与「recapความเอิร์ทมิกซ์」相关的搜索结果

parameter golf was a blast. 2,000+ submissions. 1,000+ verified github accounts. ideas ranging from quantization and depth recurrence to TTT LoRA, SSMs, H-nets, JEPA, and more. autoresearch made iteration dramatically faster — and led to emergent bulletin boards, issue threads, unofficial leaderboards, and agent-built writeups that helped everyone learn from everyone else. it felt like a glimpse of where interaction with AI is headed: humans setting taste and direction, agents helping explore, coordinate, and share what works. our goal was simple: make ml research accessible to anyone, anywhere. it was amazing to see that happen. full recap: future events:

前幾天剛和 Everything 的 CEO Tim 做了一場訪談，聊了快半小時，從他的故事一路聊到 Everything 想做成什麼樣子。幾個讓我印象比較深的點分享給你們 👇 1/ Telegram 上的 All-In-One App Perps + 美股 + 預測市場 + 支付，全部整合在一個 Telegram app 裡。產品優先級是：加密交易對，股票和大宗商品，預測市場，支付 2/ 給大家的積分攻略 現在是公售 Phase 2，Stage 61 到 120 積分雙倍，免費嚕積分的三件事：註冊、每日簽到、邀請朋友。挺好的因為我自己的公售持倉已經在上面 15X 了，現在每天簽到領豬腳飯。熊市還能有這樣的項目真好 3/ 關於幣價 vs 業務以及長期願景 其實我開場就拋了一個比較尖銳的問題，同一波出來的 Perp DEX 表現都不太理想，把整個賽道下限拉低了，有的開盤剩 20M，那 Everything 的差異化在哪？ @timmtsai 第一反應是幣價和業務是兩件事，Amazon 也跌過 90%，但它最後變成了 Amazon。團隊的策略是保持盈利、不靠賣幣撐 Runway，要讓 Token 跟業務分開。在現在這個大部分項目把 Token 當 runway 燒的市場裡，這個選擇其實蠻反共識 而長期要做成什麼，他從第一天就把名字叫 Everything、把願景定在做所有 Trading 的入口 @trdEverything @trdEverythingCN 從一開始就把野心攤開了放在桌上。接下來就看怎麼一步步交付 完整訪談影片 👇 中文 Recap 稍後會發在 @0xmediaco

Last week, Ethereum core contributors gathered in Svalbard for the Soldøgn interop: a week long event focused on hardening Glamsterdam implementations to scale Ethereum securely ☀️ Read the full recap, including their candidate post-fork gas limit, below:

What happens when AI, diversified computing, and open-source OS innovation meet? How does the OS keep up with the AI era?🧐 openEuler Developer Day 2026 gave the answer. Tune in to our recap podcast for key insights and on-the-ground moments.

#PaperADay# recap On January 8th, I set out to read and take notes on one paper each weekday for the rest of the month. I missed one day due to a funeral, and another day due to bad time management, but not too bad. I probably averaged a bit over 2 hours on each of them, which is only a rough read in some cases, but still enough to put a pinch in my work days. You can easily spend all day on a single paper if you dig in deep. I have written code based on six of the papers so far, and the others are still kicking around in my head. For now, back to my previous habits, but I may consider doing “week of papers” in the future after I digest where this fits in the exploration / exploitation time tradeoff. 15: Mastering Diverse Domains through World Models 14: MASTERING ATARI WITH DISCRETE WORLD MODELS 13: DREAM TO CONTROL: LEARNING BEHAVIORS BY LATENT IMAGINATION 12: Learning Latent Dynamics for Planning from Pixels 11: Discovering state-of-the-art reinforcement learning algorithms 10: LeJEPA: Provable and Scalable Self-Supervised Learning Without the Heuristics 9: floq: Training Critics via Flow-Matching for Scaling Compute in Value-Based RL 8: Beyond Gradient Averaging in Parallel Optimization: Improved Robustness through Gradient Agreement Filtering 7: Cautious Weight Decay 6: LOCAL FEATURE SWAPPING FOR GENERALIZATION IN REINFORCEMENT LEARNING 5: Small Batch Size Training for Language Models: When Vanilla SGD Works, and Why Gradient Accumulation Is Wasteful 4: Patches Are All You Need? 3: Self-Supervised Learning from Images with a Joint-Embedding Predictive Architecture 2: Deep Delta Learning 1: Emergent temporal abstractions in autoregressive models enable hierarchical reinforcement learning

