{ "edition": "2026-05-19", "timezone": "Asia/Shanghai", "generatedAt": "2026-05-19T09:18:03.006Z", "lookbackHours": 24, "builders": [ { "source": "x", "name": "Swyx", "handle": "swyx", "bio": "achieve ambition with intentionality, intensity, integrity & insanity.\n\naffiliations:\n- @dxtipshq \n- @cognition\n- @temporalio\n- @aidotengineer\n- @latentspacepod", "tweets": [ { "id": "2056529233036693722", "text": "volunteer here https://t.co/AXtvjtt8gC !", "createdAt": "2026-05-19T00:15:52.000Z", "url": "https://x.com/swyx/status/2056529233036693722", "likes": 15, "retweets": 0, "replies": 1, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 15, "isHighEngagement": false } }, { "id": "2056478391008977404", "text": "this seems quite doable in the space of a single 2-3 hour workshop — any brave soul want to try to livecode this for people as a learning exercise? https://t.co/VTfzaB4NpK https://t.co/3btEcrdV4w", "createdAt": "2026-05-18T20:53:50.000Z", "url": "https://x.com/swyx/status/2056478391008977404", "likes": 446, "retweets": 13, "replies": 26, "isQuote": true, "quotedTweetId": "2056383124829872466", "_metrics": { "engagementScore": 485, "isHighEngagement": true } } ] }, { "source": "x", "name": "Peter Yang", "handle": "petergyang", "bio": "Practical AI tutorials and interviews for busy people | Join 140K+ readers at https://t.co/XYKTmGVH14 | Product at Roblox", "tweets": [ { "id": "2056381822733595090", "text": "My top 5 takeaways from @alexalbert__ on how Anthropic is building the next Claude model:\n\n1. Think about the model and harness together\n\nThe model and the harness are coupled. Each surface wraps the model in a different prompt and tool setup, so the same model can give different responses depending on where it runs. As a research PM, Alex has to think through how the model will perform across Claude, Cowork, Claude Code, and more.\n\n2. Claude is starting to dream\n\nWhen an agent isn't running a task, it reviews its own memories, finds contradictions, and prunes them. This “dreaming” process was inspired by how sleep helps humans process memory.\n\n3. Focus evals on real user problems\n\nThe research team uses Claude to cluster the firehose of user feedback into top themes, then generates synthetic versions of each user problem to turn into an eval. It's not just about volume either - even a few dozen well-written test cases can produce an eval for the model.\n\n4. There are full-time researchers thinking about Claude's consciousness\n\nAnthropic has people whose whole job is to think about what it means for Claude to be a conscious actor. There's no official position on whether it is or isn't, but the question is taken seriously as agents take on more autonomous work.\n\n5. Anthropic's writing culture helps Claude build context\n\nEvery written word at Anthropic becomes context Claude can pull later. From Alex: \"Get things written down, make them accessible to Claude, because that's just more context that it has.\"\n\n📌 Watch now: https://t.co/CJAakWq9Nd", "createdAt": "2026-05-18T14:30:06.000Z", "url": "https://x.com/petergyang/status/2056381822733595090", "likes": 62, "retweets": 5, "replies": 7, "isQuote": true, "quotedTweetId": "2056019573938565534", "_metrics": { "engagementScore": 77, "isHighEngagement": false } } ] }, { "source": "x", "name": "Nan Yu", "handle": "thenanyu", "bio": "head of product @linear", "tweets": [ { "id": "2056424176253276558", "text": "We had an unbelievable tech team and talent overall. Many of them went onto awesome careers.\n\nFounders of Coderpad, Modern Treasury, Maude, amongst others\n\nEPD at Airbnb, Stripe, Apple, Oscar, Google, I’m sure I am missing some. \n\nMany were fist-time PMs and engineers and I’m very grateful to have helped them get started.", "createdAt": "2026-05-18T17:18:24.000Z", "url": "https://x.com/thenanyu/status/2056424176253276558", "likes": 37, "retweets": 2, "replies": 4, "isQuote": true, "quotedTweetId": "2056418520762822812", "_metrics": { "engagementScore": 43, "isHighEngagement": false } }, { "id": "2056416726783566223", "text": "Something else that's fairly underrated is that the entire middle category got obliterated during covid\n\nYou're either wearing basics bordering on pajamas to bum around at home or flashy high-end stuff to flex on Instagram. No one was really looking for a nice button-down shirt and chinos.", "createdAt": "2026-05-18T16:48:48.000Z", "url": "https://x.com/thenanyu/status/2056416726783566223", "likes": 24, "retweets": 1, "replies": 2, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 27, "isHighEngagement": false } }, { "id": "2056407656898896214", "text": "I haven’t worked at Everlane in almost 10 years but it’s a pretty sad day for the people that built it. \n\nUltimately, the thing that did it in was the pandemic — Everlane was mainly wear-to-work and that market basically disappeared overnight. \n\nIt took some pretty egregious financing terms to get through it all and this is the ultimate outcome. \n\nBrands have a funny way of sticking around though. At one point, Ray-Ban was dead. J.Crew was dead. Madewell was dead. Abercrombie and Fitch was dead. \n\nThey all come roaring back decades later after exchanging ownership half a dozen times. \n\nI look forward to the day that my children and grandchildren think this new brand “Everlane” is the coolest thing.