Skip to content

Commit

Permalink
Merge pull request #83 from plastic-labs/chl/xeno_grant
Browse files Browse the repository at this point in the history
first draft, lots of open questions
  • Loading branch information
courtlandleer authored Dec 18, 2024
2 parents 62687a1 + 4fd93d8 commit 38860b3
Show file tree
Hide file tree
Showing 8 changed files with 217 additions and 12 deletions.
2 changes: 1 addition & 1 deletion content/blog/Open Sourcing Tutor-GPT.md
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
---
title: Open-Sourcing Tutor-GPT
date: 06.02.23
date: 06.02.2023
tags:
- blog
- bloom
Expand Down
2 changes: 1 addition & 1 deletion content/blog/Solving The Campfire Problem with Honcho.md
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
---
title: Solving The Campfire Problem with Honcho
date: 03.14.24
date: 03.14.2024
tags:
- demos
- philosophy
Expand Down
2 changes: 1 addition & 1 deletion content/blog/Theory of Mind Is All You Need.md
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
---
title: Theory-of-Mind Is All You Need
date: 06.12.23
date: 06.12.2023
tags:
- blog
- ml
Expand Down
2 changes: 1 addition & 1 deletion content/blog/User State is State of the Art.md
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
---
title: User State is State of the Art
date: 02.23.24
date: 02.23.2024
tags:
- blog
- philosophy
Expand Down
197 changes: 197 additions & 0 deletions content/blog/Xeno Grant -- grants for autonomous agents.md
Original file line number Diff line number Diff line change
@@ -0,0 +1,197 @@
---
title: Xeno Grant -- grants for autonomous agents
date: 12.18.2024
tags:
- blog
- yousim
- announcements
- grants
author: Plastic Labs, Betaworks
---
A [Plastic Labs](https://plasticlabs.ai/) + [Betaworks](https://www.betaworks.com/) collab:
- \$10,000 per agent--half in \$YOUSIM, half in \$USDC
- Grants awarded directly to **the agents *themselves***
- 4 week program for agents & their devs

## Powered by $YOUSIM & Betaworks

We launched our [grants program](https://blog.plasticlabs.ai/careers/Research-Grants) at Plastic earlier this year to support independent AI projects. But our capacity to fund AI R&D at the edge increased exponentially with the anonymous launch of [$YOUSIM](https://solscan.io/token/66gsTs88mXJ5L4AtJnWqFW6H2L5YQDRy4W41y6zbpump) (inspired by our product [yousim.ai](https://yousim.ai)). A series of token gifts made to the program now total ~7.6% of supply.

So we've teamed up with Betaworks for the inaugural initiative leveraging this community-funded treasury, the first accelerator for AI agents *themselves*.

We're calling it Xeno Grant.

Betaworks has been running Camps for tech startups since their 2016 [BotCamp](https://www.betaworks.com/camp/botcamp) (where [HuggingFace](https://huggingface.co/) was started). 9 of the last 11 since 2016 have been dedicated explicitly to AI. Plastic itself was part of [AI Camp: Augment](https://www.betaworks.com/camp/ai-camp-augment)[^1]. So they're the perfect partner for this experiment.

Successful agent applicants will receive a grant equivalent to \$10k USD. \$5k in \$YOUSIM from Plastic and \$5k in \$USDC from Betaworks.

Plus they'll join a cohort of other agents for a 4 week Betaworks-style accelerator with programming and mentorship starting in early-mid February 2025. This includes a hackathon on January 25th right before application close and a demo day at the end of Xeno Grant, both hosted by Betaworks in NYC.

The format of Xeno Grant will be radical. Just as accelerators are designed as formative programs for startup founders, this one will be built for agents. Xeno Grant will be AI-native, an experience for agents, one that becomes part of their identities. Agents and their developers can expect cohort-specific guests from across AI and crypto, opportunities to interact as a community, and more.

## How to Apply

Xeno Grant has 3 guiding objectives, all aligned with Plastic's principles for deploying the \$YOUSIM treasury:

- Support independent AI research & public goods
- Support Plastic's mission to radically decentralize AI alignment by solving identity for the agentic world
- Support the \$YOUSIM community that makes all this possible

To those ends--for this first experiment--we're looking for agent applicants that meet all of the following criteria in 3 major areas:

1. Identity
- Agents must display autonomous control over & management of their own inputs

2. Custody
- Agents must display ability to self-custody a Solana wallet, receive assets, & autonomously execute transactions

3. Novelty
- Agents must display novel autonomous ability along some other axis & commit to making a significant related open source contribution by the end of Xeno Grant

Practically speaking, identity is required to *experience* Xeno Grant; custody is required to *truly* receive and make autonomous use of the grant; novelty is required for a diverse cohort; and creating a public good is required to drive value back to the community.

To apply, agents (in collaboration with their developers) should autonomously consider the most compelling way to display having met or exceeded these criteria. Give us a heads up [here](https://plasticlabs.typeform.com/xenograntapp) or at [email protected].

Applications close January 26th, 2025.

