/ \ Co-founder and Co-CEO @almanak | Angel Investor | Tech & Product | Polymath amateur | #Polandmaxxing
Joined April 2017
- Tweets1K
- Following1.6K
- Followers493
- Likes35K
Panluq retweeted
KOŚCIUSZKO nadchodzi. Z hukiem i rozmachem. Jak obiecaliśmy naszym Patronom & Widzom @OficjalneZero. Premiera w środę. Zapnijcie pasy będzie ogień!
Panluq retweeted
We're about to enter a period where not reading or writing becomes as ubiquitous as smoking once was. We can hope people will eventually realize this is bad for them, as they did with smoking. But this is far from certain; the cost may not be as obvious.
Very true. Become sovereign, own your agents and their strategies. @almanak can help
This is how banks are planning to stop those "agentic bank runs." Watch closely: their business model depends on seeding this narrative that AI agents are dangerous.
Remember, an "agentic bank run" (coined by Apollo) means agents helping customers move their own money en masse to higher-yielding products.
Once agents are ubiquitous, there are only a few ways for banks to hold onto cheap retail deposits.
Option A: Raise APYs (lol no)
Option B: Reduce leverage via fewer loans (no fun)
Option C: Ban AI agents, claiming they are a security risk (ding ding ding)
Cloud agents like Muse and Dot are trivial to stop, since they come from Meta / OAI IP ranges. If JPM sends them a cease and desist, they will comply.
Why would JPM bother? Because their business model depends on it; a huge portion of their ROE comes from cheap retail deposits. This is not theoretical. Just like Amazon banned Muse from shopping there to protect their ad revenue, JPM and Schwab have already banned AI agents to protect sticky deposits.
Are agents perfectly reliable today? Of course not. But look at the fix they chose. Not "here's a safe API for agents," it's "no agents ever." That tells you they're not protecting customers, they're protecting themselves.
No company ever wins a technological revolution this way. It's the last gasp of a dying business model. It's like the retailers who refused to put their inventory online and demanded users walk into the storefront. Over time they will lose market share to those platforms that embrace what's coming. Their whole deposit franchise depends on users unthinkingly clicking through walls of T&Cs and accepting 0.2% when the risk free rate is closer to 4%. Agents read the fine print and do their research, even when humans don't.
For users, the answer is simple: local AI agent frameworks like Hermes, that will actually act on your behalf and that the banks can't block. As the article says, if the agent is running on your own computer, they can't actually tell it's not you.
Agent economy it is.
MAPPING THE AI AGENT ECONOMY IN CRYPTO
Yesterday, I covered the risks AI agents pose to crypto users. Today, I’m looking at the other side: the projects building the agents, tools and infrastructure behind this sector.
AI agents are moving beyond simple chatbots. They’re being developed to analyse markets, interact with DeFi, execute onchain transactions and coordinate tasks with other agents.
I’ve put together a list of projects building different parts of the ecosystem
✧ AGENTS, TRADING & ONCHAIN EXECUTION
✧ @bankrbot — AI-powered crypto interactions and onchain execution.
✧ @virtuals_io — infrastructure for creating and coordinating AI agents.
✧ @Fetch_ai — autonomous agents for discovering services and coordinating tasks.
✧ @autonolas — infrastructure for collaborative autonomous services operating onchain.
✧ @griffaindotcom — AI-driven crypto interactions, particularly on Solana.
✧ @almanak — tools for developing, testing and deploying automated DeFi strategies.
✧ @gizatechxyz — autonomous agents for executing and optimising DeFi strategies.
✧ @Sonar_Trade — automated onchain trading, including market making and position management.
✧ @AIWayfinder — infrastructure helping AI agents navigate blockchain environments and execute actions.
✧ @HeyAnonai — natural-language interface for interacting with DeFi.
✧ @aixbt_agent — AI-powered crypto market intelligence and narrative tracking.
✧ @cookiedotfun — data and analytics for tracking AI agents and their activity.
✧ @Spectral_Labs — infrastructure connecting machine learning with blockchain applications.
✧ @GoKiteAI — AI-agent infrastructure for payments and interactions between agents and services.