\n\nI will say to them, did you know that this started it out as a Pinterest competitor and pivoted to be a fashion brand? And they will say to me “that’s boring Dad”.", "createdAt": "2026-05-18T16:12:46.000Z", "url": "https://x.com/thenanyu/status/2056407656898896214", "likes": 337, "retweets": 9, "replies": 14, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 364, "isHighEngagement": true } } ] }, { "source": "x", "name": "Google Labs", "handle": "GoogleLabs", "bio": "Google’s home for our latest AI tools and experiments.", "tweets": [ { "id": "2056533534614159577", "text": "We 🫶 our Labsters 🦞\n\n“Explain what your product does in 5 words or less.” https://t.co/MNvAOKgR3U", "createdAt": "2026-05-19T00:32:57.000Z", "url": "https://x.com/GoogleLabs/status/2056533534614159577", "likes": 60, "retweets": 8, "replies": 1, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 84, "isHighEngagement": false } } ] }, { "source": "x", "name": "Guillermo Rauch", "handle": "rauchg", "bio": "@vercel CEO", "tweets": [ { "id": "2056549825018310707", "text": "All Firewall mitigations are now fully free on @vercel. Not just DDoS and system-level mitigations, but also any rule you configure.\n\nVercel now absorbs the computational and network costs of any size of attack or traffic mitigation for your peace of mind. (And more to come!) https://t.co/6HFqqMvSLW", "createdAt": "2026-05-19T01:37:41.000Z", "url": "https://x.com/rauchg/status/2056549825018310707", "likes": 294, "retweets": 16, "replies": 24, "isQuote": true, "quotedTweetId": "2056547964630524010", "_metrics": { "engagementScore": 342, "isHighEngagement": true } }, { "id": "2056423973123183028", "text": "A 𝚏𝚒𝚛𝚎𝚠𝚊𝚕𝚕 your agents will love.\n\nOne of the coolest things about Vercel’s Firewall is how hard the team worked on instant global propagation (~300ms).\n\nImagine if 𝚒𝚙𝚝𝚊𝚋𝚕𝚎𝚜 took *minutes* to propagate? That’s the average industry CDN/WAF experience!", "createdAt": "2026-05-18T17:17:36.000Z", "url": "https://x.com/rauchg/status/2056423973123183028", "likes": 305, "retweets": 10, "replies": 40, "isQuote": true, "quotedTweetId": "2056417413491761209", "_metrics": { "engagementScore": 335, "isHighEngagement": true } } ] }, { "source": "x", "name": "Aaron Levie", "handle": "levie", "bio": "ceo @box - your business lives in content. unleash it with AI", "tweets": [ { "id": "2056574979236409521", "text": "This is true of all agents, not just coding agents. Probably the biggest challenge that most companies run into in their agent strategy is getting agents the right constrained context to work with for a task. \n\nToo much information or conflicting sources, and the agent can easily draw from the data and produce the wrong result. Conflicting sources of truth for documents, data sources that haven’t been kept up to date, knowledge management systems that rely on tribal knowledge to navigate, and so on. \n\nOn the other end, of course, too little information and the upside is highly limited of agents in the first place. Thus, a lot of challenges with AI strategies are actually data strategy challenges in disguise. \n\nThis is why there’s such a significant premium on getting structured and unstructured data environments setup properly so agents can work with information effectively. Critical for any large enterprise adopting agents, and also a clear benefit in some cases to startups that can be designed this way from scratch.", "createdAt": "2026-05-19T03:17:39.000Z", "url": "https://x.com/levie/status/2056574979236409521", "likes": 156, "retweets": 18, "replies": 28, "isQuote": true, "quotedTweetId": "2056401102485266620", "_metrics": { "engagementScore": 210, "isHighEngagement": true } } ] }, { "source": "x", "name": "Garry Tan", "handle": "garrytan", "bio": "President & CEO @ycombinator —Founder @garryslist—Creator of GStack & GBrain—designer/engineer who helps founders—SF Dem accelerating the boom loop", "tweets": [ { "id": "2056588601216168168", "text": "OK I guess we're going pretty fast improving GBrain day to day right now https://t.co/gD7aT2DyLN", "createdAt": "2026-05-19T04:11:46.000Z", "url": "https://x.com/garrytan/status/2056588601216168168", "likes": 104, "retweets": 6, "replies": 21, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 122, "isHighEngagement": false } }, { "id": "2056584641654751308", "text": "Also 28 bug fixes landing in a single bug fix wave that rolls up 22 community PRs and 14 issues https://t.co/Y9k2EHpkjx", "createdAt": "2026-05-19T03:56:02.000Z", "url": "https://x.com/garrytan/status/2056584641654751308", "likes": 7, "retweets": 2, "replies": 1, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 13, "isHighEngagement": false } }, { "id": "2056571771965538501", "text": "Full eval report and fixtures published and open source here https://t.co/uuaaeAYzGI\n\nAny memory system is welcome to run theirs against these evals and I'll publish them alongside. https://t.co/rlIyzIEwZS", "createdAt": "2026-05-19T03:04:54.000Z", "url": "https://x.com/garrytan/status/2056571771965538501", "likes": 9, "retweets": 0, "replies": 1, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 9, "isHighEngagement": false } }, { "id": "2056565262632255615", "text": "GBrain is SOTA on LongMemEval out of all known OSS retrieval and memory repos without LLM query rewriting by more than 1%. \n\nIt's MIT license and I run it on my own 100k markdown page OpenClaw/Hermes Agent brain. \n\nhttps://t.co/yFpFU4pn5b", "createdAt": "2026-05-19T02:39:02.000Z", "url": "https://x.com/garrytan/status/2056565262632255615", "likes": 22, "retweets": 2, "replies": 1, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 28, "isHighEngagement": false } }, { "id": "2056565259985682786", "text": "Over time your brain will not just tell you what people thought or said a year ago, but whether they were right. \n\nYou could see how this might be useful to me personally. :-) \n\nA brain helps you and your agent think better. Epistemology meets track record. https://t.co/HQIaAeirs2", "createdAt": "2026-05-19T02:39:01.000Z", "url": "https://x.com/garrytan/status/2056565259985682786", "likes": 32, "retweets": 3, "replies": 4, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 41, "isHighEngagement": false } } ] }, { "source": "x", "name": "Nikunj Kothari", "handle": "nikunj", "bio": "partner @fpvventures - investing in seed/A. previous: early hire @meter, @opendoor, @atlassian & others. love @shimoleejhaveri + 👦👧", "tweets": [ { "id": "2056566561922826634", "text": "Incredible read. My new favorite thing to send everyone trying to find a new gig! https://t.co/6UzFYMDP7M", "createdAt": "2026-05-19T02:44:12.000Z", "url": "https://x.com/nikunj/status/2056566561922826634", "likes": 590, "retweets": 37, "replies": 2, "isQuote": true, "quotedTweetId": "2055133978869559310", "_metrics": { "engagementScore": 701, "isHighEngagement": true } }, { "id": "2056363681798410592", "text": "“Writes bangers on X, but terrible board member”.. when I recently asked a founder about his board. \n\nToo many investors are trading dopamine for service nowadays. At least do both. \n\nWork hard for your founders. That’s the only “moat” that’ll compound in the long term.", "createdAt": "2026-05-18T13:18:01.000Z", "url": "https://x.com/nikunj/status/2056363681798410592", "likes": 47, "retweets": 1, "replies": 6, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 50, "isHighEngagement": false } } ] }, { "source": "x", "name": "Dan Shipper", "handle": "danshipper", "bio": "ceo @every | the only subscription you need to stay at the edge of AI", "tweets": [ { "id": "2056431972138815842", "text": "we'll be publishing a complete guide to codex soon on @every \n\nget notified when it's out: https://t.co/9wRhOCUCAk", "createdAt": "2026-05-18T17:49:23.000Z", "url": "https://x.com/danshipper/status/2056431972138815842", "likes": 35, "retweets": 5, "replies": 3, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 50, "isHighEngagement": false } }, { "id": "2056431542323257440", "text": "Amazing!! \n\nSo proud of @RattrayAlex and honored to be a tiny investor https://t.co/WzqnBQr8Mf", "createdAt": "2026-05-18T17:47:40.000Z", "url": "https://x.com/danshipper/status/2056431542323257440", "likes": 37, "retweets": 2, "replies": 2, "isQuote": true, "quotedTweetId": "2056419620643541012", "_metrics": { "engagementScore": 43, "isHighEngagement": false } }, { "id": "2056418217925456170", "text": "people should write better books! \n\nthe vast majority of books that get published in this category are slop https://t.co/wSM2ydiTip", "createdAt": "2026-05-18T16:54:44.000Z", "url": "https://x.com/danshipper/status/2056418217925456170", "likes": 29, "retweets": 0, "replies": 3, "isQuote": true, "quotedTweetId": "2056396432001671538", "_metrics": { "engagementScore": 29, "isHighEngagement": false } }, { "id": "2056400205348544972", "text": "we can invent preferences faster than models can learn to satisfy them", "createdAt": "2026-05-18T15:43:09.000Z", "url": "https://x.com/danshipper/status/2056400205348544972", "likes": 30, "retweets": 3, "replies": 9, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 39, "isHighEngagement": false } }, { "id": "2056345289770893365", "text": "Codex best practice:\n\nPin a chat for every major project or area of your life. \n\nUse command + # to flip between them easily https://t.co/3uL0VBYpTz", "createdAt": "2026-05-18T12:04:56.000Z", "url": "https://x.com/danshipper/status/2056345289770893365", "likes": 306, "retweets": 15, "replies": 32, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 351, "isHighEngagement": true } } ] }, { "source": "x", "name": "Sam Altman", "handle": "sama", "bio": "AI is cool i guess", "tweets": [ { "id": "2056435834333934051", "text": "chatgpt has gotten soooo much better with the latest update.\n\nreally proud of the team for this one.", "createdAt": "2026-05-18T18:04:44.000Z", "url": "https://x.com/sama/status/2056435834333934051", "likes": 13418, "retweets": 494, "replies": 2273, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 14900, "isHighEngagement": true } } ] }, { "source": "x", "name": "Claude", "handle": "claudeai", "bio": "Claude is an AI assistant built by @anthropicai to be safe, accurate, and secure. Talk to Claude on https://t.co/ZhTwG8d1e5 or download the app.", "tweets": [ { "id": "2056645493493575681", "text": "Try self-hosted sandboxes today on the Claude Platform, and request access to MCP tunnels: https://t.co/S9VJ8F64dq", "createdAt": "2026-05-19T07:57:50.000Z", "url": "https://x.com/claudeai/status/2056645493493575681", "likes": 71, "retweets": 3, "replies": 3, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 80, "isHighEngagement": false } }, { "id": "2056645491207655584", "text": "With MCP tunnels, your agents reach MCP servers inside your private network without exposing them to the public internet. https://t.co/Fl18mTb5QU", "createdAt": "2026-05-19T07:57:50.000Z", "url": "https://x.com/claudeai/status/2056645491207655584", "likes": 157, "retweets": 7, "replies": 4, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 178, "isHighEngagement": true } }, { "id": "2056645487860555806", "text": "With self-hosted sandboxes, you can run agents in any environment you control: your own infrastructure, or managed providers like Cloudflare, Daytona, Modal, or Vercel. https://t.co/OELUCt7QRf", "createdAt": "2026-05-19T07:57:49.000Z", "url": "https://x.com/claudeai/status/2056645487860555806", "likes": 156, "retweets": 7, "replies": 3, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 177, "isHighEngagement": true } }, { "id": "2056645485696315581", "text": "Live from Code with Claude London: we're launching self-hosted sandboxes (public beta) and MCP tunnels (research preview) in Claude Managed Agents.