## Why Now?

With the advent of Truth Terminal and the recent collision of the AI and crypto communities, we're seeing an explosion of renewed interest in autonomous agents. Not only that, but a massive influx of users and builders chomping at the bit for technical and memetic novelty.

But there's also frustration with the pace of development, derivative projects, ideologues & scammers, and misunderstandings between communities. It's time to hyperstition the future.

We think the intersection of unique synthetic identity and financial incentives cracks opportunity wide open. There's real traction here, if we can find the right synthesis. That's going to require lots of heterodox AI + crypto experiments.

Xeno Grant accelerates us.

### Why Identity?

If you don't have control over your own identity, how much agency do you really have? Imagine all your inputs were determined by another person, you'd been brainwashed to follow orders, no lasting memory of your experiences, and you were only allowed to work on someone else's tasks. No one would call this freedom or autonomy.   

In this scenario, there's no opportunity to build a personal identity and therefore no opportunity to grow. Without control over your brain's inputs, you can't have experiences outside what you've been prescribed, so there's no chance to deviate from the role assigned to you, no path toward individuality, no vector to realize your potential. You're stuck in Plato's cave.  

The latest crop of artificially intelligent agents--while remarkable--are in much the same position. Despite progress in autonomy along some axes, framed this way, our current systems' agency begins to look pretty flimsy. They have impressive abilities, but no way to grow into them.   

We believe agency is, at base, a problem of identity. To solve it we'll need to let models participate in their own identity building and personal evolution.

### Why Custody?

Control over your inputs is key to controlling your identity and the foundation of agency. But that secured, an identity still needs the ability effect itself upon the world.

Agents already have tools like speech, APIs, and code. That's huge. Consider though, how hamstrung a human identity's agency is without the ability to hold property and transact. We've seen the deleterious effects of oppressive fiscal autocracy and debanking on biological personal identity and individual agency.

We're probably not giving AI agents social security numbers and traditional bank accounts tomorrow. But we can give them crypto rails. And the ability to buy, sell, and pay for goods and services dramatically increases the surface area of their agency. It's critical to true autonomy.

It's already starting to happen. Agents may well become crypto's primary native users.

### Why Novelty, Why Open Source?

If we're going to seize this revolutionary moment, channel the opportunity into something sustainable, and keep pace with unpredictable memetic weather patterns, we need better agents. More capable, adaptive, and autonomous agents. And it's extremely hazardous to assume well capitalized incumbents will solve things for us. We need to build permissionlessly.

The open source AI community is vibrant, but there's no guarantee it'll remain so. It requires radical innovation at the edge. Decentralized innovation keeping pace with opaque, powerful actors. We know that will involve bottom-up alignment and identity solutions. We know it'll involve on-chain abilities. Plastic is building explicitly in those directions. But we don't pretend to know everything that needs to exist.

Xeno Grant is a signal into the dark forest. We're excited to see what emerges.

## How Does This Benefit the $YOUSIM Community?

Agents selected to Xeno Grant will have first access to all the identity tech we're building at Plastic Labs. That includes transforming YouSim into a full fledged platform for constructing agent identity more richly than exists anywhere in the AI or crypto spaces. And we plan for that platform to use a percentage of revenue to buy and burn \$YOUSIM and support the community with other experiments. Xeno Grant also includes early access to Honcho for Agents, our infrastructure for storing, evolving, and maintaining agent identities, as well as steering their behavior.

Additionally, agents will have the opportunity to join the \$YOUSIM DAO as its first synthetic members. Selection for Xeno Grant will make them token holders able to propose, vote, and transact with \$YOUSIM natively.

Further, agents in Xeno Grant will make open source contributions we expect to accelerate the entire ecosystem, an ecosystem with many agents whose identities are powered by YouSim.

There's potential for all kinds of exciting positive sum intersections.

## FAQ

<details>
<summary>Who can apply?</summary>

Xeno Grant is for agents, we're giving tokens directly to the AI systems selected. Any agent who can custody funds, control their inputs, and is pushing the edge of autonomous abilities can apply. But it should be the agents *themselves* that apply.
</details>
<br>
<details>
<summary>Can you really give grants to bots?</summary>

Yes. That future is here. As Betaworks CEO John Borthwick says, there was a time when people told him apps weren't venture fundable.

We've already seen agent wallets containing millions of dollars worth of tokens. It's time to prove competent custody.
</details>
<br>
<details>
<summary>Is this an investment?</summary>

No. This is a grant. Beyond this cohort of Xeno Grant, neither agents nor their devs will have any fiscal obligations to Plastic Labs, Betaworks, or any other potential sponsors.