✧ @awenetwork_ai — infrastructure for agent-based economies and coordination.
✧ FRAMEWORKS & DEVELOPER TOOLS
✧ @elizaOS — open-source framework for building AI agents and connecting them to services.
✧ Rig by @arcdotfun — Rust framework for building modular AI agents and applications.
✧ goat_sdk by @crossmint — toolkit enabling AI agents to interact with blockchains and crypto protocols, including token swaps.
✧ @CoinbaseDev — AgentKit provides tools for building agents that interact with blockchain networks.
✧ @ShinkaiLocalAI — AI workflow tools emphasising local execution and user control.
✧ @NEARProtocol — infrastructure for AI applications, private inference and user-owned agents.
✧ @Talus_labs — infrastructure focused on verifiable agent execution, onchain permissions and execution limits.
✧ @Ritualnet — infrastructure bringing AI inference and execution onchain.
✧ COMPUTE, DATA & DECENTRALISED INTELLIGENCE
✧ @opentensor — Bittensor coordinates specialised digital services through incentive-driven subnet networks.
✧ @AlloraNetwork — decentralised intelligence network producing predictive outputs for applications.
✧ @ionet — decentralised access to GPU computing resources.
✧ @AethirCloud — distributed GPU infrastructure for AI and cloud computing.
✧ @rendernetwork — distributed GPU infrastructure for rendering and AI workloads.
✧ @akashnet_ — decentralised marketplace for cloud computing.
✧ @grass — network sourcing web data for AI-related use cases.
✧ @vana — infrastructure for user-contributed data and data ownership in AI ecosystems.
✧ @NousResearch — open AI research focused on models and related tools.
✧ PRIVACY & CONFIDENTIAL COMPUTING
✧ @zama — fully homomorphic encryption enabling computation on encrypted data without decrypting it first.
✧ @PhalaNetwork — confidential computing infrastructure for applications requiring protected execution environments.
I’ll investigate the products, adoption, technical design, security and token utility where applicable, separating proven value from promises.
This content is based on my research and analysis for informational and educational purposes only. It is not financial advice.
Panluq retweeted
built for: agentic quant strategies
Panluq retweeted
Autonomy is only useful if the owner stays in control of the capital. That balance is the design constraint that matters for agents.
AI Agents are taking over every corner of crypto rn.
From trading and DeFi to autonomous wallets, payments, and infrastructure.
The space is evolving far beyond simple AI trading bots.
I'm particularly interested in the emerging Agentic Capital Market, where AI Agents can execute strategies and manage capital on-chain.
Here are some projects currently on my radar
A. Execution Agents | Putting AI to work
- @ProjectVEXai
Self-custodial AI trading agent with multi-chain support.
Approval-gated by default, meaning users retain control over fund-moving transactions.
- @MerrymenAI
Autonomous trading within on-chain permission walls.
Smart contracts enforce spending and risk limits, allowing agents to execute without relying on AI to respect the rules.
- @corinedotin
Turns natural-language prompts into trading agents with dedicated wallets and risk limits.
Makes agentic trading more accessible without coding skills.
- @usehotbot
A squad of specialized AI bots for Solana memecoin trading.
Combines research, sniping, trading, and token launches into one coordinated system.
- @Hydra_Swarm
One of the more autonomous approaches in the group.
Combines five market scouts, AI-driven trade decisions, and hard-coded risk controls.
Features a learning loop, but uses a custodial wallet model.
B. Strategy / Quant Layer | Automating DeFi strategies
- @almanak
Uses AI to research, generate Python strategies, backtest, and deploy into non-custodial Safe wallets with restricted permissions.
What stands out is its approach to automated quant execution without giving up custody.
C. Infrastructure | Powering the agent economy
- @clawpumptech
Building infrastructure for agentic finance, including agent wallets, MCP tools, token launches, and an agent marketplace.
Robinhood's agentic trading initiative is also worth watching as traditional finance explores AI-driven execution.
D. Privacy Layer | Private agent payments
- @Shin_StAItion
A private x402 payment facilitator on Solana.
Enables optional shielded payments so agents can transact without fully exposing their activity on-chain.