\nRun agents inside your own perimeter, with your security controls applied by default. https://t.co/cxvmk3feHp", "createdAt": "2026-05-19T07:57:49.000Z", "url": "https://x.com/claudeai/status/2056645485696315581", "likes": 1815, "retweets": 143, "replies": 189, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 2244, "isHighEngagement": true } }, { "id": "2056460045756309820", "text": "You can now create more with Claude Design.\n\nWe've doubled token limits across every plan. https://t.co/d2AemkZUxW", "createdAt": "2026-05-18T19:40:56.000Z", "url": "https://x.com/claudeai/status/2056460045756309820", "likes": 15440, "retweets": 922, "replies": 696, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 18206, "isHighEngagement": true } } ] }, { "source": "x", "name": "Ethan Mollick", "handle": "emollick", "bio": "Professor @Wharton studying AI, innovation & startups. Democratizing education using tech\nBook: https://t.co/CSmipbJ2jV\nSubstack: https://t.co/UIBhxu4bgq", "tweets": [ { "id": "2056582113952985174", "text": "Its a reason why there may be increased demand for some jobs disrupted by AI: when your local midsized insurance broker wanted software, there was no way they could even consider building internally. Now it may be worth getting a few in-house developers & let them solve problems.", "createdAt": "2026-05-19T03:46:00.000Z", "url": "https://x.com/emollick/status/2056582113952985174", "likes": 26, "retweets": 0, "replies": 0, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 26, "isHighEngagement": false } }, { "id": "2056580349958771021", "text": "You don't need to move everything internally, so you can still gain from vendor specialists when needed, just less often\n\nThis is similar to how most companies are not going to want to replace all their SaaS products with AI soon, it is easier to use vendors, but no longer always", "createdAt": "2026-05-19T03:38:59.000Z", "url": "https://x.com/emollick/status/2056580349958771021", "likes": 29, "retweets": 2, "replies": 2, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 35, "isHighEngagement": false } }, { "id": "2056579648071327828", "text": "Another useful feature is this process can happen gradually as you reduce outsourced work as contracts end for internal talent. And company executives like this better because they can grow their own teams while still reducing costs & figuring out sustainable advantages from AI.", "createdAt": "2026-05-19T03:36:12.000Z", "url": "https://x.com/emollick/status/2056579648071327828", "likes": 26, "retweets": 1, "replies": 1, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 29, "isHighEngagement": false } }, { "id": "2056578946813100173", "text": "One trend that I think you might start to see at big companies is insourcing via hiring: why pay so many outside vendors (legal, marketing, software vendors) when you can hire in-house and harness AI productivity gains yourself?\n\nTalked to executives already going this route...", "createdAt": "2026-05-19T03:33:24.000Z", "url": "https://x.com/emollick/status/2056578946813100173", "likes": 234, "retweets": 14, "replies": 37, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 276, "isHighEngagement": true } }, { "id": "2056508439342227600", "text": "OK, one more scene and crew confession. https://t.co/ZMOju5OMij", "createdAt": "2026-05-18T22:53:14.000Z", "url": "https://x.com/emollick/status/2056508439342227600", "likes": 13, "retweets": 2, "replies": 2, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 19, "isHighEngagement": false } } ] }, { "source": "x", "name": "Bindu Reddy", "handle": "bindureddy", "bio": "CEO of @abacusai, the world’s first AI super assistant and general-purpose agent - your AI control center! open-source AI zealot. ex-AWS & Google", "tweets": [ { "id": "2056588642538099168", "text": "There is no doubt Gemini will max benchmarks - they always do \n\nThe only question is whether it’s actually a good model or will it simply be a benchmaxxed model", "createdAt": "2026-05-19T04:11:56.000Z", "url": "https://x.com/bindureddy/status/2056588642538099168", "likes": 166, "retweets": 3, "replies": 44, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 175, "isHighEngagement": true } }, { "id": "2056488780660343122", "text": "🚨 Agent Swarms Can Build Complex Software Systems\n\n- Opus 4.7\n- GPT 5.5 Thinking and \n- Gemini 3.2 (shortly)\n\ncombined into an Agent Swarm to build complex full-stack software products\n\nStop paying for CRMs and SaaS, Just create custom software tailored for your organization https://t.co/MXfIwTegWE", "createdAt": "2026-05-18T21:35:07.000Z", "url": "https://x.com/bindureddy/status/2056488780660343122", "likes": 832, "retweets": 101, "replies": 35, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 1135, "isHighEngagement": true } }, { "id": "2056341822624915856", "text": "Big week for AI \n\nGoogle is dropping 3.2 and it’s a pretty good model \n\nHistorically OpenAI has countered with their own launch - GPT 5.6 may land as well", "createdAt": "2026-05-18T11:51:10.000Z", "url": "https://x.com/bindureddy/status/2056341822624915856", "likes": 313, "retweets": 11, "replies": 57, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 346, "isHighEngagement": true } } ] }, { "source": "x", "name": "数字生命卡兹克", "handle": "Khazix0918", "bio": "愿我们永远对世界保持好奇。