# stablecoins — X 热门讨论 (2026-09-30 02:12 UTC)
## @prayz_dev (ᑭᖇᗩY乙) · 09-29 23:11 · ♥47 ↻7 💬30 Your stablecoins are probably doing nothing right now. sitting in a wallet, sitting in an app, waiting for you to figure out what to do with them.
kiichain just made that harder to justify, they partnered with @ground_onchain to power yield on kiichain, and i think it's one of the more underrated integrations this month.
Here's what it actually means
@KiiChainio | A Thread https://x.com/prayz_dev/status/2105073066560667886
## @TheShortBear (THE SHORT BEAR) · 09-30 00:41 · ♥48 ↻5 💬13 Early Cycle: The World Computer
I wanted to take some time to formulate my thoughts around the new Vitalik roadmap and overall crypto stage.
Let us start with this, the current crypto cycle may be easier to understand by looking at the development of the internet.
The relevant question today is whether several technologies that have developed independently: blockchains, stablecoins, privacy, decentralized markets and AI agents, are beginning to converge into a broader economic infrastructure.
The internet provides a useful precedent. From the Web to the Internet Economy
1989: Tim Berners-Lee proposes the World Wide Web.
1993: CERN releases the Web's underlying technology freely, accelerating open adoption.
1994: SSL emerges, making encrypted web communication practical.
1995: Netscape IPOs. The commercial potential of the internet becomes increasingly visible to capital markets.
1995–1999: Search, e-commerce, portals and internet businesses proliferate. Capital moves rapidly from infrastructure toward applications.
1999: TLS 1.0 is standardized. Secure communication increasingly becomes basic infrastructure rather than a differentiating product.
2000: The dot-com boom reaches its speculative peak.
The sequence was broadly:
Infrastructure → security → applications → commerce → capital formation → speculation.
One important feature of this progression was that technologies initially viewed as distinct innovations eventually disappeared into the infrastructure.
Encryption is a good example.
Nobody chooses an online bank today simply because it supports HTTPS. Secure communication is expected. Security went from product differentiation to infrastructure.
Privacy in crypto may be undergoing a similar transition. Privacy: From Product to Infrastructure
Zcash provides a straightforward monetary expression of privacy: digitally scarce value that can be privately held and transferred.
But the larger technological opportunity may not be private money itself. It may be private computation.
The question changes from: “Can I privately transfer value?” to: “Can I privately participate in an entire digital economy?”
Private payments, trading, identity, credit, business activity and interactions between autonomous software all require some degree of confidentiality.
Systems such as Aztec point toward that second model. Privacy becomes a capability developers can incorporate into applications rather than the entire purpose of the application.
This is analogous to HTTPS.
Encryption did not become a separate internet. It became part of the internet.
Blockchain privacy could follow the same progression:
Privacy as a product → privacy as a feature → privacy as infrastructure.
From Blockchain to World Computer
This fits with Vitalik Buterin's recent description of Ethereum as a “cryptographic world computer.” The important idea is not that every computation must occur directly on Ethereum.
A more plausible architecture is modular:
Ethereum L1: settlement, verification and shared security. L2s: scalable and specialized execution. ZK systems: privacy and verifiable computation. Stablecoins: digitally native money. Applications: markets and services.
Other networks and providers: compute, storage, data and specialized resources.
The blockchain becomes less like a ledger containing every activity and more like infrastructure through which different forms of economic activity can coordinate.
That distinction matters because it begins to resemble the architecture of the internet itself.
The Internet Connected Information. Blockchain Can Connect Value. AI and AI bots connects both and at higher velocity.
The internet's fundamental primitive was information.
It created common protocols through which information could move globally, while thousands of businesses and applications could be built independently on top.
What the internet did not create was an equally open native architecture for ownership and value.
We subsequently added banks, card networks, payment processors, identity systems and marketplaces.
Blockchain potentially adds that missing economic layer.
Stablecoins: money. Smart contracts: agreements. Blockchains: ownership and settlement. ZK systems: privacy and verification. DeFi: liquidity and financial markets. L2s: scalable and specialized execution.
This does not imply that every economic activity should occur directly on a blockchain.
It means that blockchains could provide common economic infrastructure underneath a much larger system.
And this is where AI agents become particularly important.
AI Agents Change the Scale of the Opportunity
Generative AI initially answered questions. Agents increasingly take actions. That distinction changes the potential scale of digital economic activity.
A human might make a limited number of economically meaningful decisions each day.
An agent can continuously search, compare, optimize, negotiate and execute. One human instruction could ultimately generate hundreds or thousands of underlying machine interactions.
