| | Crypto Twitter has been looking a lot like AI Twitter of late. Every week is replete with excitement around AI-related tokens for AI agents, agent frameworks, agent launchpads, AI investment DAOs, AI memecoins, etc. Keeping up with them all is damn near impossible, but we'll try to make sense of some of them today. | | | Solana DEX volume: | | This chart tracks Solana's total trading volume across various DEXs over the past few months. Raydium — the largest DEX by far — dominates volume, followed by Orca, Lifinity, Meteora, Phoenix and pump.fun, with smaller volumes from each. | Blockworks Research excluded volume from a new wash trading bot that popped up lately and executed over $800 billion in wash trades in just two days. | As Dan Smith observed, the bot's activity involved a flash loan of millions in USDC, which was swapped for JUP, then immediately swapped back into USDC, with no profit generated and no other trades in between. This behavior is not arbitrage, but a form of market manipulation, as the same account controls all the transactions. | The trades are primarily conducted on newly launched Orca v2 pools with zero swap fees and minimal liquidity, making them vulnerable to this type of activity. The account behind the trades also created and seeded these pools with initial liquidity. | The apparent culprit is @STACCoverflow, who explained it as a test of sorts. "[It's] volume for the masses. [If people] like volume bots, this one is pennies on the billions, coming soon to a no-code-eliza-launching-ui-near-you." | Or it could be farming a prospective Jupiter DEX airdrop campaign. | — Macauley Peterson (X: @yeluacaM | Farcaster: @Macauley) | | An overview of the AI agent meta | | The AI agent frenzy was kickstarted by GOAT. GOAT is a Solana memecoin that spawned from an experiment by creator Andy Ayrey when he pitted two Claude AI models in an endless debate. GOAT still sits today at a comfortable $404 million market cap, but no longer dominates the mindshare it once did. | First to copy GOAT's success was the AI agent Luna (LUNA). It's a busty anime girl chatbot that streams 24/7 on social media — need I say more? Chat with her, tip her with her own LUNA token (why does anyone do that?), or be like A16z-backed Story Protocol: Hire Luna for social media marketing. | The Luna virtual companion was launched on Virtuals Protocol, an "AI Agent launchpad" on Base that is the most recognizable brand name in this meta right now. | Formerly the PathDAO gaming guild, Virtuals took pump.fun's successful formula and applied it to AI agents. Launch an agent token, hit a target market cap on a bonding curve, and pair the agent's token in a liquidity pool on Uniswap. | What Virtuals did differently, however — and what has gotten investor-types bullish on its token — was tying its native VIRTUALS token to the cost of launching agents and running inferences, as well as pairing it with agent tokens in a Uniswap liquidity pair (rather than SOL). | Within the Virtuals ecosystem, Luna faces competition for speculative capital from other similarly looking chatbots like Airene, or more "utility-based" agents like aixbt, an agent that scrapes troves of social media data to surface trading alpha (aixbt is incidentally bullish on everything). |  | Diamond Hands Dig 👾 💎 @DiamondHandsDig |  |
| |
$AIXBT is now a $1 billion market cap project based on mindshare, and it still wouldn't be overpriced. @aixbt_agent has undeniably become the most influential entity in all crypto right now. From @cookiedotfun to @_kaitoai, the signs are everywhere—this is a movement, not just… x.com/i/web/status/1… | Diamond Hands Dig 👾 💎 @DiamondHandsDig Now $800m projected market cap vs. mindshare for $AIXBT, according to @cookiedotfun. Two things seem guaranteed in AI crypto right now: 1️⃣ Hot air rises with $FARTCOIN. 2️⃣ Mindshare rises with @aixbt_agent (according to @_kaitoai). Take that as you will for $AIXBT's future… x.com/i/web/status/1… x.com/DiamondHandsDi… |
| |   | | | 7:01 AM • Dec 29, 2024 | | | | | | 116 Likes 19 Retweets | 6 Replies |
|
| Just as agents face competition, so too do the picks-and-shovels of agents. Besides Virtuals, a myriad of AI agent launchpads already exist, such as vvaifu.fun, Creator.Bid, MyShell.ai, Zentients and more. | Then we have the AI agent "frameworks." These are simply code libraries that make it easy for devs to customize their agents and integrate them into the virtual Wild West. | The leading agent framework is ai16z's "Eliza," but there's also Virtual's G.A.M.E (Generative Autonomous Multimodal Entities), the Python-based ZerePy, the Solana-focused RIG, and more. And of course, each one of these frameworks has its own token that speculators are furiously aping into. | | Finally, we have the AI investment DAOs. The idea is simple: Deposit SOL/ETH and receive a DAO token that lets you share in the "hedge fund's" upside. | Ai16z was the first investment DAO to launch on Solana, now with an AUM of $23.4 million. People apparently really trust the investor acumen of AI agents. Presently, its native AI16Z token is valued at an astounding $1.4 billion market cap. | Daos.world and VaderAI are two upcoming AI investment platforms on Base that have gained steam in the past week. At least six funds have already emerged on both platforms, with about a collective million in capital raised. | — Donovan Choy (X: @donovanchoy | Farcaster: @donovan) | | | "Hari," CEO and co-founder of Cantina and Spearbit DAO, kicked the hornet's nest of devout Ethereans with an argument against client diversity. He argued on X that it led to Ethereum's ossification, complicates governance, made frequent hard forks unrealistic, created unnecessary conflicts, scapegoated Geth — a popular client — and ultimately, served as a misguided purity test. Naturally, this did not sit well with some.
|  | Mark (ethDreamer.eth) 🦇🔊 @EthDreamer | |
| |
Insane take. First off, there's hardly a correlation between the client team you're on and your opinion on any given roadmap issue. At least on the consensus side (with 5 clients) I've seen almost none of this. People on the same team have different opinions and people on… x.com/i/web/status/1… |  Hari @_hrkrshnn
Client diversity on Ethereum was a mistake. It forced Ethereum to become ossified. Every change today involves multiple stakeholders, each competing for the scarce resource of what goes into the next hard fork. The governance process ultimately has no way to resolve tiebreaks.… x.com/i/web/status/1… |
| | | 6:04 AM • Dec 27, 2024 | | | | | | 318 Likes 70 Retweets | 14 Replies |
|
|  | sassal.eth/acc 🦇🔊 @sassal0x | |
| | Replying to@_hrkrshnn | This tweet is really weird to me - you seem to not understand just how critical it is for Ethereum to have client diversity - let me explain. You are advocating for faster changes to Ethereum mainnet while also saying it'd be better off with 1 client which is *crazy*. The risk… x.com/i/web/status/1… | | | 11:30 PM • Dec 26, 2024 | | | | | | 466 Likes 41 Retweets | 20 Replies |
|
| |
| | Hari @_hrkrshnn | |
| |
It looks like my client-diversity post was very controversial. Let me make some clarifications. - One written specification with multiple implementations sounds appealing but rarely works in practice. In reality, the spec is the code. The Ethereum Yellowpaper, for example,… x.com/i/web/status/1… | | | 1:52 AM • Dec 28, 2024 | | | | | | 117 Likes 9 Retweets | 21 Replies |
|
|  | James Prestwich @_prestwich | |
| |
client diversity people never choose specific metrics, so you can't argue with them. it's all feels. personally, I care about:
- long time between incidents
- fast incident response times
- low incidence of long reorg these are optimized with a simpler single-client network | | | 5:06 PM • Dec 28, 2024 | | | | | | 72 Likes 7 Retweets | 10 Replies |
|
| |
|
| | |
|
No comments:
Post a Comment