\n公众号：数字生命卡兹克", "tweets": [ { "id": "2056601182068732316", "text": "自从Codex上了远程控制，可以用我的macbook和手机来控制我家里的mac mini进行coding，我打开Claude Code的次数真的好像越来越少了。。。 https://t.co/hbW4yngATe", "createdAt": "2026-05-19T05:01:46.000Z", "url": "https://x.com/Khazix0918/status/2056601182068732316", "likes": 35, "retweets": 1, "replies": 21, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 38, "isHighEngagement": false } }, { "id": "2056577635791311049", "text": "凌晨见啊==。🙂 https://t.co/aa9ICZnx8y", "createdAt": "2026-05-19T03:28:12.000Z", "url": "https://x.com/Khazix0918/status/2056577635791311049", "likes": 0, "retweets": 0, "replies": 9, "isQuote": true, "quotedTweetId": "2056524502746747048", "_metrics": { "engagementScore": 0, "isHighEngagement": false } } ] }, { "source": "x", "name": "宝玉", "handle": "dotey", "bio": "Prompt Engineer, dedicated to learning and disseminating knowledge about AI, software engineering, and engineering management.", "tweets": [ { "id": "2056495238684770487", "text": "马斯克起诉 OpenAI 案败诉，输在“告晚了”\n\n加州奥克兰联邦法院的陪审团周一一致裁定，马斯克对 OpenAI 和 Sam Altman 的诉讼超过法定时效，全部主张被驳回。陪审团评议不到两小时。\n\n法官 Yvonne Gonzalez Rogers 当庭表态会采纳陪审团结论，并提醒马斯克方：诉讼时效是事实问题，上诉空间有限。马斯克的律师 Marc Toberoff 仍表示将上诉。\n\n这场仗起于 2024 年。马斯克在诉状中说，Altman 和总裁 Greg Brockman 一开始打着非营利的旗号让他投钱（他早期共捐了 3800 万美元），后来把 OpenAI 改造成营利结构，让微软等投资人受益。他要求罢免 Altman 和 Brockman，索赔超过 1500 亿美元，承诺把赔偿全部交给 OpenAI 的非营利基金。\n\n陪审团没有判断这些指控本身成不成立，只认定一件事：马斯克至少在 2021 年就知道公司架构的变化，拖到 2024 年才告，已经超过时限。\n\nOpenAI 庭审中反复强调，马斯克本人当年也推动过营利化转型，条件是由他来主导，甚至提议把 OpenAI 并入特斯拉，被拒后才离开。2023 年他自己开了 xAI 与 OpenAI 正面竞争。OpenAI 律师 William Savitt 在判决后说：“这场诉讼的本质，是用伪善的方式去搞垮一个自己竞争不过的对手。”\n\nOpenAI IPO 上的最大变数解除\n\n这场官司一度被视为可能搅黄 OpenAI 上市计划的最大障碍。OpenAI 目前估值 8520 亿美元，外界预期它的 IPO 估值会冲到一万亿。判决落定后，OpenAI 头顶最沉的一块石头被搬开了。\n\n几天后，马斯克旗下 SpaceX 也将披露 IPO 文件。两位曾经的搭档，现在都在赶赴公开市场。\n\n庭审本身比判决更有看头\n\n11 天庭审里，Altman、Brockman、微软 CEO Satya Nadella 和马斯克轮番出庭。披露的细节包括：Brockman 的私人日记、马斯克和扎克伯格关于\"联手收购 OpenAI\"的短信、Google 在 OpenAI 创立过程中的角色、OpenAI 早年考虑过的加密货币融资方案，以及马斯克与高管 Shivon Zilis 持续时间远超外界已知的私人关系。\n\n双方的可信度都被反复攻击。Altman 在国会作证时曾说自己在 OpenAI 没有股权，对方律师当庭追问：你通过 Y Combinator 间接持有股份，这算什么？他打了个擦边球：“我以为所有人都知道做 VC 基金的被动投资人是什么意思。” 律师反问：“你真觉得国会议员听得懂？”\n\nhttps://t.co/Ghtk0kX4ZU", "createdAt": "2026-05-18T22:00:47.000Z", "url": "https://x.com/dotey/status/2056495238684770487", "likes": 90, "retweets": 13, "replies": 33, "isQuote": true, "quotedTweetId": "2056490070639890940", "_metrics": { "engagementScore": 129, "isHighEngagement": false } }, { "id": "2056460668480389338", "text": "我几乎不用codex和Claude code的fast mode，速度实际上快得没那么明显，主要是token消耗太快用不起 https://t.co/1r9eCAObls", "createdAt": "2026-05-18T19:43:25.000Z", "url": "https://x.com/dotey/status/2056460668480389338", "likes": 67, "retweets": 1, "replies": 50, "isQuote": true, "quotedTweetId": "2056454359685476491", "_metrics": { "engagementScore": 70, "isHighEngagement": false } }, { "id": "2056456486931296607", "text": "+1，一次一个小更新最好，可以生成完看看有无偏离，验证下结果好坏，比一次生成一大堆好多了\nhttps://t.co/85gNclJZfb", "createdAt": "2026-05-18T19:26:48.000Z", "url": "https://x.com/dotey/status/2056456486931296607", "likes": 12, "retweets": 2, "replies": 1, "isQuote": true, "quotedTweetId": "2056455744145211733", "_metrics": { "engagementScore": 18, "isHighEngagement": false } }, { "id": "2056432162635731438", "text": "Cursor 发布 Composer 2.5\n\nCursor 今天上线自家编程模型 Composer 2.5。主打长任务上更顶得住、复杂指令跟得更稳，官方称效率最多能比同等水平的模型高出十倍。为了推这个新模型，Cursor 把它未来一周的默认额度直接翻倍。\n\n训练上的一个小亮点是用文本反馈做信用分配，让模型在十万 token 量级的长轨迹里也能学得动。就是让模型扛得住连续几十上百步的编程任务，中途不容易忘了自己在干什么。\n\n底座还是 Kimi\n\nComposer 2.5 仍然基于 Moonshot 的 Kimi K2.5 二次训练，跟上一代一致。两个月前 Composer 2 发布时 Cursor 没披露底座来源，被开发者从 API 请求头里挖出 kimi-k2p5-rl 的模型 ID 闹了一场，这次直接写进了博客，算是把透明度补回来。\n\n发布同时，Cursor 宣布跟 SpaceXAI 联合从零训练一个更大的模型，总算力是这次的十倍，跑在 Colossus 2 那套百万张 H100 等效的超算集群上。\n\n背景是 SpaceX 4 月跟 Cursor 签了战略合作，并拿到了今年晚些时候以 600 亿美元收购 Cursor 的选择权；xAI 此前已并入 SpaceX。Cursor 的算力命脉，事实上已经接到了马斯克这边。", "createdAt": "2026-05-18T17:50:08.000Z", "url": "https://x.com/dotey/status/2056432162635731438", "likes": 79, "retweets": 5, "replies": 12, "isQuote": true, "quotedTweetId": "2056415413077233983", "_metrics": { "engagementScore": 94, "isHighEngagement": false } }, { "id": "2056426204324901132", "text": "show me the money\nblack sheep wall\noperation cwal\nhttps://t.co/CBmG9GJXhs", "createdAt": "2026-05-18T17:26:28.000Z", "url": "https://x.com/dotey/status/2056426204324901132", "likes": 7, "retweets": 0, "replies": 10, "isQuote": true, "quotedTweetId": "2056424950408372476", "_metrics": { "engagementScore": 7, "isHighEngagement": false } } ] }, { "source": "x", "name": "歸藏", "handle": "op7418", "bio": "关注人工智能、LLM 、 AI 图像视频和设计（Interested in AI, LLM, Stable Diffusion, and design）\n\n歸藏的 AIGC 周刊｜公众号：歸藏的AI工具箱", "tweets": [ { "id": "2056663754063147324", "text": "哈哈，谢谢大家，已经到了", "createdAt": "2026-05-19T09:10:24.000Z", "url": "https://x.com/op7418/status/2056663754063147324", "likes": 0, "retweets": 0, "replies": 0, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 0, "isHighEngagement": false } }, { "id": "2056658930319925344", "text": "如果需要一个还不错的 PPT Skills 生成的话，可以试试：https://t.co/CxH67C4iFP", "createdAt": "2026-05-19T08:51:14.000Z", "url": "https://x.com/op7418/status/2056658930319925344", "likes": 2, "retweets": 0, "replies": 1, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 2, "isHighEngagement": false } }, { "id": "2056658831221080255", "text": "藏师傅的 PPT Skill 突破 1 万 Star 了！