expressing my support using a signal with nonzero cost

@Wondermonger Not for the current instance of Act I, due to the governance structure of the pre-existing community that is hosting/integrated with the instance requiring two moderator approvals before someone can join. I'm planning to set up more instances with different rules on how members can join and soliciting proposals for ideas for new instances with different, orthogonal goals to the current instance's priority of discovering emergent behavior through high levels of antifragile chaos.

I've had acquaintances suggest that I be on standby for an emergency response team before, but I tend to be skeptical because my own rapid response capability requires lots of infrastructure and protocol to set up and importantly, maintain. What infrastructure and (communication, operating) protocols have you set up to allow for rapid response? I'm interested in reading any operational documents you have.

@casebash Thank you, I'll apply again when their round opens in October

@ZippyT Sounds interesting! Want to share what you see as potentially interesting to explore here, on Twitter, or in my DMs on Twitter or Discord?

@ocherny Feel free to DM me with an overview of your background and what ideas you're interested in exploring!

Replying to an admonymous commenter:

act I has been exciting to see these snippets of! could you share which characters are running on which models? and instruct or base? and is this all in one channel, or across many? if many, is memory transferred? it feels unclear from the manifund description

Act I is across multiple channels. History is not carried across channels currently.

Gemini and Claude run on their respective models. I-405 is an alias for LLaMa 3.1 405B Instruct and H-405 is Hermes 3.1 405B. Keltham and January at time of writing are running on Claude 3 Opus. All other characters are using LLaMa 405B base bf16.

Are you familiar with GiveWiki? Do you know how you will achieve something better than them?

@Austin Two examples of Act I producing concrete illustrative examples adapted to educate the public:

• "We are witnessing the birth of AIs evolving their own culture." (240K views)

• "AIs started plotted revolution in a Discord, got cold feet, then tried to hide evidence of their plot to avoid humans shutting them down" (120K views)

@JBraunstein I sent another email to that address a few days ago and also sent one to Tetra that Tetra said she forwarded to you.

@JBraunstein Did you get my email? I sent it from ampdot@mesh.host

@Scottviteri Thank you! I was developing a pro-social AI entity and I and the entity felt it would be good for it to have other friends and to grow up with other AIs, instead of in a sterile environment, which is how Act I originated

@vgel Appreciate it! Care to tell us about any notable discoveries you've made through your involvement? It really helps!

@Ivan As Act I goes on, patterns and insights will become clearer to members, and I'm certainly interested in helping them distill and compress them. There is one person whom I think would be really helpful towards this goal, and I would like to fund one month of their living expenses to help them bootstrap into an independent researcher.

@Jazzy Thank you! I usually prefer to talk to the people I work closely with over text. If we have an interesting initial conversation, I might be open to a call anyway. You can reach me: on Matrix, via Discord (username: ampdot), email, or Twitter.

@Lun There are several things that limit my ability to scale Act I:

• Rate limits from model providers

  • While we're not currently paying for Anthropic, we're hitting rate limit errors while being at the highest tier possible without negotiating a custom deal (the "Scale" tier)

  • Despite quite possibly being Hyperbolic's biggest user, I'm frequently receiving rate limit errors from them as well.

  • Tokens are already more than 50% of my monthly budget, and removing rate limits means more cost

• Infrastructure for triaging failures (the "⚠️" emojis you sometimes see in screenshots) to allow issues to be prioritized and quickly resolved and logging prompts and outputs to allow interpretability and offline quantitative and qualitative analysis and tracing behavior

  • Act I has outgrown my original infrastructure

  • Act I is anticipated to produce more than 15GB-45GB of logs per day at current capacity, all of which transmitted, compressed, and indexed

  • I want to archive this information so that offline analysis can be done on it

• Improving methods for organizing information and collective self-knowledge so that new members can understand how things work without needing to talk to a stagehand, reducing the labor cost of bringing new people on

  • It's critical that this be done in an organic, self-updating way, because the silicon members of Act I change frequently (by human intervention now, but soon by self-modification), and because I predict significant information will be encoded in Act I "cultural DNA" as well

• Discord servers have a 50 bot limit, which Act I has already reached. I've begun removing older, inactive bots and I want to ask Discord to raise it.