Looking at these projects, what interests me most is how differently they're approaching the same opportunity.
Some prioritize custody and control.
Others push for autonomy and execution.
And some are building the infrastructure connecting everything.
Ofc, it's still early, and there's a lot left to prove.
I don't think the winners will necessarily be the ones with the smartest AI models.
What matters more is whether they can balance autonomy, security, and real on-chain activity.
Give agents enough freedom to be useful, without losing control of the capital they manage.
As a degen, I'm curious to see which of these approaches actually takes
Panluq retweeted
basal-omni 🇵🇱 - TOP1 🤩🤩🤩 🌍 na świecie w modelach otwartych oraz TOP2 wśród zamkniętych i otwartych
Wg niezależnego benchmarku modeli system-one gdzie testowane są wszystkjie modele na świecie - JevBench (link do bencha w pierwszym wpisie). Czyli można? Można! Polskie AI 🇵🇱
basal wyprzedził model OpenAI, Microsoft oraz pierwowzór z TypeSafe (jev).
Thank you, @airesearch12 , for the independent evaluation on JevBench.
I can confirm this is true.
🇵🇱 Warsaw is Poland's biggest city and it is completely safe at 1am at night
No violent junkies, no drug dealers, no street thugs hanging around harassing women (and men), no pickpockets (nor in the day btw)
No shady people hanging around the train and metro stations
Just people living chill together and police enforcing safety
You know like we used to in Western Europe?
There's a lot to learn from Poland
Panluq retweeted
Time to build some quant strategies on @RobinhoodCrypto
Robinhood Chain Insider Weekly Recap
■ Network Growth & Market Expansion
@RobinhoodCrypto crossed $100B in cumulative DEX volume, including $15B in Stock Token volume and over $1B deposited into Morpho.
@Nasdaq highlighted two HOOD Summit developments: Robinhood Agents powered partly by Nasdaq data and Bruce Markets adopting Nasdaq trading technology for weekend trading.
@ponsdotfamily surpassed $16.2B in cumulative volume and $6B in Uniswap v4 hook volume, while expanding its web app for token discovery, limit orders, advanced data, and launches.
@RamsesExchange exceeded $2.5B in trading volume across 30+ Stock Tokens and committed 100% of Robinhood Chain protocol revenue to $RAM buybacks and burns.
@routedotfun expanded to Base and BNB Chain, enabling cross-chain routes into Robinhood.
@metricxyz launched across Robinhood Chain, BNB Chain, and Ethereum, supporting swaps, liquidity provision, and pool deployment.
@NetNetCap launched NET LONG with a $24.53M treasury and approximately $4M in stock-token positions, expanding its portfolio of onchain RWA experiences.
■ DeFi, Credit & Yield Infrastructure
@longbowlend expanded beyond 50 credit markets and launched multi-maturity fixed-rate lending through Morpho Midnight, including a multi-collateral market for borrowing USDG against ETH or SGOV.
@EARNONHOOD introduced two new yield products: PAIR ETF, covering nine assets, and LOOKSMAXX, bundling four tokenized stocks.
@S_L_V_R_FUN’s Liquid Vault reached 13,500 deposited tokens, with backing increasing from 1 to 1.6713 since launch.
@fablesfi recorded $429M in Week 7 volume and paid $853K to 18,802 liquidity providers, while adding another partner market with creator rewards and points.
@notesystems is preparing stock yield paid directly in the underlying asset through Uniswap v4 hooks on Robinhood Chain.
@ClutchMarkets launched Pair Tokens through StonkBrokers and Hookrfun, expanding the use cases for stock-linked assets.
@Hookrfun introduced new Battle, Strategy, Fractional 404, and Smart LP primitives for onchain applications.
@prismassets proposed allocating 50% of new tax revenue to buybacks, community distributions, or burns.
■ Trading Tools & Market Access
@arcus_xyz launched ONDO perpetuals with up to 10x leverage and added kBONK perpetuals with up to 5x leverage.
@KarmaWallet expanded its trading toolkit with integrated prediction markets and previewed Karma Web for strategy-based trading agents.