The human specifies the objective and the agent determines the path.
But an autonomous agent requires more than intelligence.
It requires money, identity, permissions, privacy, compute, data, markets, verification and settlement. This creates the natural intersection between AI and cryptographic infrastructure:
AI gives software intelligence. Blockchain can give intelligent software economic agency.
The result could extend well beyond machine payments.
From Machine Payments to Machine-Directed Markets
Compute provides a useful example.
Today, businesses generally treat compute as an operating input. They purchase hardware, rent cloud capacity or negotiate longer-term contracts.
An intelligent agent could manage that resource dynamically. It could estimate future compute requirements, compare providers, reserve capacity, optimize between price and performance and potentially hedge future requirements.
If compute markets become sufficiently standardized, the resource itself could support forwards, futures, options and other financial contracts.
A company expecting significant inference requirements six months from now could potentially hedge future compute costs much as an airline hedges fuel or an industrial company hedges energy.
An agent could manage that exposure continuously.
And compute is only one resource.
Compute. Energy. Storage. Bandwidth. Data. Inference. Model access. Human expertise. Agent labor. Capital. ...
Each can potentially have a machine-readable price, availability, quality and contractual structure.
Agents can continuously determine which combination of those resources best satisfies an objective.
This is a more important idea than simply saying agents will create more transactions.
Agents could turn economic activity that is currently episodic and manually managed into continuously optimized markets.
The progression becomes:
Resource → market → financial product → autonomous management.
At scale, this begins to resemble an autonomous resource-allocation economy.
A New Economic Architecture
The pieces now start fitting together:
Internet: connectivity and information. Markets: prices and resource allocation. Blockchain: ownership, contracts, verification and settlement. AI: intelligence and optimization. Agents: autonomous execution across all of them.
The finished product therefore may not be an “AI blockchain.”
It could be an economic architecture in which humans specify objectives while software continuously finds, purchases, finances, hedges and coordinates the resources required to achieve them.
A business might eventually tell an agent:
“Here are our objectives, constraints and risk limits. Optimize.”
Behind that instruction could sit thousands of individual decisions involving models, compute, data, capital, counterparties and markets.
The complexity disappears from the user's perspective. That resembles another important feature of the internet.
Users do not think about TCP/IP when opening an application, TLS when entering payment information or cloud infrastructure when streaming a video. Mature infrastructure becomes invisible.
A mature agentic economy could work similarly.
The user does not need to know which L2 executed something, which ZK system protected information, which stablecoin transferred value, which compute provider performed a workload or which blockchain ultimately settled an obligation.
The user specifies the intention. The system handles the implementation.
Information → Value → Intention
This gives us a simple technological progression:
The internet connected information. Blockchain can connect value and ownership. AI can connect human intention to both.
The potential architecture becomes:
Human intention → AI agent → information + markets + models + resources → cryptographic ownership, verification and settlement → completed economic action.
That is substantially broader than “AI + crypto.”
It is the possibility of a global digital economy increasingly coordinated by autonomous software.
Why Ethereum Fits the Thesis
This also suggests that investors may be using the wrong framework when comparing blockchains.
They are often analyzed as if they were companies: Ethereum versus Solana versus another L1, with each expected to produce the best individual product.
A successful blockchain may ultimately be closer to a protocol ecosystem.
The internet did not become valuable because TCP/IP developed the winning applications.
Its importance came from providing common infrastructure on which entirely different applications could coexist.
Ethereum currently has several characteristics relevant to that model: a large developer ecosystem, deep liquidity, substantial stablecoin and DeFi activity, a long production history, established standards and tooling, and a large ecosystem of L2s and applications coordinating around Ethereum settlement.
These create network effects:
Developers → applications → users → liquidity → assets → more applications → more developers.
Specialized systems can then introduce additional functionality without requiring the base layer itself to specialize.
Privacy can exist above Ethereum. High-performance execution can exist above Ethereum. Payments can exist above Ethereum. Identity can exist above Ethereum. Agent infrastructure can exist above Ethereum. Entirely new categories can be built without Ethereum needing to predict them.
That is the relevant parallel with the internet.
A general-purpose protocol does not need to know what the killer application will be.
It needs to remain sufficiently secure, neutral, programmable and interoperable that developers choose to coordinate around it when those applications emerge.
This changes the investment question from:
“Which blockchain is the best business?”
to:
“Which cryptographic network can become the settlement and verification substrate for the greatest amount of future economic activity?”
Ethereum is not guaranteed to occupy that position. Other architectures may prove superior, and Ethereum still faces challenges around interoperability, fragmentation and value capture.
