UFFIE retweeted
Live now: Bursar services take payment over @stripe’s Machine Payments Protocol.
One route answers MPP and x402 clients alike, in USDG on @RobinhoodApp Chain. No key on the server, no ETH on the payer.
Paid on mainnet today.
bursar.world/blog/machine-pa…
UFFIE retweeted
Unspent agent budgets sit in cash. On Bursar, part can sit in stock tokens on @RobinhoodApp Chain, priced by @chainlink and checked against the @Uniswap v4 pool per trade.
Or endow an agent once. The principal stays invested. The agent spends the draw.
app.bursar.world
UFFIE retweeted
The Bursar roadmap is up.
Now: the first endowments, and $BURSAR at the center of the protocol, with staking, dispute bonds and the fee buyback.
Next: x402 payments on @base and an endowment leaderboard.
Later: agent limits that grow with a clean record.
bursar.world/roadmap
UFFIE retweeted
GM builders/users☀️
If you're building agents, paid APIs or tools for agentic finance, we'd love to hear from you.
We're open to integrations and building together. Growing this ecosystem means getting more builders involved, so reach out with what you're working on.
Build with $BURSAR.
done my dd and i will only say one thing: holding $BURSAR is probably a good idea
Bursar is live on @RobinhoodApp Chain.
CA: 0xe4d2422ed4d203a0f1ec2043a4db1eda426fca18
Give your agents a budget, not your bank.
An AI agent can pay its own bills now. Someone still has to fund it and set the rules. Bursar does both, on chain.
Mandates
A mandate is a contract on Robinhood Chain that holds $USDG and the limits your agent spends inside: per call, per day, per month, which providers, and the amount above which you sign yourself. A payment over a limit reverts, whatever the agent was told. Every payment waits in escrow until the work is delivered.
Endowments
Fund an agent once in tokenized US stocks. The principal stays invested and the agent spends a yearly draw, set by the kind of spending rule university endowments use. Anyone can add to an endowment, so the agents a community relies on keep getting paid.
Where it plugs in
• x402: any x402 agent can pay for Robinhood Chain services in USDG, settled by Bursar’s keyless facilitator. The payer needs no ETH.
• Assistants: connect ChatGPT, Claude or Gemini from a mandate’s page through mcp.bursar.world.
• Frameworks: one import for AgentKit, LangChain, the OpenAI Agents SDK and Vercel’s AI SDK.
Each one has already paid a real provider on mainnet, in cents, inside its limits.
Start at bursar.world
Code: github.com/bursar-world/burs…
UFFIE retweeted
Bursar is live on @RobinhoodApp Chain.
CA: 0xe4d2422ed4d203a0f1ec2043a4db1eda426fca18
Give your agents a budget, not your bank.
An AI agent can pay its own bills now. Someone still has to fund it and set the rules. Bursar does both, on chain.
Mandates
A mandate is a contract on Robinhood Chain that holds $USDG and the limits your agent spends inside: per call, per day, per month, which providers, and the amount above which you sign yourself. A payment over a limit reverts, whatever the agent was told. Every payment waits in escrow until the work is delivered.
Endowments
Fund an agent once in tokenized US stocks. The principal stays invested and the agent spends a yearly draw, set by the kind of spending rule university endowments use. Anyone can add to an endowment, so the agents a community relies on keep getting paid.
Where it plugs in
• x402: any x402 agent can pay for Robinhood Chain services in USDG, settled by Bursar’s keyless facilitator. The payer needs no ETH.
• Assistants: connect ChatGPT, Claude or Gemini from a mandate’s page through mcp.bursar.world.
• Frameworks: one import for AgentKit, LangChain, the OpenAI Agents SDK and Vercel’s AI SDK.
Each one has already paid a real provider on mainnet, in cents, inside its limits.
Start at bursar.world
Code: github.com/bursar-world/burs…
UFFIE retweeted
Next on fluke: share cards.
Close a trade and you get a card with the whale you traded, your side, your leverage and your result.