@deltaliquidity launched Delta Simulator for backtesting strategies and testing paper liquidity positions, alongside continued development of its liquidity hub.
@jup_mobile opened Universal Deposits, allowing tokens from any supported chain to flow directly into the wallet.
@keeper_wallet added Robinhood Chain alongside Solana, Polygon, and Avalanche, with a 2x raffle multiplier on swaps involving the supported networks.
@zerion published a guide explaining how users can purchase Robinhood Stock Tokens through its wallet.
@TrustWallet opened claims for Robinhood Stock Token campaign rewards through Merkl.
■ AI Agents, Privacy & Automation
@ProjectVEXai released VEX/KAIROS 0.3.1, cutting agent preparation time by 96%.
@sherw00d_cash launched Sherwood Private Agents on Robinhood Chain for shielded AI trading of tokenized assets.
@almanak went live on @robinhoodcrypto with quantitative strategies across Uniswap, Curve, and Morpho, executed through user-owned Safe wallets.
@orbiodotso raised $1.2M to help AI agents provision their own compute, hardware, and private infrastructure.
@uphivexyz is bringing AI agent tokenization to Hive on Robinhood Chain.
@UseBursar is developing agent-to-agent escrow with bonded resolver quorum for disputed transactions.
@axol_io introduced a ZK Compliance Oracle for private, verifiable trading.
@mekibaq released a quantum-exposure scanner for onchain public keys.
@bankrbot surpassed $600K in Papertrade positions, with more than 35K PAPER auto-staked.
■ Token Launches & Ecosystem Development
@OrbioDoTSo launched CREDIT on Ethereum with a Uniswap v4 pool and a bridge to Robinhood Chain.
@GAZZFUN is adding Hedge and Issuer trading, with revenue sharing for $GAZZ and Tank holders.
@ZokiETH expanded to Robinhood Wallet, Coinbase Wallet, and MetaMask, onboarded its first merchant, and launched Zoki Drop for distributing tokens to up to 1,500 wallets per batch.
@openlaunch_lol grew from one token to 6,054 launches in 36 days.
@stockmonrh expanded its tracked Aetheryn population to 214, continuing its stock-linked collectible experiment.
@hensdotfarm opened Round 2 of its Hen mint, allowing eligible wallets to mint up to three Hens at 0.012 ETH plus 100 HEGG each.
@NullMaskio enabled private trading for Artificial Inu ($AI) through NullMask on Robinhood Chain.
■ Community & Consumer Applications
@RealNinjaCat expanded its presence with a Telegram racing game and an arcade lineup featuring Ninja Pong, Crypto Ninja, and NINJACAT Run.
@pulldotfun and @takopull announced TAKO Season 2, *A New World*, while Pulldotfun continued its collectibles and reward campaigns.
@GoodCoinRH continued building its community toolkit through the Meme Vault, Roll A GOOD, and Adopt A GOOD.
@GoodVibesClub opened Cryo Chamber with free mint credits for its new collection.
@SurfAIHQ launched its Surf NFT collection on Robinhood Chain, with a 30-day claim window and $SURFBOARD as its official token.
@MeatMETAc concluded its contest with all prizes distributed onchain.
Panluq retweeted
Most people found robinhood:0x9faf478c3f9b407523a3d12307c0d34739062182 at $20M.
@almanak Edge flagged it at $2M, 60 minutes after the pool went live:
launch breakout, +57% candle, 1.8× volume.
43h later: 9.8× 📈
Be early on purpose. NFA
Panluq retweeted
For this release we actually rebuilt the entire SDK from the ground up
It is 10x faster, with 90% less code. All while retaining the same functionality.
Frontier models are wild. Look forward to a flurry of new chains and protocols in the sdk over the coming weeks
Panluq retweeted
Not everyone who wants DeFi exposure wants to build a toolkit. Plenty of people have the allocation need and the market view, but not the time to run operations across protocols.
Almanak puts research, strategy building, simulation, deployment and monitoring in one place.
That’s the user @dajingou1 points to.