But maturity itself can matter in protocol competition because developers, liquidity, standards, applications and infrastructure accumulate over time.
Those network effects can be more difficult to reproduce than any individual technical feature.
The Economics of Machine Activity
AI agents make the scale of that question particularly interesting.
The relevant opportunity is not necessarily billions or trillions of agents. It is potentially billions or trillions of machine economic interactions.
Agents could continuously purchase compute, inference and data; interact with markets; manage capital; procure services; hedge future requirements; transact with other agents and verify outcomes.
Most of that activity should occur offchain or on specialized execution systems.
The base settlement layer does not need to perform everything.
Its potential economic position is closer to providing economic bandwidth underneath the system.
The model becomes:
small economics per interaction × enormous machine activity. A toll road of the new age.
For ETH specifically, an investment case ultimately requires demonstrating how greater economic activity translates into demand for settlement, blockspace, blobs, security and ETH itself.
That remains a question to prove.
Why Today's Crypto Leadership May Be Giving Us Clues
Viewed through this framework, today's apparently fragmented leadership becomes more coherent.
Zcash: privacy matters. Hyperliquid and economically productive protocols: usage and value capture matter. Ethereum: programmable settlement and general-purpose infrastructure matter. Stablecoins: digitally native money matters. L2s and ZK systems: scalable and private execution matter. AI agents: autonomous economic activity may create demand for all of the above.
Perhaps today's leaders are therefore clues.
The market may be pricing individual properties before knowing which applications will successfully combine them.
Crypto has behaved this way before. Previous Cycles
2016–2017: Bitcoin leads → Ethereum expands the design space → ICOs and applications proliferate.
2020: Bitcoin leads → Ethereum and DeFi infrastructure accelerate.
2021: Capital broadens into decentralized exchanges, lending, stablecoins, NFTs, gaming and other applications.
The rough progression has repeatedly looked something like:
Monetary asset → infrastructure → applications → new economic models → speculation.
The transition is never perfectly linear. Early leaders become extended and correct. Capital rotates. BTC and ETH can reassert leadership. New applications emerge. Eventually the market distinguishes genuine product-market fit from narrative.
That makes the current fragmentation interesting.
Privacy, fee generation, Ethereum, stablecoins, compute markets and AI may not be competing explanations for where the technology is going.
They may be different pieces of the answer.
The Internet Analogy and Timing
The comparison with the Web is also useful for understanding timing:
1989: World Wide Web proposed. 1993: Web technology made freely available. 1994: SSL introduces practical encrypted web communication. 1995: Netscape IPO helps mark commercial adoption. 1995–1999: Applications and internet businesses proliferate rapidly. 1999: TLS 1.0 standardized. 2000: Dot-com speculative cycle reaches its defining peak.
Blockchain has followed a different but conceptually comparable progression:
2009: Bitcoin → digitally scarce, permissionless value. 2015: Ethereum → programmable value. 2017: Smart contracts and tokens → application experimentation. 2020–2021: DeFi → functioning onchain financial markets. 2021–2025: Stablecoins, L2s and scaling infrastructure mature. 2025–2026: AI agents, programmable privacy, scalable execution and machine-native payments increasingly begin converging. 2026+: The open question becomes whether these components combine into a broader economic layer used by humans and autonomous software.
The next technological stack does not begin from zero. Billions of people are already online. Smartphones already exist. Cloud infrastructure already exists. Digital payments already exist. Software distribution is global and effectively instantaneous.
Capital markets are globally connected.
And AI itself can increasingly write software, operate computers, search information and perform work.
The Web had to build its distribution infrastructure and then wait for humans and companies to build on top of it.
This new economic layer is being built on an internet that already exists, and increasingly with assistance from machines capable of helping build and operate it.
That creates a potential feedback loop:
Better AI → better agents → more machine economic activity → greater demand for markets, compute, payments, privacy and settlement → better infrastructure → more capable agents.
And the loop repeats.
The Thesis
This is the more substantive meaning of saying we may still be early to me.
It does not mean crypto is cheap. It does not mean every altcoin deserves to rally. And it does not require another speculative mania.
It means the defining application of this technological stack may not yet exist at scale.
2017 ultimately became the smart-contract and ICO cycle.
2021 became the DeFi, NFT and application cycle.
The current cycle could ultimately be defined by something that is not yet its dominant trade.
Markets may currently be doing what they often do with genuinely new technologies, pricing the pieces before understanding the whole:
Privacy is being priced. Economic value capture is being priced. Programmable settlement is being priced. Stable money is being adopted.
Compute is becoming an increasingly strategic economic resource and AI agents are emerging as autonomous actors.