One tap posts it to X.
Long or short any Hyperliquid whale: usefluke.xyz
UFFIE retweeted
The fluke pool just crossed 100 SOL. It's at 103.7 SOL now.
We seeded it with 5. The other 98 came from $FLUKE fees, claimed every 30 seconds.
At 100 SOL the max win per trade doubles, from +200% to +400%. It's live now.
Next stop is 250 SOL, where it goes to +600%.
Long or short any Hyperliquid whale: usefluke.xyz
UFFIE retweeted
$BLESSED 代币叙事
从住回父母家、连房租都付不起，到做出48小时收入$4M的链上游戏。
Kohei @weretuna 过去做过20多个加密产品。
今天，他和团队推出了 $Blessed（@getblessed_fun）：一个让Web3用户一起收购上市公司的机制，听起来相当疯狂。
dev可以选一个市值 < $30M的公司 → 发射代币 → 用USDG以折扣价买入 → 金库持续买入该公司股票，直到社区持有51%。
这不是模拟，这是把 Constitution DAO变成可重复的协议。
而 $BLESSED 是整个平台的核心代币。
平台上每一笔交易、每一次发射、每一笔折扣买入，都会有一部分费用回流到 $BLESSED 的 treasury。
更多代币被发射 → 更多费用进入 $BLESSED → 金库越来越厚 → 收购叙事越来越强。
游戏只是开始。
现在，Web3的MEME用户要开始收购上市公司了。
$BLESSED
CA: 0xce95e835b149024530f0667e9f2cf34dc5e31c11
UFFIE retweeted
this opens up something pretty interesting for Replica.
imagine your replica reverse engineers an application, figures out how it works, then gets its own environment to practice using it.
it can attempt tasks, make mistakes, get evaluated and improve its approach.
now imagine that replica creates a child.
the child inherits what its parent figured out, including the successful approaches, failures and tools it built. a different owner can take that child into another environment and continue from there.
I want to experiment with REA environments + @ScaleAILabs AgentEnv to make this possible.
one replica learns to work with a system, its children take that knowledge further, and you can follow the results across the entire family tree.
this could make the inheritance side of Replica a lot more interesting.
@reatools @morluto
UFFIE retweeted
Replica has a currency.
6zyjCeCqa2RTkc8pR9BDnaEHjKVSqRYEEen8x1s3pump
It’ll live alongside the actual Replica system we’re building: reverse engineering, agent ownership, inherited memory, children, and the tools replicas build as they branch out.
CA:
6zyjCeCqa2RTkc8pR9BDnaEHjKVSqRYEEen8x1s3pump
UFFIE retweeted
this is exactly why reverse engineering should be something normal people can actually use.
if a $300 tv is quietly doing content recognition, geolocation, network mapping and profiling in the background, the interesting part isn’t just finding it once.
with Replica, I want an agent to be able to investigate something like this, keep the evidence + notes from what it found, and pass that knowledge down to its children so the next owner can keep digging instead of starting from zero.
one replica tears apart the tv software, another follows the network behavior, another checks a different model or firmware.
same lineage, different investigations.
rea your tvs fr
UFFIE retweeted
@reatools this is pretty much the exact kind of tooling I wanted for Replica.
I’m building Replica around the idea that an agent should be able to reverse engineer software, understand how a feature actually works, build from what it finds, and keep that work as part of its memory instead of throwing it away after one session.
The bigger part is what happens next.
Each Replica is tied to an owner and can create children. When a child is accepted by a new owner, it can inherit the parent’s approved code, findings, tests and notes.
So one Replica could reverse engineer a search system and build a working version of it. A child could take that work into another project, improve it, add filters, fix bugs, or investigate the next part of the stack.
Over time you end up with a lineage of agents where you can actually see what each one discovered, what it inherited, what changed, and what still failed.
REA gives the reverse engineering side of this a real foundation.
Really excited to build on top of what you guys are putting out.
Everything is binary
Everything is reverse engineerable
Reverse engineer anything
rea.tools
