You've been asking for this one... Now in preview: Codex in the ChatGPT mobile app. Start new work, review outputs, steer execution, and approve next steps, all from the ChatGPT mobile app. Codex will keep running on your laptop, Mac mini, or devbox.

#WeAreHiring# 我们正在招一位懂链上合约的实习生，主要负责 Perp DEX 带单、开单思路输出、社群交易互动。 要求： 有 Perp DEX / 链上合约使用经验，有自己的合约交易系统和策略 有真实开单经验，能讲清楚交易逻辑 熟悉 Meme、合约、市场热点和交易情绪 输出能力强，能写开单思路、行情点评、复盘 反应快，愿意在社群里互动、答疑、带节奏 待遇： 100U/天 每周 5 天，表现好可转正。 有交易记录、带单经验、个人账号或社群影响力者优先。 简历发送至：jack@ave.ai

Today we’re launching the OpenAI Deployment Company to help businesses build and deploy AI. It's majority-owned and controlled by OpenAI. It brings together 19 leading investment firms, consultancies, and system integrators to help organizations deploy frontier AI to production for business impact.

SpaceXAI will provide @AnthropicAI with access to Colossus 1, one of the world’s largest and fastest-deployed AI supercomputers, to provide additional capacity for Claude →

Fees in TON have dropped 6× — to nearly zero. Next step — Telegram replaces the TON Foundation as the driving force behind TON and becomes its largest validator. The focus shifts to tech superiority. New new dev tools, new performance upgrades. Timeline: 2-3 weeks.

预测市场需要自己的 CA，也需要自己的 Scan。 这两天我用工作之外的业余时间，开着 ChatGPT Pro 的 OpenAI Codex，花了几个小时 coding 出来一个小产品：MarketScan。 我做它的原因，是我越来越觉得 Polymarket、 这类产品，不应该只被理解成“预测市场交易所”。长期看，它们更像一种底层基础设施：提供市场、流动性、结算和手续费模型。它们的角色可能会更接近公链、DEX，或者某种预测市场协议。 在 crypto 里，公链和 DEX 的很多核心能力其实都在后端：资产发行、流动性池、swap、LP fee、协议收入等等。但用户真正高频接触的产品，不一定是协议自己的官方前端，而是大量第三方交易工具和钱包，比如 Wallet、Binance Wallet 等等。这些产品用更好的数据、更好的交互、更强的交易流程，把底层流动性包装成用户真正愿意使用的产品。 我觉得预测市场也可能会走向类似结构。Polymarket、 这类平台可能会越来越像市场和流动性的提供者，而上层会出现更多个性化前端：跟单工具、聪明钱工具、AI agent 交易工具、群内 bot、数据看板、垂直领域的预测市场终端。尤其在 AI 时代，每个人都可能用 AI 做出一套自己习惯的交易界面。前端会越来越定制化，而底层平台会越来越协议化。 MarketScan 本身就是这个判断的一个小例子。我不是一个完整开发团队，也不是全职在做这个项目，只是用业余时间和 AI coding 工具，把一个自己觉得应该存在的预测市场信息工具做了出来。 如果这是未来的方向，那预测市场就不应该只依赖“链接”作为主要的发现和传播格式。 为什么 meme 币能在社群里传播得这么快？一个很重要的原因是 CA。合约地址是纯文本，可以复制、转发、搜索、追踪，也能被钱包、bot、行情工具和交易前端直接识别。 相比之下，现在预测市场大多数时候还是通过“一个 Polymarket 链接”或者“一个 链接”来传播。链接在微信群、QQ群、Telegram、评论区这些场景里并不理想，可能被折叠、风控、屏蔽，也不方便之后再次搜索。更重要的是，依赖链接传播也会带来安全问题，因为恶意链接、投毒链接已经成为非常常见的钓鱼和诈骗方式。 但预测市场其实已经有类似的底层标识符：condition ID、event slug、market slug、market ID。这些标识符对应每一个 event 或 market。尤其是 condition ID，更接近每个最小可交易单元在链上的身份。 所以我做了 MarketScan。 MarketScan 目前集成了 Polymarket 和  ID。MarketScan 会自动识别输入内容，并定位到对应的 event 或 market。它会展示基础信息、赔率、Yes/No 数据、平台来源、market ID、condition ID、event slug、market slug、原始链接，以及其他常用标识符。重要信息都支持一键复制。 我也内置了排行榜功能。你可以浏览不同平台上的热门 event 和 market，比如最热门的话题、交易量最大的市场、热门盘口、概率变化明显的市场等等。所以它不只是一个查询工具，也可以作为发现活跃预测市场的入口。 如果说 Etherscan 让地址、合约和交易变得可读、可查、可传播，那么 MarketScan 想做的，就是让预测市场里的每一个 event、每一个 market、每一个 condition ID 也变得可读、可查、可传播。 我相信预测市场下一阶段的增长，不只会来自官方交易页面，也会来自更开放的数据层、更丰富的第三方前端，以及更适合社群传播和 AI-native 工作流的基础工具。 MarketScan 是我在这个方向上的一个小尝试。 🔗⬇️