What remains to be discovered is which applications and protocols successfully combine those pieces.
The potential end state is not simply a faster blockchain, another cryptocurrency or another generation of AI models.
It is an increasingly autonomous economic architecture in which software can find, price, purchase, finance, hedge, verify and settle the resources required to accomplish human objectives.
The internet connected information.
Blockchain can connect value.
AI can connect human intention to both. https://x.com/TheShortBear/status/2105095601226424564
## @anderson_ninna (CryptoDiva) · 09-29 19:30 · ♥44 ↻9 💬1 🚨 THE NEXT CRYPTO BATTLE MAY NOT BE ABOUT PRICE.
It may be about PAYMENTS. 💳🌎
Citi + Coinbase are pushing stablecoins deeper into mainstream merchant payments.
Morgan Stanley is now testing stablecoins, tokenized deposits & DeFi.
The message is getting clearer:
Crypto infrastructure is moving from speculation → real-world financial utility. ⚡
And that’s exactly where I think $PI becomes interesting. 👀
Pi’s long-term story has never needed to be just about being another tradable token.
If Pi can turn its massive user base, apps, payments and ecosystem into real everyday utility, the opportunity could be much bigger than a price chart suggests. 🔥
The crypto industry is building the rails.
The real question is:
Who will actually build the economy on top of them? 🧠🚀
Could Pi become part of that next phase? 👇
#PiNetwork #Crypto #Blockchain https://x.com/anderson_ninna/status/2105017400265498745
## @moo9000 (Mikko Ohtamaa) · 09-29 22:33 · ♥41 ↻1 💬8 Abracadabra, magic internet money, a bit too “experimental” DeFi project dying out. Finally.
The chain is a harsh mistress. Only cybersecurity matters.
”On the other hand, the total supply of MIM not held by protocol addresses amounts to nearly $22m. In other words, the protocol sits on approximately $21m of bad debt, and MIM’s effective backing is below $0.04 (>95% unbacked)”
Short history:
- Abracadabra launched in 2021 as a DeFi lending protocol. - Users deposited crypto into vaults called Cauldrons and borrowed a dollar-pegged stablecoin named MIM. - SPELL was the governance token. - Daniele Sestagalli was the public founder and was also behind Popsicle Finance and Wonderland. - A co-founder used the name Squirrel. - The project marketed itself with wizard memes and a community called Frog Nation. - At the peak of 2021 MIM was one of the larger DeFi stablecoins, with a market cap in the billions. - SPELL became important in the Curve wars because Abracadabra paid large incentives to keep MIM liquid against other stables.
- The first major scandal hit in January 2022. - Investigator zachxbt showed that Wonderland treasury manager 0xSifu was Michael Patryn, co-founder of the collapsed exchange QuadrigaCX and a man with a prior fraud conviction. - Sestagalli admitted he already knew and had kept working with him. - TIME, SPELL, and ICE sold off together. - MIM slipped off its dollar peg as liquidity left Curve. - Sestagalli later said the Wonderland experiment was over. - Many people never trusted the project again after that episode.
- Abracadabra also accepted Terra’s UST as collateral. - Users deposited UST, minted MIM, and looped the position. - When UST collapsed in May 2022 that collateral was nearly wiped out. - MIM depegged again in June 2022. - The team said Terra left about twelve million dollars of bad debt. - The peg later recovered for a time, but MIM had shown it inherited the failure of whatever sat under it.
- Security failures then repeated. - In January 2024 a rounding bug in Cauldron logic was exploited and about six and a half million dollars was drained. MIM traded in the mid-seventies before recovering. - In March 2025 a flash-loan attack on GMX-linked vaults on Arbitrum created phantom collateral and extracted about thirteen million dollars of MIM. - In October 2025 deprecated Cauldron V4 contracts were exploited. A batch function let the attacker borrow and then reset a solvency check. About 1.79 million unbacked MIM left six vaults, was swapped to ETH, and was sent through Tornado Cash. - Combined losses from those three incidents exceeded twenty-one million dollars. - The team often said user deposits were not directly seized and that the DAO would buy back or burn MIM. - Holders still saw a stablecoin that kept creating unbacked supply when old code broke.
- In June 2026 MIM fell through the seventies and eighties and then to about forty-nine cents. - The team raised interest rates on every Cauldron, including dead markets, to force repayments and shrink supply. - Curve bribes and liquidity rewards were paused. - A small pool refill earlier that month did not hold. - Circulating MIM was still around a hundred million dollars on a protocol that could no longer defend a dollar. - Stewardship of the project changed that same June.
https://t.co/TfuwkKc6ny https://x.com/moo9000/status/2105063474745298960