem to appear in the material. [5]
Tricky
- Help with PauseAI US Advocacy.
- Contribute to a technical verification project.
- Developed questions about practical feasibility and governance.
- Did some learning, no contribution yet.
- PauseAI US is controversial.
- Again, technical projects weren't a thing in Part 1. But there's always Part 2![6]
Afrika Bambaataa
- Develop project proposal.
- Prepare for hackathon [7] research sprint.
- Discussed cyber risks, HW/SW combinations, government funding.
- Did FFT on GPU
- No proposal recorded.
- On a smaller dataset where I suspect CPU performance would be comparable. [8]
Massive Attack
- Finish course
- Conduct additional research.
- Articulate views on Substack and in job interviews.
- Land a role.
- Support progress towards a multilateral instrument.
- Mitigate catastrophic risks from AI at the global level.
- Done.
- Done.
- Done.
- Nearly done.
- *single tear rolls down one eye* ... so cracked.
- cracked, fissured, and groovy.
- Will not accept an endorsement on LinkedIn 😔.
Björk
No original goal documented.
Got a good baseline, alongside a better understanding of hardware verification's weaknesses.