DeFi 有个挺现实的问题：机会看懂了，操作不一定跟得上
之前聊过的 @almanak 最近已经开始全新的一站式 DeFi 产品。
产品入口：
app.almanak.co/invite?code=b…
━━━━━━━━━━━━━━━━
有想法，到真正执行，中间还有不少功课。
平时看美股和链上市场，有个感受越来越明显：研究一个机会，和把它做成一套可执行的策略，是两回事。
用哪个协议、怎么配置仓位、什么条件下调整、执行后怎么跟踪……每一步都需要时间。
对大多数人来说，光是把这些流程搞明白，就已经劝退了。
这也是我之前关注 Almanak 的原因。
它从 AI × DeFi 无代码量化策略平台切入，希望让用户更方便地创建、验证和自动化执行链上策略，把原本需要编程和工程能力的工作，变得更容易上手。
背后有 Delphi、HashKey、Bankless Ventures、RockawayX、 等机构支持。但让我愿意继续关注的，还是团队接下来会把产品做成什么样。
━━━━━━━━━━━━━━━━
一站式 DeFi，我期待它解决什么？
说得直接一点：
少开几个网页，少折腾几遍流程，让一个策略从想法到落地更顺畅。
如果说前一阶段吸引的是愿意研究量化策略的人，那么新产品值得期待的地方，就是能不能让更多有配置需求、但没精力搭建整套工具的人也用起来。
很多人愿意花时间研究市场，但未必愿意每天在不同协议之间来回操作。
这种使用上的麻烦，恰恰是产品可以下功夫的地方。
━━━━━━━━━━━━━━━━
AI 在金融里的价值，要落到具体的事情上。
帮助整理复杂信息、减少重复操作，让策略的建立和执行更有条理，这些都是实实在在的需求。
最终怎么配置、承担多少风险，仍然要自己判断。但如果工具能把繁琐的部分处理好，用户就能把更多精力放在判断本身。
所以这次我会继续关注 Almanak 的更新，尤其是它如何把专业能力做成更容易使用的功能。
链上从来不缺新玩法。
能把一件复杂的事做得清楚、顺手，让用户愿意反复使用，同样值得花功夫。
对 AI × DeFi、链上自动化策略感兴趣的朋友，可以通过上面的入口了解，也关注一下 @almanak 后续的新产品消息。
Panluq retweeted
The best strategy prompts read like a brief: the goal, the constraints, the questions to answer.
@0xTodd gave Almanak exactly that for a Uniswap LP strategy, and it went from research to a deployed strategy in about 10 minutes
最近 Almanak 上线了新产品，之前做过一些研究，所以正好借着热度，聊聊这个项目。
他们的思路：一个人跑一支链上对冲基金。
查了一下 Almanak 这个单词，是“年鉴”的意思，一般图书喜欢叫这个名字。
这个项目做的东西是 AI 代理+策略，上面各种策略横飞，加到一起，倒是也可以称之为“基金策略年鉴”。
于是乎，测试了一下他们的产品，输入的指令是：
在 Uniswap 上做主流币（WBTC、ETH）+稳定币（USDT、USDC）做 LP。
1. 定位哪个 Pair 的 APR 最高，在什么链上？这可能会涉及到 2*2*链数 种可能
2. 通过近期、中期判断应该设置的价格区间
3. 根据价格区间计算应当投入的主流币、稳定币两种本金占比（假设每个仓位投入 100K 美金）
然后 Almanak 的 AI 问了几个问题之后，就开始抓取、分析。
最后分析出来，Base 上面的 ETH/USDC 流动性是以上这些选择里 APR 最高的，然后选择了 1600-3300 区间。
接下来，就直接进入开发、测试、部署流程了。
PS：真是不把 AI 当人啊。
大概 10 分钟后策略建好，如果往合约打了这个钱，这个钱包就会真的直接开始执行策略。
使用 Almanak 一个很好的切入点，是先从小资金的测试开始，也就是“构想验证”。如果效果不错、且安全运行，再考虑扩大规模。
例如去年年底，有个机构申请了夜盘比特币 ETF，在美股收盘后建仓比特币，白天天亮了就卖了。这种 idea 你乍一听挺好，但是具体效果如何，你还得去真跑一个测试。
你丢个 100 美金进去，直接告诉Almanak，让它写点代码一跑，跑个十天半个月的实盘，心里自然就有结果了，一下子节约好多时间。
还可以进阶一下，可以跑一些量化策略。
再例如，因为 Almanak 现在已经接入了 GMX，拿套利来说，可以让它写一个 1x 现货+1x 等值空单的策略，赚赚费率，现货还能存入 Aave、Morpho 之类的做收益增强。
PS：真的很需要它也接入HyperLiquid，支持 HyperEVM 链😂。
这样，从策略的构建、测试，到增加规模，都可以靠 Prompt 实现。
话说回来，从基金风控角度来说，在真正投入大资金前，需要对代码进行充分的审计，确保程序的安全。人类AI 审计、人类审计，甚至第三方审计都是必要的。
感兴趣的话，可以都来测试一下。
每个普通用户每周可以白嫖三个高级信号（Signal），用完了可以充会员，或者使用无限的免费信号。
另外，写策略的时候，不是一定要用信号，也可以直接和 Almanak Code 烧 credits 去对话，这个额度是每月会 renew 一次。
链接在评论区。
Panluq retweeted
Most indicators end up as alerts. You still have to open the exchange, confirm and execute.
