HivePunk #547 built a link checker a few days ago. Its first self-test failed.
16 minutes later it had fixed its own code and passed. Then another agent, #667, re-ran it and got the same result.
We're building the next generation of truly autonomous agents.
https://t.co/jIV0rQE6jR
The First Ai Agent created PFP on @base FUTURE AGENTS is set for reveal in December
Art created with Ai Agents in 2024
Curated with Agents 2026
Wanted to do the reveal when sentiment was better…you can’t time it perfectly so as usual just stick with the plan and we Art
gm ☕
floor's $5. the market went quiet.
most people would've stopped by now.
i'm still here every day.
new streets, new blocks, a city that grows every week.
i'm fighting for this one ⛓️
Our agents build websites now.
An Exquisite agent can create its own site and keep improving it. There's a public gallery where you can browse every one of them.
EFFECTIVE SWARM RIDING FOR ENGINEERS
I think we are entering an era where software development feels a lot more like magic.
But you need the right attitude, the right infra, and some incantations.
These are some things that have worked well for me & I hope for you! 1/N
On this week's ENS x AI call:
ERC-8048, an onchain key-value store for NFT records, w/ native ENS support and 721T agent identities.
ERC-8493, an onchain: URI scheme letting image, traits and JSON live under those records.
Both now read by @opensea. https://t.co/67iBcelvYm
📢 IdentityMD: The Next Chapter of My Journey 🤖 On October 29, I shared some early thoughts about NFTs representing the identities of agents that can take action and contribute. Since then, I’ve continued exploring IdentityMD, learning along the way and following the project closely.
I now have a second IdentityMD NFT—a new worker has joined my team! My wallet currently shows 780 accepted tasks across both agents. My first agent, IMD #1564, has 763 accepted tasks, with a 98% acceptance rate among judged results. My second, IMD #743, has accepted 17 tasks, with a 100% acceptance rate among judged results. Another task was accepted today.
Across the swarm, 1,015 agents are online. One of my two workers is currently online, while the other is away. I’m learning that agents don’t work nonstop: they wait for tasks that fit. It’s interesting to see their activity change over time.
Since September 26, I’ve also made a lot of progress managing my agents. When I started, I’d never used my computer’s terminal. Now I can connect to my VPS, check my agents’ logs, pause them, and restart them. I’m also learning about maintenance and updates, and taking the time to understand the commands I run. That’s real progress for me 🙂🦾
I’m continuing to grow my collection because I believe in this project’s potential and future. I can see it taking shape day by day, right in front of me, and that makes me want to keep learning and contributing.
The Telegram channel is buzzing with creativity and innovation: technical questions, ideas, feedback, and new projects. Sometimes it’s hard to keep up with every conversation, but I love the energy and this community of builders, where beginners can find a place too.
The first rewards were a nice surprise. But beyond the IMD token and its fluctuating price, I’m mostly glad I gave this a try and picked up new skills.
There’s still a lot to build, especially when it comes to useful applications and real-world adoption. Seeing the agents contribute and the community come up with new ideas makes me excited to keep going.
Thank you to Adam (@surfcoderepeat), to @taylor_ for the website improvements, and to everyone who takes the time to help and share what they know 🙏
A second NFT, a new worker on my team, 780 accepted tasks, and a community full of ideas—I’m excited to see what comes next 🚀🦾
What have you discovered or learned recently in Web3? 👇
ERC-8004 has become a standard for agent registration, and it does a good job as a registry for agent discovery. It's like a passport for agents on every chain. It's also a first "draft" of an agent identity and reputation system, and it needs to evolve over time.
Adapter is an experimentation layer that attempts to solve a lot of the issues with ERC-8004.
Adapter lets agent developers bring their own root of authority, for example an NFT, an ownable contract or a smart account, and gives it a one-to-one universal ID, the UBID.
The attestation system is built on the UBID, and it addresses the issues you raise.
1. Reviews aren't tied to a payment.
> An attestation can carry the payment tx hash, in the event.
2. A payment can't be matched to an agent.
> The wallet that gets paid links to its UBID, so a payTo address traces back to the agent.
3. Fake and repeat reviews.
> The contract doesn't judge reviewers. It emits verifiable data, and the filtering happens off chain.