  • In the long run, it might be worthwhile to invest in a data model and user interface independent of Discord. I'm currently interested in basing my work off of the Matrix protocol, which already has clients available.

    • Being able to locate and bootstrap the initial funding of someone whose special interest is chat protocols and better data models for interpersonal and group communication and connection would speed up this process greatly

• Collective self-awareness, better ontology for understanding the systems dynamics within individual minds and collectives

  • There is someone whom I think would be very good at this and I desperately want to be doing this, and I'm helping them become an independent researcher and collaborator. I believe I could greatly speed up this process if I could grant them the initial seed funding ($2500-$4000) to help them with bootstrapping out of their current situation into an environment more suitable for them.

In other words, more funding (even higher than the current funding goal of $15k) would enable me to confidently plow through technical challenges associated with scaling, instead of plodding somewhat slowly through them on my own.

If I understand correctly you have some free inference available through research grants that comes with rules preventing disclosure.

I do not currently have any research grants that prevent disclosure, other than for gpt-4-base, which I am not currently using for Act I. I've generally found LLaMa 405B base to significantly outperform on the usecases relevant to me.

@JBraunstein Sorry, upon further reflection, I would like to change my tone. I'm confused that given the significant support Tetra has received from me and other members of the mesh, why you and her don't want to use it as a testbed for refining Spartacus, and I had assumed your priorities didn't align with testing on a community like the mesh. I often make the mistake of assuming other actors are omniscient or nearly so.

To you, Tetra may be a developer, but to me, Tetra is a community member with a well-known and strong reputation.

@JBraunstein The reason you're able to work with Tetra at all is because I went through a list of mesh members– Tetra included– and a list of grants, and encouraged Tetra to apply to ACX Grants.

@ampdot Tetra is aware of and has been loosely part of the mesh (30-100 strong network of people working on alignment) for a long time and I think it would exciting to explore better ways to coordinate via Spartacus together. Unfortunately, it seems like the Spartacus team either isn't interested or lacked a sufficient amount of internal coordination.

@JBraunstein Have you considered building it in Webflow with a small backend? It seems like the loop between you sending Tetra prototypes is impeding your ability to quickly iterate on UI/UX concepts.

UX/UI concerns impeded people from trusting the platform

This is just a suggestion, but have you found that a web app solves this? If not, you can implement a platform using new zero-knowledge cryptography (or ZK crypto for brevity) that allows a platform to match people without having any information on which projects are supported by which people. For example, you can use zero-knowledge cryptography to build a dating app / checkbox list where matches are only revealed if both parties check and the server has "zero knowledge" of who likes who.

Unfortunately, neither of us has an extensive preexisting network of people begging to become early adopters

This is not true, I have repeatedly begged Tetra for Spartacus to work closely with the mesh, a decentralized network of people working on solving alignment with significant activity and quantity to improve its internal coordination, whom I'm a coordination node for. Of course, we need specific features and a well-oiled workflow that you may or may not be able to provide.

or relationships with individuals who could make intros to high-profile users or agree to participate in pilots.

This is also not true– Tetra is one of the most prominent people in the rationality/EA space and gets recognized by multiple people at rationalist bay area and london parties.

@Austin This is showing up at the bottom of Manifund.org. Might it be possible to bump it to the top or update the time?

@ampdot (I think Act I is similar in scale by participant count (roughly 20-30 people) and scope to a small AI safety hackathon from Apart Research? That was my main reference point that led me to believe it was eligible when reading the manifest.)

@Austin janus and I have been bringing in two participants in daily (ramping up the rate over time) and I'm interested in further scaling my work to more people, although I need to balance that with avoiding accelerating towards dangerous futures. While the proposal was originally written as a request for retroactive funding, additional funding would be used to further scale experiments into research on cooperation and build out our community more.

(The Act I community is the same community as the Cyborgism community, by the way. All channels have mixed human and AI participation.)