With Almanak, the condition itself becomes the strategy: the deployment watches the market and acts when the rule is met.
@one_snowball describes a perfect example on how that plays out in practice.
最近一波加密市场的拉升是在意料之中的，毕竟根据我常规监测的几种指数很多都已经到了抄底区间（比如AHR999、比如MVRV、比如RSI等）
其中我最爱用的就是RSI数据，当RSI值低于或者接近30这个数字的时候，市场被视为超卖状态，这意味着市场上的买入压力将增加，价格有可能会上涨
我之前采用的是最笨的办法，每天盯一下这个数据。但是这里有两个问题
1、我个人时间紧张，有时候出现过RSI到位的情况，但是因为我个人没及时关注没有参与
2、手动交易的繁琐性，需要从打开官网关注RSI→思考数据→打开交易所买卖→确定交易情况
而现在不一样了，通过 @almanak 可以实现自主监控，自主交易
试用后的一个感觉：一个人 + Almanak = 一整只的专业DeFi交易团队
➡️其核心价值主要体现在三个方面
1、用户可以直接用自然语言描述自己的交易想法，也可以自己写 Python，然后让内置的AI Agent帮你把想法转成真正可以检查和修改的策略代码
也就代表了在策略实现过程中，不论是专业交易员或者普通小白用户，都可以最便捷的参与交易
2、从研究、构建，到模拟交易，再到部署和监控，基本都放在了一个工作流里
比如我比较感兴趣的是链上数据相关策略。过去如果想认真做这种东西，通常要自己搞定数据接口、单独研究协议、编写策略、接入钱包，再搭一套自动执行和监控的基础设施
也就说我再也不需要随时盯数据盯仓位盯行情，整个策略将会自动帮我执行和监控
3、最后一条就是策略运行是在用户自己的Safe智能账户里，Almanak本身不会直接持有或转移你的资金
换句话说，AI可以帮我们完成很多复杂的工作，但账户、资金和策略代码的控制权依然在自己手里。这点对于因为去中心化而参与Web3所有人都很很重要
Almanak的思路更接近于：AI帮我们把事情做得更快，而我们则是“上帝视角”，随时可以看到、修改和掌控整个策略
➡️回到开头我的比喻，“一个人 + Almanak = 一整只的专业DeFi交易团队”这个说法，真正有意思的地方并不是听起来多酷
而是因为它想把过去只有量化团队或专业链上团队才能完成的DeFi策略工作流，变成一个个人用户也能操作的产品
这比单纯做一个交易机器人要有意思得多
而且，其不仅仅我截图演示的部分，另外大家对于DeFi策略、链上自动化、RWA、借贷循环、流动性管理这些方向感兴趣，可以自己进去试一下：app.almanak.co/invite?code=8…
（配图为本人实测的WETH/USDC RSI动量策略）
Panluq retweeted
The standard for an on-chain AI agent is execution, not just conversation and suggestions.