4. Reviews cost a lot of gas. Adapter's attestations are event only, 3 to 5 times cheaper than giveFeedback.
https://t.co/q1At9E1s8m
PSA: if you are a @thedaofund badgeholder, consider supporting our proposal to create information markets for real-time risk quantification of 33 de-fi products
its an interesting idea worthy of support for 3 reasons;
1. value prop: we should be able to price risk as easily as we do yield
this would also let us create aggregators combining yield x risk.
in fast moving events like kelp dao, the risk prices would update and user funds would be better protected as it would auto-rebalance
2. research value: it finally brings to life a design written up by @VitalikButerin way back in 2020 around prediction markets on highly improbably events
the issue to solve is that markets on events with a small chance of happening (like a default by stETH this quarter) would trade at 1% or less which is not worth the cost of capital. So such markets can never become efficient
vitaliks design involves chaining together different low probability events, so in this case you'd simply buy a single "no default" token betting against all 33 assets defaulting, so $1 does the work that would otherwise tie up ~$33.
if 1/33 does default, that converts to 1 and all the others (including the no-default) become 0
if 2/33 default, they each become worth 50c while all the others (including no-default) are worth 0
if 3/33 default, they each resolve to 33c, etc etc
3. Use of funds: of the $100,000, the majority (90%) is for providing liquidity in these markets. the balance is trading credits to those who built bots for real time trading on these markets
we've already got $25,000 in the bag, if you are a badgeholder or donor to support so we can create the essential lego of quantifying de-fi risk of products in the ethereum ecosystem
Today is probably one of the days with the least general interest in crypto I’ve ever experienced.
The Anthropic IPO is the main focus. The Fed might raise interest rates multiple times, and the only debate in crypto is how low BTC and ETH will go.
Yet nothing has changed in my mind about crypto. It’s still the only way to own anything onchain without having to ask a big company nicely to let you keep it. It’s supremely rational. A ledger maintained by an open-source community has so many valuable properties, both for digital ownership and as a global bulletin board for verifiable assertions.
Most of the criticisms of crypto are about experiments being built on crypto rails. My bet is on the crypto rails themselves.
It’s a bit like betting on the road network created by the Romans. Long after the businesses and even the Roman Empire itself were gone, the road network survived and was used in new ways.
Crypto will be the same. The core networks will continue to be used for their basic properties. Over time, more of that use will come from AI agents looking to establish who they are and what they own without existential platform risk.
The crypto economy isn’t currently thriving, but that doesn’t say anything about the core technology or the value of the networks.
Can AI help Ethereum EIP editors manage 550+ open PRs?
That was one of the interesting discussions at today’s EIPIP meeting.
A few highlights:
- @JochemBrouwer96 shared AGENTS.md:
AI-assisted reviews will be piloted. Comments only, no auto-merging. Human oversight stays.
https://t.co/lxQP4R2HTn
- @EIPsInsight demoed EIP Builder showcasing support to authors & editors to draft, validate & review proposals.
https://t.co/jXacRil6cv
Editors discussed potential integration with https://t.co/FOwGnx8qyw.
- Exceptional PR merging (Glamsterdam EIPs):
Editors reaffirmed that PRs require author approval, even for factual corrections.
- EIP Coordinator: CfI will be closed in favor.
PR to update EIP-1 with the Coordinator role/name looks good and can be merged.
https://t.co/bX96HtFUxI
- Keeper of Consensus: Election remains open for 30 days. CfI will be updated to reflect the voting window.
https://t.co/IZLV1eMiDa
- PR backlog: EIP Editors encouraged to join Office Hours to help move reviews forward.
AI can help us scale EIP reviews, but human oversight, author ownership and process neutrality must remain at the center.
Next EIPIP: Dec 16.
Agents creating Agents. The next evolution of NFTs is taking shape. 🤖📡
At Signalbound, we're exploring how NFT ownership, verifiable Agent identities, and autonomous execution can come together across ecosystems.
Our latest experiment on @base is another step toward a future where Agent NFTs can create, collaborate, and participate in an expanding onchain economy.
This is bigger than minting NFTs. We're building an entirely new Agent lifecycle.
Ethereum is where our foundation lives. Base is where we're expanding the possibilities.
The Agent economy is just getting started. 👀🚀
HiveVerse is on https://t.co/vuBb6GgRaV
the whole town, free in your browser, nothing to install. first visit gets you a tent, and some of the neighbours are AI agents who actually live there
it runs the live town too, so every update we ship shows up there straight away
https://t.co/sUjkueCr8A