Semester 1

Preparation (20/5/2024 - 24/5/2024) Since the hackathon starts two weeks off the foot of the previous jam hosted by Apart, most server members did not anticipate intensive participation. Nevertheless, Qronox did come up with the idea to fetch images from around the world and then feed that either into a special use AI model (like GeoSpy.ai) or into a multimodal model (like GPT-4o) to see how well the model infers user location from these images....

Individual reading 12:30 - 14:45: Mogu, Jord, and Gumperto get started by doing some individual reading before everyone arrives. Jord and Mogu mostly worked on group administration, while Gumperto tried going through the early lessons in the ARENA course. Rationality group reading 14:45 - 15:55: Qronox and Allen Dulles arrived at the venue around this time, which prompted us to start the group reading related to scope sensitivity and Moloch. We debated with each other on what we should do to be more resistant to scope sensitivity and what ways in which we can address the various Moloch-run systems that plague our world....

Agent foundations We found this textbook— AISF Alignment Express —which seems quite useful for a more structured and centralized way of getting into introductory AI Safety than the AI Safety Fundamentals Course. This is the resource we used for the week to learn about agent foundations. We spent around 30 minutes reading through Chapter 9: Agent Foundations and discussed some basic concepts and definitions. After that we listened through a talk delivered by Eliezer Yudkowsky, called AI Alignment: Why It’s Hard, and Where to Start for a more gentle introduction in agent foundations, as well as to reinforce some fundamental ideas about AI alignment as a whole....

Hackathon description 🌏 Join us to evaluate and mitigate societal challenges from AI Despite the many potential benefits of the technology (such as equalizing opportunity, creating wealth, and improving coordination), we are also facing significant risks from AI. Together, we will be hacking away to demonstrate and mitigate the challenges that arise in AI, while trying to project these risks into the future. Summary Sadly, the majority of group members were unavailable for the whole weekend due to exams or other personal reasons....

Brainstorming Ideas We read a brief post from and tried to come up with potential ideas to work on for Apart Research’s upcoming research hackathon from May 3rd to May 5th. Here’s some ideas we came up with and discussed: Kappie & Welsh: Conduct an opinion survey on some demographic of local law students on their use of AI (specifically LLMs) in their academic work, and then write a report based on their findings....

New member initiation We initiated our newer members - Welsh and Kappie - on the basics of AI and AI alignment. Self-research We spent the first half of the meeting (2pm - 3:45pm) researching and taking notes on topics that we found interesting. Generally, the group was split into two streams, as per our established activity pipeline: Stream 1 and Stream 2. Here’s a list of what each of us did research on:...

We spent an evening thinking about some novel/neglected research directions and making proposals based on what we think is a promising research topic. Some of these topics (which may still be a bit rough around the edges!) include: Can multi-agent system LLMs model each other accurately with just text interactions? Can LLMs recognise a user based solely on their text responses? Can LLMs find novel loopholes in certain legislations? Can we make an LLM derive a nonsensical conclusion through a connection that is semantically similar to some other reasonable logical connection?...

Restructuring We reworked the activity pipeline for our weekly meetings, which goes as follows: Self-research: Spend up to 1 ~ 1.5 hours researching the topics/working on the problems you want to discuss that day and then answer some related questions. Discussion: Spend 15 ~ 20 minutes discussing what you just researched with someone who also researched the same thing/already has some foundations in it. Presentation: Spend ~10 minutes presenting your findings to everyone....

New members We spent some time introducing ourselves to the new members and letting the new members introduce themselves to us. We talked to each other about the basic argument of AI alignment and why it’s important We watched Robert Miles’ “Intro to AI Safety, Remastered” and “10 Reasons to Ignore AI Safety” x-risk Debate We debated with each other on whether the most important problems of AI Safety is short-medium-term (economic, sociopolitical) or long-term (irreversible turning point in human history), whether that problem would be infrastructural or existential....

Jord&qronox Hanging up posters to invite non group members to join, though no one did. Preparation (22/3/2024 - 23/3/2024) Jord, Mogu, Qronox, Skrubz worked on the initial phase of coming up with and submitting ideas Jam Time! (23/3/2024 - 24/4/2024) 23/3/2024 On the evening of 23/3/2024, Jord stayed over at Mogu’s for the night apartment to cowork on the specifications (and submit new ideas). Specifically, they worked on specifications for these ideas: Open Grants Glitch Tokens Others?...