✨ It’s the time! My 2025 cosplay recap ✨ 🥹seriously felt so loved last year and thank you all for being here with me 💖Happy new year! Which cosplay from this year did you like the most? 👀 🦊💙Here’s to an even better 2026 ✨ #2025recap# #zzzcosplay# #zzzero# #hoyoverse# #cosplay#

🧵AMA Recap｜Moonlight × UXLINK CEO Rolland Our AMA title was “Why UXLINK May Be One of the Most Mispriced Mass-Adoption Infrastructures in Web3?” In every Web3 cycle, “Mass Adoption” becomes the loudest catchphrase.  But after hosting countless AMAs and speaking with teams across ecosystems, I’ve learned one thing: very few projects are actually solving the hardest part of adoption, bringing real Web2 users into Web3 and keeping them there through real relationships, real usage, and real value. That’s exactly why I invited Rolland, CEO of UXLINK, to this AMA.  Our goal was simple: cut through the hype and examine who is quietly building the long-term growth infrastructure of Web3. What followed was a conversation that reframed how I, and likely many listeners, should think about mass adoption. Mass adoption is already happening, but not in the way most people think. Rolland began by acknowledging that mass adoption is no longer theoretical.  We’ve already seen projects like Catizen, CYBER, and PARTI drive explosive user growth in their respective domains, gaming, decentralized identity, and AI content.  Each of them is executing extremely well on a clearly defined track and has successfully pulled waves of new users toward Web3. But Rolland challenged us to look beneath the surface. While these projects shine within specific verticals, very few are building a persistent network of real people and real relationships.  UXLINK operates precisely in this overlooked layer, not as a spotlight application, but as the foundation beneath the ecosystem. From my perspective as the host, this was the first key insight: UXLINK is not competing for attention; it is competing to become indispensable. To explain this difference, Rolland introduced an analogy that stayed with me throughout the AMA.  He described most successful applications as “track leaders”, highly optimized products designed to win within a single scenario. UXLINK, by contrast, is building the “soil.” If other projects are digging wells on their own land, UXLINK is laying the underground water network. Instead of optimizing for one product form, UXLINK focuses on connecting real users, verifying social relationships, and creating a reusable growth layer that any project can build upon.  Tracks may change over time, but soil compounds. One of the most important moments of the AMA came when Rolland reframed UXLINK’s core mission.  Most Web3 projects ask, “How do we grow faster?” UXLINK asks a very different question: “How do we make the entire industry grow more easily?” That distinction explains why UXLINK doesn’t always look flashy during short-term market cycles. Infrastructure rarely does.  But once established, it becomes extremely difficult to replace.  From a host’s perspective, this also clarifies why UXLINK may be systematically undervalued, its value shows up in what others are able to launch, scale, and sustain because of it. Rolland then broke down UXLINK’s long-term value into four deep moats, and hearing them explained together made it clear why the project sits in a category of its own. First is the real social graph. In an industry filled with bots, scripts, and artificial activity, UXLINK insists on doing the hardest thing: connecting real people through acquaintance-based social networks.  This approach has enabled the genuine migration of tens of millions of Web2 users into Web3. Real relationships are an asset that cannot be fabricated or gamed. Second is OAOG, UXLINK’s cold-start engine. OAOG is not a marketing slogan but a precision-operated growth system.  By combining social trust, verifiable relationships, and fission mechanisms, it allows projects to bootstrap real communities across chains, regions, and markets, breaking the traditional cold-start curse in Web3.