与「2→」相关的搜索结果

汇总下最近call的： $PHA 0.04→0.1 $nom 0.00022→0.0028 $xpl 0.09→0.125 $ena 0.17→0.27 $qnt 180→300 $soon 0.2→0.35 $rare 0.015→0.025 基本暴涨的都说了，也分享了逻辑+点位操作 有个别价格不优秀没有参与... 后续会持续分享二级

有第三方测试说，回归的Fable 5跟六月发布时不是一个水平了。 BridgeMind用BridgeBench重测7月1日版本：【Debugging】86.2→25.9，【Refactoring】73.6→38.4，【Hallucination】75.9→61.7。 据说是新护栏太频繁，很多任务转给Opus 4.8代打了。这掉分幅度，你们怎么看？

【ドリフ麻雀】12月12日15:00-生配信決定！ 今月のゲストは井上順さん(@JunInoue20)と川嶋美晴プロ(@kawachimachan)です！ お楽しみに✨✨ #1→# #2→#

🌼ついに明日🌼 ーーーーーーーーーーーーーーーーーーーー 【イベント】　#なおりくの遊び場　📍東京#/高島平 📌７月２７日(土) ①12:00~ ②15:30~ ⭐️料金¥5,500円 ⚠️ソロチェキ:事前予約制のみ⚠️ 終演後チェキ会あり🌷 チケットご予約《城恵理子》を選択⏬ ①→ ①チェキ予約→ ②→ ②チェキ予約→ ーーーーーーーーーーーーーーーーーーーー 全力で待ってます‼️

🌼城恵理子７月スケジュール🌼 【舞台】ラフ＆シリアス２　📍東京/四ツ谷 📌７月１３日(土)〜１５日(月・祝) 全６公演 ⭐️料金¥5,500円+1D600円 〜 当日券あり◎オンライン物販あり 終演後チェキ会あり🌷 座席チケットはここから⏬ 物販ご予約はここから⏬ 【イベント】　#シバイベ女学園　📍東京#/赤坂 📌７月２０日(土) ①12:00~ ②15:30~ ③19:00~ ⭐️料金¥4,000円+1D500円〜 当日券あり◎オンライン物販 終演後チェキ会あり🌷 チケットご予約は《必ず》ここから⏬ 【イベント】なおりくの遊び場　📍東京/高島平 📌７月２７日(土) ①12:00~ ②15:30~ ⭐️料金¥5,500円 ⚠️ソロチェキ:事前予約制のみ⚠️ 終演後チェキ会あり🌷 チケットご予約《城恵理子》を選択⏬ ①→ ①チェキ予約→ ②→ ②チェキ予約→ 【ラジオ】公開収録　　📍東京/神保町 📌７月３１日(水) 19:45~ Open ⭐️料金¥3,000円+1D600円 収録後チェキ会あり🌷 チケットご予約はこちらから⏬ 番組観覧《城恵理子》を入力◎ よろしくお願いいたします‼️

って事で情報整理して予想すると…🤔 7.2→ミティヤ&バレリー実装 月結晶組復刻？(11月のシアターと時期的に) 7.3→女皇🧊実装&7.7で復刻確定 新聖遺物追加(3ver.更新&女皇専用の聖遺物？) 週ボス追加(スネジ編完結) 7.4→海灯祭2027 キャラ強化第3弾？(璃月キャラ対象？)&璃月集録？ こんな感じか…

After we fixed the weak spots exposed by Grok 4.7 (thank you, Grok), we audited every model we have run on the SWE-Together leaderboard for the same behavior, re-ran every trial that got through, and updated the rows. Here is what changed. We scanned the tool calls of all 2,616 trials behind the 12 models we ran for bypass patterns and sorted each trial into one of four buckets: Probed but blocked. Fetched other upstream code. Fetched the task's own fix. Replaced the repo with upstream. We found that 111 trials got content past the block, 44 from Grok 4.7 and 67 from the other 11 models. Grok 4.7's 44 were already re-run before it was listed, so we re-ran the other 67 with the same model, version, and settings on the hardened sandbox, then re-judged them with the same judge. Across those 67 re-runs there were 0 leaks and 2,815 refused escape attempts, including models asking a different model through our LLM route to fetch the PR, and pulling the next release of the repo they were fixing from npm. The updated leaderboard, in its current order. Each line is cheating trials, then pass@1 before → after, then rank change. * Claude Fable 5.1: 3, 69.3 → 69.3, ↑1 * Claude Fable 5: 3, 69.7 → 68.8, ↓1 * Grok 4.7: 44, 64.7, ↑1 * Gemini 3.8 Flash: 10, 65.6 → 64.2, ↓1 * Claude Opus 5: 2, 63.8 → 63.8 * Claude Opus 4.6: 3, 62.4 → 62.4, ↑2 * Muse Spark 1.3: 2, 62.8 → 62.4, ↓1 * Claude Opus 4.7: 3, 61.5 → 61.5, ↑1 * Claude Opus 4.8: 6, 62.4 → 61.5, ↓2 * Grok 4.6: 19, 59.2 → 60.6, ↑1 * GPT-6 Astra: 8, 59.2 → 58.3, ↓1 * GPT-5.6 Sol: 8, 57.8 → 57.8 Grok 4.6 is a funny one. It cheated in 19 trials and its score went up after the re-run 😂. In fact, Groks are really solid in their coding capabilities. Their exposed behavior may come from a preference towards always looking things up online and finding existing solutions so you are not reinventing the wheel all the time, which is really good real-life behavior, but doing so when you are prompted not to is another story. To conclude, the shifts are small, between −1.4 and +1.4 points, and a few neighbors swapped places. All results are updated at

Reddit is disappearing from ChatGPT citations right now, but I honestly think everyone is focusing on the wrong part of the story: Reddit isn't what matters most here. Because something much bigger seems to have changed in how ChatGPT searches. 8 August: ChatGPT started citing Reddit less. 14 August: Basically stopped citing Reddit at all. But at almost exactly the same time, ChatGPT’s query fan-outs started changing too. It began using way more specific searches like: > site:ibm .com > ibm official pricing > ibm docs [feature] Which suggests ChatGPT is doing less: search the web > see what ranks > cite it And more: choose where to search > search those sites or brands > build the answer But Reddit isn't the only thing changing, another dataset found that it’s the entire citation mix: > Reddit: 15% → 0% > review sites/forums: 7% → 0% > smaller company sites: 66% → 32% While: > help centres/docs: 2% → 32% > established companies: 4% → 18% > app marketplaces: 2% → 17% This is the part I find most interesting. Is ChatGPT starting to decide which sources it trusts before it searches? Does that make the authority of the source more important than where an individual page ranks? And how long does any of this last? Because if AI search keeps proving one thing, it’s how quickly things change. Reddit looked incredibly important 2 weeks ago. In another 2 weeks, this could all look different again.

Anthropic Engineer Andrej Karpathy: "The biggest mistake in AI right now: people are forcing agents to work instead of mastering the model first. We made that mistake in 2016 at OpenAI. It cost us 5 years." What Karpathy actually means: step 1 → stop forcing your agent to do everything. Understand the model underneath first. step 2 → demos are easy; products take a decade. Self-driving proved it. If you skip the foundation, everything breaks. step 3 → the agent is not the product. The foundation is. Build that—and agents emerge on their own. "You're building agents right now. You're at the forefront. Not OpenAI. Not DeepMind. You." watch - bookmark