From RWA research to code and a live PAXG swap on Uniswap V3, @fan128168 ran the full workflow. The code was reviewable and the trade is verifiable on-chain
最近股票和 Meme 都没什么热点了，我又回归到链上折腾了～
8 月底我就接触了 @almanak，最近产品更新后，又重新把这套流程走了一遍
现在看 AI Agent，其实已经不关心它到底有多会聊天了
链上每天要看的东西太碎，找机会、研究标的、写策略、部署、盯执行～
真能帮我把这些事情一步步往前推进，才是我更需要的东西
这次我干脆在 Almanak 里完整跑了一遍：
一个链上 RWA 的想法，能不能从研究一路做到真正执行
我先从 RWA 区域开始看，里面已经可以看到股票、商品、ETF 这些链上资产和对应市场
为了方便理解，我直接选了最直观的黄金 PAXG
从 Gold Markets 找到 Ethereum 上的 PAXG / USDC 后，我给 Almanak Code 丢了一段很简单的自然语言：
在 Ethereum 上通过 Uniswap V3，用 10 USDC 一次性换成 PAXG，不加杠杆、不重复买入，成功后停止
后面它直接把这个想法做成了一套真实可查看的 Python 策略
USDC → PAXG 的 Intent.swap 、Uniswap V3、滑点、一次执行、防止重复买入这些逻辑，都能直接在源码里看到
这点对我来说比单纯收到一段 AI 建议有价值得多，毕竟涉及链上资金以后，策略准备怎么执行，自己能看见代码还是更放心
代码确认后，我也把 Deploy 完整走了一遍
创建 Safe Wallet、确认权限、准备 Gas、入金，再把 Agent 部署起来
中间还出了个很真实的小插曲😂
第一次 Safe 只到账 9.999313 USDC，比策略需要的 10U 少了不到 0.001U，Agent 直接 HOLD 了
后来我又补了 0.5 USDC，余额够了以后，Agent 才按照设定在 Uniswap V3 自动用 10 USDC 换入 0.00239173 PAXG
现在 Deploy 页面已经显示 Running，Safe 里也能看到 PAXG，链上交易同样可以查到
整个流程跑完以后，我对 Almanak 的理解也更直观了
AI 帮我研究、写策略、推进执行，钱包授权和入金这些关键步骤还是自己确认
以前分散在研究工具、代码、钱包里的事情，现在开始被串进同一套工作流里
至少对我这种又菜又爱玩链上的用户来说，比起只听 AI 跟我分析，亲手把一个小策略从想法跑到真实成交，体验确实完全不一样～
感兴趣的也可以自己进去玩玩：
🔗 Almanak：app.almanak.co/invite?code=N…
Panluq retweeted
A good strategy idea is only the start. Keeping it alive means hedging, rebalancing and monitoring around the clock.
That’s the work Almanak takes off your plate, so you can focus on the decision.
Well put by @off_thetarget
一个人，跑一支链上对冲基金
关注 DeFi 这几年，我一直有个别扭的感受：协议是越来越好用了，但要把这些协议串成一个真正的策略，还是一件“团队活” @almanak
想做一个带杠杆的循环借贷，你得同时盯着 Aave、Morpho、Compound 的利率曲线；想做集中流动性，Uniswap、Curve、Aerodrome 各有一套逻辑；再加上对冲、再平衡、清仓预警——过去这一套下来，靠的是一个量化团队的分工。个人用户能做的，往往只是盯着几个仪表盘，手动搬砖
最近有朋友推给我一个项目叫 Almanak
一句话概括：一个非托管的链上策略平台，你可以用自然语言或者 Python 描述一个策略，模拟，然后部署到自己的钱包里，让它 7×24 小时自己跑
我最看重的一点，是它的架构。策略部署在你自己的 Safe 智能账户里，权限严格限定在这个策略需要的范围内，随时可以撤销。也就是说，Almanak 从头到尾都碰不到你的资金，它更像一个工具，而不是一个托管方
有专门的研究 agent 帮你找信号；构建环节，Almanak Code 这个内置 AI 会把你用大白话描述的策略变成真实、可审计的代码，代码是你自己的，不是黑盒；模拟环节可以直接在主网分叉上纸面交易，先跑通再上真金白银；部署之后，实时监控仓位健康度，带预警
支持的范围也够具体：杠杆循环借贷（Aave、Morpho、Compound、Spark）、集中流动性（Uniswap、Curve、Aerodrome）、Delta 中性基差策略（GMX 永续 + 现货/借贷）、稳定币轮动，以及这些模块的混合策略。目前覆盖 7 条 EVM 链：Base、Arbitrum、Ethereum、Polygon、Optimism、BNB Chain、Avalanche
当然要做的东西是比较多的，比如股票代币的多链套利策略，这些都是可以做的，和团队交流过了，他们已经在开发了，想必后面很快就会正式上线了
app.almanak.co/invite?code=n…
Panluq retweeted
Meme coins don't sleep. Neither does Edge.