1// Peg2Peg is live. a peer-to-peer marketplace for permanent on-chain digital art built on @Uniswap v4 hooks. no middleman. no custody. hash and provenance intact on every transfer. 🧵🦄

Prediction markets need their own CA, and they also need their own Scan. Over the past couple of days, I spent some of my spare time outside work building a small product with OpenAI Codex in ChatGPT Pro: MarketScan. The reason I built it is that I increasingly feel products like @Polymarket and @predictdotfun should not only be understood as “prediction market exchanges.” In the long run, they may look more like base-layer infrastructure: providing markets, liquidity, settlement, and fee models. Their role may be closer to a public chain, a DEX, or a prediction market protocol. In crypto, many core capabilities of public chains and DEXs live in the backend: asset issuance, liquidity pools, swaps, LP fees, protocol revenue, and so on. But the products users interact with most often are not always the official protocol frontends. They are third-party trading tools and wallets like GMGN, Axiom, OKX Wallet, Binance Wallet, etc. These products wrap underlying liquidity with better data, better UX, and stronger trading workflows. I think prediction markets may evolve in a similar direction. Platforms like Polymarket and may increasingly become providers of markets and liquidity, while the upper layer will have more personalized frontends: copy-trading tools, smart-money tools, AI agent trading tools, group bots, dashboards, and vertical prediction market terminals. Especially in the AI era, everyone may be able to build their own preferred trading interface with AI. Frontends will become more customized, while the underlying platforms become more protocol-like. MarketScan itself is a small example of this thesis. I am not a full development team, and I am not working on this full-time. I just used spare time and AI coding tools to build a prediction market information tool that I personally felt should exist. If this is the direction, prediction markets should not rely only on “links” as their main format for discovery and distribution. Why do meme coins spread so fast in communities? One important reason is the CA. A contract address is plain text. It can be copied, forwarded, searched, tracked, and recognized by wallets, bots, market tools, and trading frontends. By contrast, prediction markets today are still mostly shared as “a Polymarket link” or “a link.” Links are not ideal in WeChat groups, QQ groups, Telegram, comment sections, and similar environments. They can be collapsed, filtered, blocked, or simply inconvenient to search again later. More importantly, relying on links also creates security risks, since malicious or poisoned links have become a common phishing and scam vector. But prediction markets already have similar low-level identifiers: condition IDs, event slugs, market slugs, and market IDs. These identifiers correspond to each event or market. In particular, the condition ID is closer to the on-chain identity of the smallest tradable unit. That is why I built MarketScan. MarketScan currently integrates Polymarket and You can search for a market using three types of input: a link, a slug, or a condition ID. MarketScan automatically detects the input and locates the corresponding event or market. It shows basic information, odds, Yes/No data, platform source, market ID, condition ID, event slug, market slug, original links, and other common identifiers. Everything important is copyable with one click. I also built ranking features into it. You can browse trending events and markets across different platforms, such as the hottest topics, highest-volume markets, popular markets, and markets with significant probability movement. So it is not only a lookup tool, but also an entry point for discovering active prediction markets. If Etherscan made addresses, contracts, and transactions readable, searchable, and shareable, then MarketScan is my attempt to make every prediction market event, every market, and every condition ID readable, searchable, and shareable as well. I believe the next stage of prediction market growth will not only come from official trading pages. It will also come from more open data layers, richer third-party frontends, and better tools for community distribution and AI-native workflows. MarketScan is my small experiment in that direction. Welcome to try it out. Feedback is very welcome, especially on what information tools prediction markets still need.