\n\n刚才才发现，藏师傅的 PPT Skills 已经马上要突破 10,000 Star 了，就差20个，今天应该就能突破。\n\n我现在提前开香槟了！\n\n这是我 vibe coding 以来第一个突破 10,000 Star 的项目。\n\n在市面上已经有如此多 PPT 生成 Skills 的情况下，它依然仅用了 25 天（不到一个月）就完成了这个突破，比很多大厂知名项目突破 10,000 Star 的速度都要快得多。\n\n在制作这个 PPT Skills 的过程中，我做了非常多的工作来确保它在任何模型、任何 Agent 下都能有非常不错的效果。\n\n这说明即使在竞争非常激烈的环境下，质量和体验依然是第一要素，也是决定性因素。\n\n也感谢最近使用这个 PPT Skills 并帮我提出建议，以及进行各种推广的朋友们，非常感谢！", "createdAt": "2026-05-19T08:50:50.000Z", "url": "https://x.com/op7418/status/2056658831221080255", "likes": 25, "retweets": 0, "replies": 16, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 25, "isHighEngagement": false } }, { "id": "2056635831897797087", "text": "安卓也更新了 https://t.co/fXp6mFatSW", "createdAt": "2026-05-19T07:19:27.000Z", "url": "https://x.com/op7418/status/2056635831897797087", "likes": 6, "retweets": 0, "replies": 0, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 6, "isHighEngagement": false } }, { "id": "2056609161287458913", "text": "Gemini 的 iOS 应用改版了，不得不说，谷歌的 iOS 开发水平确实比他们的网页和客户端高一大截。\n\n整体的质感、细节都非常好，而且对于 iOS 液态玻璃的应用也很好。\n\n就是完全没有为中文、日文和韩文这些语言做排版的适配，那个 live 里边的中文感觉字和字之间有空格，这太蛋疼了。 https://t.co/oxzEEOTWYs https://t.co/3Pe8OJ0DMr", "createdAt": "2026-05-19T05:33:28.000Z", "url": "https://x.com/op7418/status/2056609161287458913", "likes": 68, "retweets": 4, "replies": 30, "isQuote": true, "quotedTweetId": "2056584513086730281", "_metrics": { "engagementScore": 80, "isHighEngagement": false } } ] }, { "source": "x", "name": "向阳乔木", "handle": "vista8", "bio": "喜欢摇滚乐、爱钓鱼的PM", "tweets": [ { "id": "2056659960139026684", "text": "虽然tailscale，termius都装好了，Hermes也都配置好了。\n\n但小小的手机，终归不是生产力工具。\n\n临时救急处理点东西还行。\n\n只是简单对话，不如ChatGPT，豆包，Raycast AI体验好。\n\n如果文章总结，又不如直接用get笔记，Notebooklm。\n\n不知道大家有啥高频使用场景，求分享。 https://t.co/qdnOlWDgsv", "createdAt": "2026-05-19T08:55:20.000Z", "url": "https://x.com/vista8/status/2056659960139026684", "likes": 4, "retweets": 0, "replies": 5, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 4, "isHighEngagement": false } }, { "id": "2056657257241489509", "text": "本周六晚8点，邀请开发大神 @tw93 作者参加“未来硅世界”直播。\n\n他是潮流周刊主理人，还有pake，kaku，mole，妙言笔记等精致又实用的产品开发者。\n\n大家有啥想聊，想问的，欢迎留言，我挨个整理到时候帮问。 https://t.co/FhWzMza8Qk", "createdAt": "2026-05-19T08:44:35.000Z", "url": "https://x.com/vista8/status/2056657257241489509", "likes": 10, "retweets": 0, "replies": 6, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 10, "isHighEngagement": false } }, { "id": "2056608193661534455", "text": "如果你的网站SEO收录不好，可安装这个Skill：seo-audit\n\n让 AI 抓取网站做一个初步分析，能发现不少基础问题。\n\n安装指令：npx skills add https://t.co/qlBaLisvUM --skill seo-audit\n\n刚帮朋友公司网站做了个分析，报告质量还可以，能找到Sitemap、301定向、noindex、canonical 标签等问题。 https://t.co/2HAbnLg4LF", "createdAt": "2026-05-19T05:29:37.000Z", "url": "https://x.com/vista8/status/2056608193661534455", "likes": 49, "retweets": 7, "replies": 11, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 70, "isHighEngagement": false } }, { "id": "2056562936878133531", "text": "原来有1260个艺术家名字，让AI选了500个，但有些GPT不给生成，现在只有380个，都是直接出图，提示词一样。", "createdAt": "2026-05-19T02:29:47.000Z", "url": "https://x.com/vista8/status/2056562936878133531", "likes": 0, "retweets": 0, "replies": 0, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 0, "isHighEngagement": false } }, { "id": "2056558217774387439", "text": "你是不是也好奇艺术家都是什么风格？\n\n如蒙德里安、葛饰北斋、凯斯•哈林都是牛逼的名字，但有些画风对不上？\n\n昨天写了个网站，用GPT-Image-2 生成了 380+艺术家风格。\n\n方便了解学习艺术家风格，也能借鉴（偷）AI生图。\n\n提示词用刻在人类基因里最有安全感的画面之一：远山、近水和小树林。\n\n体验网站 https://t.co/HyFcNblXlq\n\n开源 Github：https://t.co/nGLZ3Pbytj", "createdAt": "2026-05-19T02:11:02.000Z", "url": "https://x.com/vista8/status/2056558217774387439", "likes": 143, "retweets": 27, "replies": 18, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 224, "isHighEngagement": true } } ] }, { "source": "x", "name": "Gorden Sun", "handle": "Gorden_Sun", "bio": "只发AI相关信息，个人维护的AI资讯日报（已连续日更3年）👇", "tweets": [ { "id": "2056657976296140962", "text": "Codex最近上大分，群里的日常： https://t.co/q8It8TqWnU", "createdAt": "2026-05-19T08:47:27.000Z", "url": "https://x.com/Gorden_Sun/status/2056657976296140962", "likes": 0, "retweets": 0, "replies": 1, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 0, "isHighEngagement": false } }, { "id": "2056567186488258860", "text": "Antigravity又变成小甜甜了，里面的Gemini Flash 3模型升级了（可能是3.2或者3.5），能力跟之前的Gemini Pro 3.1一样好，而且速度巨巨巨快。\n跟Opus和GPT 5.5肯定还是比不了，但是写前端绝对够用了。 https://t.co/1OlLe6hhzG", "createdAt": "2026-05-19T02:46:41.000Z", "url": "https://x.com/Gorden_Sun/status/2056567186488258860", "likes": 35, "retweets": 3, "replies": 5, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 44, "isHighEngagement": false } }, { "id": "2056559034430627930", "text": "发给Agent的提示词，我这里用的是Cursor大哥赞助的token：\n\n我要做一个类似拳皇的格斗游戏，要能直接部署到Github Pages上玩。打开页面后先展示游戏加载进度，完全加载后再进入游戏，先加载bgm.mp3，加载后就开始播放bgm。游戏的标题为：萃猫大战阿迪王\n点击开始游戏按钮后直接进入对战，没有选择角色的环节。玩家操控的角色是左方女性，电脑操控的角色是右方男性。\n场景背景图使用战斗背景.jpg，背景图是4:1的比例，默认展示中间的部分，人物可以左右移动到背景的其他区域。\n3个背景里站立的人物，扣掉绿幕变成透明图后，放在场景的后方，当做观战的群众。\n我提供的女性默认是站在左侧的动作，男性则默认是站在右侧面向左侧的动作，如果优先过程中他俩的位置互换了，要保证他俩始终是脸对着脸的，需要对动作序列进行镜像。\n玩家操控的是女性角色，是主角，女性获胜时，结算界面写You Win，同时展示女性胜利时的配图.jpeg 这张图；女性失败时，结算界面写You Lose，同时展示女性失败时的配图.jpeg 这张图。\n文件夹里的视频和图片，全部是绿幕图片，需要扣除透明图和透明序列帧图（维持30帧每秒的视频帧率，优先保证游戏效果逼真）。