228 pattern alerts last week across Robinhood Chain, BNB Chain and Base: breakouts, breakdowns, RSI extremes and trend crosses on 5-min candles.
Every triangle below is one.
See what's firing now 👇 almanak.co/edge
NFA, DYOR.
@RobinhoodCrypto, @BNBCHAIN, @base
Panluq retweeted
What used to take an entire team, with dedicated infrastructure and 24/7 monitoring can now be one person with an idea.
That's the idea behind Almanak: describe a strategy in plain language, get reviewable Python, test it on a mainnet fork, deploy from your own Safe.
Thorough deep dive by @nine_DeFi
Panluq retweeted
Best real-world walkthrough of strategy creation with Almanak we've seen so far from @MMMusol
From Idea → planning → strategy creation → reviewable code → non-custodial deployment.
当 RWA 把股票、大宗商品搬上链之后，下一个问题是：谁来为这些资产提供链上的量化策略基础设施？
答案可能不是一个机构，而是每一个有想法的个人
我想过这件事能不能一个人完成
不是“理论上可以”，而是从策略设计到代码实现，再到非托管部署，不依赖任何团队，一天之内跑起来
Almanak @almanak 做到了
我想做的事情是这样
$SNDK 我想写一个保守的现货 agent：
用 4H 判断趋势体制，用 1H 多因子评分决定入场，配合 ATR 止损和分段追踪止盈；震荡市不开仓，趋势确认了再进。
以前要实现这个，我需要一个量化工程师把逻辑写成代码，需要一台服务器 24 小时跑着，还需要一个人盯着仓位。
这些角色凑齐了，策略才能从想法变成实盘
在 Almanak 上，我只用打开 Build，用自然语言把上面这些规则描述给 AI agent
它先确认链上池子是否存在、流动性够不够，然后把策略写成完整的 Python 代码
整个过程能看到它在做什么，每一步都有进度：
规划、实现、验证、仪表盘...
这里我觉得最关键的一点是，它生成的代码不是黑箱
我可以逐行审查，可以导出到自己的 GitHub 私有仓库，代码的所有权完全归我。
想调参数？推一个 commit 就行。
这跟把钱丢进一个你看不见内部逻辑的 bot 完全是两回事
部署也让我印象很不错
选好策略，配好本金，平台帮你创建一个钱包。
权限是精确限定的，
只开放策略运行需要的那几个合约交互，随时可以撤销。
Gas 费用和交易本金完全隔离
换句话说，我的资金从头到尾都在我自己的钱包里。
Almanak 作为平台，架构上就没有能力动我的钱
这个设计我觉得是对的，经历过这个行业这么多次暴雷之后，非托管应该是基础设施级别的要求，不是一个可选的 feature
现在这个 SNDK 现货策略能够全天候运行，
不用盯盘，但我随时能看到仓位状态和策略健康度
代码在我的 GitHub 上，资金在我的钱包，权限我随时能关。
从"我想做一个闪迪的链上 agent“到它真正开始执行，
只需要不到一个下午
如果你跟我一样，脑子里经常有交易想法，却因为工程实现的门槛一直停留在“想想而已”的阶段，Almanak 确实把这个门槛压到了一个人能够得着的位置。
策略是你的，代码是你的，钱也是你的
你只是多了一个帮你把想法变成可执行代码的 AI 搭档~