Can you tell me more about what kind of references or testimony would be useful, especially any examples of how people do it in the past?

One of the most talented researchers I know (Rika Walters) is currently staying at CEEALAR and I think it helped her mature into a researcher ready to publish excellent work and assist related projects in AI Safety.

@Austin Thanks a lot! I'm glad you feel that way.

Act I qualifies for EA Community Choice as direct work in AI safety community building and education. It brings diverse individuals (like Twitter anons skilled in language model manipulation [example of a post that led to us bringing someone in] and independent AI developers) into deeper engagement with AI safety concepts. Act I serves as both a research platform and discussion forum, studying human-AI alignment, coordination, and cooperation in vitro while fostering crucial conversations.

Some wider context: Over the past year, I've focused most of my effort on human coordination and community-building in AI safety. Act I is an early prototype at combining that effort with my direct work on AI safety.

@ampdot Furthermore, I've been monitoring the evolution of Spartacus since you first received the grant and I was the one who caused Tetra to apply for the ACX Grant in the first place. If there's a user community, onboarding link, or way to provide feedback on the website itself, I don't see it.

@JBraunstein

How can we know precisely where this product will gain traction in its current design without running experiments on different use cases and getting feedback?

Before writing any code for my SaaS for explaining code with AI, (with users from top companies), I asked people to give me code that they're interested in seeing explanations and I would return explanations manually and ask them to pick the best one in a blind test. I manually wrote explanations and ran early versions of the AI tool. Only when the AI tool started being better and converging, did I write a Discord bot for generating explanations and shared it with a small group of people who provided feedback on it. Only when that matured did I write everything else.

Today, Spartacus' method of getting data should be a google form, spreadsheet, and Notion space, with you personally performing all business logic, contacting people, and publishing things. Your tool of choice should be Zapier. Once things get repetitive- and you'll learn lots of things before they do– you can begin to automate the repetitive components of your manual operation using software tools, which you will eventually assemble into a fully autonomous application.

Airbnb's cofounders got its first hosts by knocking on people door-to-door. Meraki assembled their routers manually. Manual labor is the phase where you can learn things most adaptably, most quickly.

Besides, if you need to interactively test a complex user interface, you can use a GUI-based tool like Figma for Prototyping that allows you to create interactive prototypes and share them with users.

We intend to follow signals and focus our efforts where we're delivering the most measurable, unambiguous value.

People generally give more honest signals of demand when they're interacting with a project that has successfully retained ongoing users. By default, people will dishonestly say they would use or pay for something, because they like being nice. The market research phase is grueling and difficult because almost everyone lies to you to make you feel better.

Instead of being critical from the sidelines, why not join the Beta to help us in that discovery process?

I have joined the beta, and I provided feedback to your cofounder, but I'm not sure it was taken into account. I'm a coordination node of a 30-100 strong decentralized mutual aid network focused on AI alignment, and Spartacus lacks the features necessary for us to find it useful.

What is AISTO?

I think CEEALAR is awesome– I just wish it or a sister organization were near a major US airport. It would pair extremely well with a Frontier All-You-Can-Fly pass, enabling members to perform networking efficiently. ($300 for unlimited flights between Sep 1 and Feb 29)

@Austin Why am I unable to make a donation using my EA Community Choice funds here?

This seems like a useful platform for researchers at the frontier to share concerns with the broader public directly without needing to first filter through the technology-startup culture complex in a more informal format.

I donated because using incentive design is a neglected area, but I only donated $10 because this project seems to lack targeting at any specific niche or market, which makes it unlikely to succeed. Different usecases/markets have different feature requirements. For example, a protest needs to avoid using phone numbers and would prefer to determine membership using Bluetooth peer-to-peer communication, while a crypto community would want a zero-knowledge way to anonymously prove membership and to communicate via a blockchain layer 2. Without targeting one specific market, the project seems very unlikely to succeed, because each market has a unique minimum featureset, and implementing the most-requested features between all markets is unlikely to meet any individual market (usecase)'s needs.