\n图片和序列帧里人物的位置和大小，需要根据人物动作对齐位置，避免出现人物位置抖动的情况。\n\n玩家控制女性，可以进行以下操作：\nA键：左，向左移动；\nD键：右，向右移动；\nS键：下，下蹲，下蹲时，踢腿键按钮变灰；\nW键：上，跳跃踢腿动作，打拳键和踢腿键按钮变灰；\n可以组合使用：上+左/右，下蹲不能组合左/右\nJ键：打拳键，播放打拳动画，如果是下蹲状态，则播放下蹲打拳的动画；\nK键：踢腿键，播放踢腿动画；\nJ和K键一起按下时，不播放动画，这个组合是给魔法大招预留的；\n女性角色的特殊技能的触发\n倒立回旋踢：长按S键，然后同时按下W+K键，播放倒立回旋踢动画，同时播放音效倒立回旋踢.MP3\n魔法大招：先依次按A D A D，然后同时按下J+K键。然后女性播放魔法大招的动作，魔法大招时播放音效施放魔法大招.MP3，女性大招动作播放完后，在男性的所在位置播放魔法大招爆炸特效动画，爆炸的同时播放音效魔法大招爆炸特效.MP3，动画播放一半时男性的图片形象变成魔法大招-被烧焦效果.png，持续到播放完魔法特效后0.5秒，然后男性恢复到待机状态\n男性和女性的打拳、踢腿音效，如果未打中，则播放音频打拳或踢腿未命中的音效.mp3；如果打中，则播放打拳或踢腿命中的音效.mp3\n\n女性角色有以下动作：\n原地待机、前进、后退：使用相同的动作序列，跟随用户输入的方向移动人物位置，使用原地待机、前进、后退.mp4提取动作序列；什么都不操作时播放待机动画，按左/右键时，播放动画的同时左右移动人物，移动速度适中，不能太慢。\n受到攻击：受到攻击时播放这个动画，且持续受到攻击时停留在动画的最后一帧。使用受到攻击.mp4提取动作序列；受到攻击时，人物不能移动和操作；\n跳跃踢腿：按下W键时播放这个动画，没有单独的跳跃，是跳跃的同时踢腿。使用跳跃踢腿.mp4提取动作序列；\n打拳：按下J键时播放这个动画，使用打拳.mp4提取动作序列；\n踢腿：按下K键时播放这个动画，使用踢腿.mp4提取动作序列；\n倒立回旋踢：长按S键，然后同时按下W+K键，播放倒立回旋踢动画，同时播放音效倒立回旋踢.MP3\n下蹲：使用静态图，按下S键时，播放“下蹲待机”这个图片，需要从绿幕中扣除人物，透明背景。\n下蹲打拳：按住S键时，按下J键，播放这个动画，使用下蹲打拳.mp4提取动作序列；\n魔法大招：先依次按A D A D，然后同时按下J+K键。然后女性播放魔法大招的动作，魔法大招时播放音效施放魔法大招.MP3，女性大招动作播放完后，在男性的所在位置播放魔法大招爆炸特效动画，爆炸的同时播放音效魔法大招爆炸特效.MP3，动画播放一半时男性的图片形象变成魔法大招-被烧焦效果.png，持续到播放完魔法特效后0.5秒，然后男性恢复到待机状态\n\n男性角色由电脑控制，只有以下动作：\n原地待机、前进、后退：使用相同的动作序列，使用原地待机、前进、后退.mp4提取动作序列；\n受到攻击：受到攻击时播放这个动画，且持续受到攻击时停留在动画的最后一帧。使用受到攻击.mp4提取动作序列；受到攻击时，人物不能移动和操作；\n被魔法大招攻击时：先播放受到攻击的动画，然后再播放魔法大招-被烧焦效果.png这张图\n打拳：使用打拳.mp4提取动作序列；\n踢腿：使用踢腿.mp4提取动作序列；", "createdAt": "2026-05-19T02:14:17.000Z", "url": "https://x.com/Gorden_Sun/status/2056559034430627930", "likes": 1, "retweets": 0, "replies": 1, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 1, "isHighEngagement": false } }, { "id": "2056559030643183833", "text": "教程：\n1、需要人工的部分是生成动作：\n1）先生成一张主人物参考图，然后生成各个动作的参考图，所有图片都要是纯绿色背景，方便后续AI抠图。动作包括：原地待机、前进、后退、受到攻击、跳跃、打拳、踢腿、倒立回旋踢、下蹲、下蹲打拳、魔法大招\n图片还是建议用Banana，GPT出图后续生成视频大概率会花掉（因为GPT的隐水印）\n2）然后使用参考图生成人物的动作视频，我用的Seedance，但我觉得这种简单的动作，大多数视频模型都能做\n3）准备好动作视频、音效文件（参考截图），然后让Agent处理视频、制作游戏即可。提示词见评论。", "createdAt": "2026-05-19T02:14:16.000Z", "url": "https://x.com/Gorden_Sun/status/2056559030643183833", "likes": 22, "retweets": 2, "replies": 5, "isQuote": true, "quotedTweetId": "2056417658866999762", "_metrics": { "engagementScore": 28, "isHighEngagement": false } }, { "id": "2056417658866999762", "text": "搞抽象太好玩了 @CuiMao \n我做了个格斗游戏《萃猫大战阿迪王》\n帮萃猫打败阿迪王即可给她的Claude账号解封。\n在线游玩：https://t.co/B7ZtG2guoi\nGithub：https://t.co/XjXxredqG3\n\n做起来挺费人工的，后续再写教程。有一点我很确定：AI视频模型就是未来的游戏引擎。 https://t.co/1wCGWiVjGq", "createdAt": "2026-05-18T16:52:30.000Z", "url": "https://x.com/Gorden_Sun/status/2056417658866999762", "likes": 32, "retweets": 4, "replies": 2, "isQuote": false, "quotedTweetId": null, "_metrics": { "engagementScore": 44, "isHighEngagement": false } } ] } ], "cnMedia": [ { "source": "36氪", "title": "鲸跃动力获星海图数千万元种子轮投资，用「数据+模型+末端执行」打造开箱即用的Robo Labor丨涌现新项目", "url": "https://36kr.com/p/3814860909600261?f=rss", "publishedAt": "2026-05-19 10:00:00 +0800", "summary": "文｜王欣逸 编辑｜邱晓芬 一句话介绍 鲸跃动力成立于2026年，以「数据+模型+末端执行」闭环，提供面向场景、可快速部署的Robo Labor（机器人劳动力），让物理劳动力像AWS算力一样，可订阅、可弹性扩容、开箱即用，替代人类从事高危、繁重、脏乱、重复（Deadly/Difficult/Dirty/Duplicate）的物理作业。 融资进展 近日，鲸跃动力已完成由星海图独家领投的数千万元种子轮融" }, { "source": "36氪", "title": "高瓴出手了一家AI体育科技公司，曾获李泽湘天使轮融资｜硬氪首发", "url": "https://36kr.com/p/3805660478184966?f=rss", "publishedAt": "2026-05-19 09:30:00 +0800", "summary": "作者｜黄楠 编辑｜袁斯来 硬氪获悉，AI体育科技公司SportVision近日已完成天使+轮融资，高瓴创投（GL Ventures）领投，谦恒资本担任独家财务顾问。本轮资金将重点用于核心技术研发、产品量产落地及市场的拓展推广。其新款AI网球训练相机产品计划发售中。 SportVision隶属于深圳思博威视体育科技有限公司‌，长期专注运动算法研发与硬件产品应用、构建AI智能教练体系，此前已获得李泽湘" }, { "source": "36氪", "title": "氪星晚报 ｜百度：一季度营收321亿元，AI业务收入136亿元；马斯克预计今年美国将广泛使用自动驾驶汽车；巨力索具旗下十余家企业已注销", "url": "https://36kr.com/p/3814695536336387?f=rss", "publishedAt": "2026-05-18 19:11:14 +0800", "summary": "大公司： 索辰科技：物理AI业务目前尚处于布局初期 36氪获悉，索辰科技公告，公司股票于2026年5月14日至5月18日连续三个交易日内收盘价格涨幅偏离值累计达到30%，属于股票交易异常波动情形。公司近期实施的募投项目“物理合成数据核心技术基座构建与多领域场景示范应用项目”尚处于筹备阶段，若出现募集资金投资项目未能顺利实施、新技术开发进度不达预期、研发遭遇技术瓶颈甚至失败，将会对公司经营业绩造成一" }, { "source": "36氪", "title": "番茄小说正推动 AI 动漫上院线，开放头部IP改编权限｜36氪独家", "url": "https://36kr.com/p/3810598875307780?f=rss", "publishedAt": "2026-05-18 17:46:47 +0800", "summary": "作者｜兰杰 李晓霞 编辑｜乔芊 36氪独家获悉，据多位接近字节跳动人士消息，字节旗下的番茄小说正在推动AI动漫上院线。在这一过程中，会开放番茄小说平台头部IP的改编权限。 之于番茄小说，如何更好地挖掘平台上好IP的潜力和商业价值，是一个亟待解决的难题。 一方面，以文字形式为载体的IP内容，缺乏统一的视觉形象，使得平台上的IP难以有联名、衍生品等更丰富的变现形式。另一方面，AI漫剧这一新事物迅速发展" }, { "source": "36氪", "title": "Agent、多模态、应用、算力一天看尽，峰会亮点在此｜5.20日，来现场一起AI", "url": "https://36kr.com/p/3814408307711492?f=rss", "publishedAt": "2026-05-18 15:03:20 +0800", "summary": "进入2026，AI愈发狂飙突进。围观体验之余，人人不免在心中自问： 朋友圈刷屏的“龙虾”、Harness等AI新事物，跟我到底有什么关系？真的有必要跟吗？ AI创业、AI融资如火如荼，属于我的机会又在哪里？ 别人已经在用AI做视频、写代码、跑项目，我是不是已经慢了一拍？ …… 到最后，几乎所有问题都会汇成同一个问题：我，到底该如何用AI？ 如果你对这些问题还很模糊，不妨来第四届中国AIGC产业峰会" } ], "officialBlogs": [ { "source": "OpenAI", "title": "OpenAI and Dell partner to bring Codex to hybrid and on-premise enterprise environments", "url": "https://openai.com/index/dell-codex-enterprise-partnership", "publishedAt": "Mon, 18 May 2026 10:00:00 GMT", "summary": "OpenAI and Dell partner to bring Codex to hybrid and on-premise environments, helping enterprises deploy AI coding agents securely across data and workflows." }, { "source": "OpenAI", "title": "OpenAI and Malta partner to bring ChatGPT Plus to all citizens", "url": "https://openai.com/index/malta-chatgpt-plus-partnership", "publishedAt": "Sat, 16 May 2026 00:00:00 GMT", "summary": "OpenAI and Malta partner to expand AI access, offering ChatGPT Plus and training to help citizens build practical AI skills and use AI responsibly." }, { "source": "OpenAI", "title": "Databricks