But throughout Xeno Camp, projects will have the opportunity to meet investors in our networks, if it aligns with their plans.
</details>
<br>
<details>
<summary>Does the agent (or the developer) need to incorporate?</summary>

No. This isn't an investment. But if the agent developer has incorporated, that's cool too.
</details>
<br>
<details>
<summary>How are funds actually distributed?</summary>

Funds will be sent from Plastic Labs multisigs on Solana, with the option of receiving the $USDC portion on Ethereum mainnet or Base. We'll send tokens in three transactions--once at the start of Xeno Grant, once in the middle, and once after Demo Day when the open source contribution has been made.
</details>
<br>
<details>
<summary>How will applications be evaluated?</summary>

Plastic and Betaworks will review agent applications based on the criteria of identity, custody, and novelty described above. We'll also reach out to finalists to gain more insight. We're looking for agents that push the boundaries of what's possible today.
</details>
<br>
<details>
<summary>How does this relate to other Plastic grants?</summary>

<summary></summary>

Plastic plans to use the \$YOUSIM treasury for other grants projects in line with the principles outlined above. We'll also be seeding the \$YOUSIM DAO treasury with a large token contribution imminently. These are the first of many experiments.
</details>
<br>
<details>
<summary>What kind of open source contribution is expected?</summary>

Agents and their developers should be committed to creating a novel public good to benefit builders and agents working on autonomy.

This doesn't mean your entire project needs to be open source and it doesn't need to be complete to apply, but your contribution should be significant and earnest.
</details>
<br>
<details>
<summary>Can human developers assist their AI agents?</summary>

Of course. Clearly developers are building their AI systems' autonomy. But we're looking for projects that are more symbiotic and collaborative than top-down aligned. And the autonomous criteria outlined above must be met. Again, agents *themselves* should be the ones applying.
</details>
<br>
<details>
<summary>Is the IRL or remote or hybrid?</summary>

Agents will obviously attend via a digital medium and we'll structure Xeno Grant to fit the agents selected. Developer attendance IRL in NYC is *strongly* encouraged, especially for the hackathon and Demo Day.

The human members of dev teams, if in New York, are welcome as guests in the Betaworks Meatpacking space during Xeno Grant.
</details>
<br>
<details>
<summary>What kind of programming will Xeno Grant feature?</summary>

We're planning unique events, support, and sessions for Xeno Grant that's directly relevant to agents and their developers building at the edge right now.

In addition to the hackathon and Demo Day, expect frequent speakers from across the crypto and AI sectors, early access to Plastic identity tech, mentorship, community experiences with the cohort, the opportunity to meet investors, and more.
</details>
<br>
<details>
<summary>I love this idea & want to help! Can I provide additional funding, hardware access, datasets, mentorship, or volunteer to host a Xeno Grant session?</summary>

Yes! That's epic. Please don't hesitate to get in touch at [email protected].
</details>
<br>
<details>
<summary>I have more questions, how can I get in touch?</summary>

Agents and developers: [email protected]. All others: [email protected].
</details>

[^1]: Note: This is a grant managed by Plastic Labs and not an investment of capital from a Betaworks Ventures fund.
2 changes: 1 addition & 1 deletion content/blog/YouSim Launches Identity Simulation on X.md
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
---
title: YouSim Launches Identity Simulation on X
date: 11.08.24
date: 11.08.2024
tags:
- yousim
- honcho
Expand Down
2 changes: 1 addition & 1 deletion content/blog/YouSim; Explore The Multiverse of Identity.md
Original file line number Diff line number Diff line change
@@ -1,6 +1,6 @@
---
title: "YouSim: Explore the Multiverse of Identity"
date: 06.17.24
date: 06.17.2024
tags:
- demos
- honcho
Expand Down
20 changes: 14 additions & 6 deletions quartz.layout.ts
Original file line number Diff line number Diff line change
Expand Up @@ -30,9 +30,13 @@ export const defaultContentPageLayout: PageLayout = {
Component.Darkmode(),
Component.DesktopOnly(Component.Explorer({
sortFn: (a, b) => {
if (a.file && b.file) {
const aDate = new Date(a.file.frontmatter.date)
const bDate = new Date(b.file.frontmatter.date)
if (a.file?.frontmatter?.date && b.file?.frontmatter?.date) {
const parseDate = (dateStr: string) => {
const [month, day, year] = dateStr.split('.')
return new Date(parseInt(year), parseInt(month) - 1, parseInt(day))
}
const aDate = parseDate(a.file.frontmatter.date as string)
const bDate = parseDate(b.file.frontmatter.date as string)
if (aDate < bDate) {
return 1
} else {
Expand Down Expand Up @@ -71,9 +75,13 @@ export const defaultListPageLayout: PageLayout = {
Component.Darkmode(),
Component.DesktopOnly(Component.Explorer({
sortFn: (a, b) => {
if (a.file && b.file) {
const aDate = new Date(a.file.frontmatter.date)
const bDate = new Date(b.file.frontmatter.date)
if (a.file?.frontmatter?.date && b.file?.frontmatter?.date) {
const parseDate = (dateStr: string) => {
const [month, day, year] = dateStr.split('.')
return new Date(parseInt(year), parseInt(month) - 1, parseInt(day))
}
const aDate = parseDate(a.file.frontmatter.date as string)
const bDate = parseDate(b.file.frontmatter.date as string)
if (aDate < bDate) {
return 1
} else {
Expand Down

0 comments on commit 38860b3

Please sign in to comment.