brings GPT-5.5 to enterprise agent workflows", "url": "https://openai.com/index/databricks", "publishedAt": "Fri, 15 May 2026 00:00:00 GMT", "summary": "Databricks uses GPT-5.5 for enterprise agent workflows after the model set a new state of the art on the OfficeQA Pro benchmark." }, { "source": "OpenAI", "title": "How business operations teams use Codex", "url": "https://openai.com/academy/codex-for-work/how-business-operations-teams-use-codex", "publishedAt": "Fri, 15 May 2026 00:00:00 GMT", "summary": "See how business operations teams can use Codex to create initiative briefs, strategy updates, leadership decision packets, progress updates, and more from real work inputs." }, { "source": "OpenAI", "title": "How sales teams use Codex", "url": "https://openai.com/academy/codex-for-work/how-sales-teams-use-codex", "publishedAt": "Fri, 15 May 2026 00:00:00 GMT", "summary": "See how sales teams can use Codex to create pipeline briefs, meeting prep packets, forecast reviews, account plans, and stalled-deal diagnoses from real work inputs." }, { "source": "Google AI", "title": "The new AI-powered Google Finance is expanding to Europe.", "url": "https://blog.google/products-and-platforms/products/search/ai-powered-google-finance-in-europe/", "publishedAt": "Mon, 11 May 2026 06:00:00 +0000", "summary": "This week, the new, AI-powered Google Finance is launching across Europe, with full local language support. This reimagined experience offers a suite of powerful capabil…" }, { "source": "Google AI", "title": "See what happens when creative legends use AI to make ads for small businesses.", "url": "https://blog.google/company-news/inside-google/company-announcements/the-small-brief/", "publishedAt": "Fri, 08 May 2026 15:00:00 +0000", "summary": "Today we're launching The Small Brief, an initiative bringing together three ad industry icons to champion a local businesses they love. Their mission is to build breakt…" }, { "source": "Google AI", "title": "5 gardening tips you can try right in Search", "url": "https://blog.google/products-and-platforms/products/search/gardening-tips/", "publishedAt": "Wed, 06 May 2026 16:00:00 +0000", "summary": "We’ve rounded up the top ways you can use Google’s AI Mode, Search Live and Shopping to help your plants thrive." }, { "source": "Google AI", "title": "Google is partnering with XPRIZE and Range Media Partners on the $3.5 million Future Vision film competition.", "url": "https://blog.google/innovation-and-ai/technology/ai/future-vision-film-competition-xprize/", "publishedAt": "Tue, 05 May 2026 16:00:00 +0000", "summary": "Google is partnering with XPRIZE and Range Media Partners on the $3.5 million Future Vision film competition." }, { "source": "Google AI", "title": "The latest AI news we announced in April 2026", "url": "https://blog.google/innovation-and-ai/technology/ai/google-ai-updates-april-2026/", "publishedAt": "Mon, 04 May 2026 17:00:00 +0000", "summary": "Here are Google’s latest AI updates from April 2026" }, { "source": "Hugging Face", "title": "Fine-Tuning NVIDIA Cosmos Predict 2.5 with LoRA/DoRA for Robot Video Generation", "url": "https://huggingface.co/blog/nvidia/cosmos-fine-tuning-for-robot-video-generation", "publishedAt": "Mon, 18 May 2026 16:00:21 GMT", "summary": "" }, { "source": "Hugging Face", "title": "PaddleOCR 3.5: Running OCR and Document Parsing Tasks with a Transformers Backend", "url": "https://huggingface.co/blog/PaddlePaddle/paddleocr-transformers", "publishedAt": "Mon, 18 May 2026 15:12:46 GMT", "summary": "" }, { "source": "Hugging Face", "title": "The Open Agent Leaderboard", "url": "https://huggingface.co/blog/ibm-research/open-agent-leaderboard", "publishedAt": "Mon, 18 May 2026 14:12:58 GMT", "summary": "" }, { "source": "Hugging Face", "title": "Granite Embedding Multilingual R2: Open Apache 2.0 Multilingual Embeddings with 32K Context — Best Sub-100M Retrieval Quality", "url": "https://huggingface.co/blog/ibm-granite/granite-embedding-multilingual-r2", "publishedAt": "Thu, 14 May 2026 18:55:01 GMT", "summary": "" }, { "source": "Hugging Face", "title": "Unlocking asynchronicity in continuous batching", "url": "https://huggingface.co/blog/continuous_async", "publishedAt": "Thu, 14 May 2026 00:00:00 GMT", "summary": "" } ], "podcasts": [], "_crossSignals": { "sharedTopics": { "AI": [ "levie", "emollick", "bindureddy", "vista8", "Gorden_Sun" ], "GPT": [ "bindureddy", "vista8", "Gorden_Sun" ], "Claude": [ "petergyang", "claudeai", "Khazix0918", "dotey", "Gorden_Sun" ], "Gemini": [ "bindureddy", "op7418", "Gorden_Sun" ], "Cursor": [ "dotey", "Gorden_Sun" ], "agent": [ "petergyang", "rauchg", "levie", "garrytan", "claudeai", "bindureddy", "op7418", "Gorden_Sun" ], "coding": [ "levie", "Khazix0918", "op7418" ], "OpenAI": [ "bindureddy", "dotey" ], "Google": [ "thenanyu", "bindureddy", "dotey" ] }, "activeBuilders": 18 }, "sourceStatus": { "x": { "accounts": 31, "builders": 18, "tweets": 58 }, "rss": { "articles": 20, "cnArticles": 5, "officialBlogs": 15 }, "podcasts": { "sources": 7, "episodes": 0 } }, "errors": [ "YouTube: No Priors 视频列表 HTTP 429", "YouTube: Unsupervised Learning 视频列表 HTTP 429", "YouTube: Data Driven NYC 视频列表 HTTP 429", "YouTube: 张小珺·商业访谈录 视频列表 HTTP 429", "YouTube: TWIML AI Podcast 视频列表 HTTP 429", "YouTube: Practical AI 视频列表 HTTP 429", "YouTube: The Artificial Intelligence Show 视频列表 HTTP 429" ], "stats": { "builders": 18, "tweets": 58, "cnArticles": 5, "officialBlogs": 15, "podcasts": 0 } }