🔥 Search Hot Tweets
Search and analyze hot tweets from KOL accounts list (list: https://x.com/i/lists/1961235697677017443) within 6 hours. Use SoPilot plugin to quickly comment and occupy the comment section.
Yeah!!! 上电视啦!!! 感兴趣的朋友一定要听呀! https://t.co/Qe24i37AOD https://t.co/ghujwz4cj0
韩国黑色电影《无可奈何》,有点当年《寄生虫》的感觉了,好看啊! 中年失业大叔为了找工作,把其他竞争对手一一干掉。 人生真的是一念天堂一念地狱啊。 https://t.co/tTcBENVAFJ
妈妈你为什么带精灵的帽子? 帽子的第一性原理不是保暖吗? https://t.co/cleU30rW4e
上次被那堆愚蠢的、记错日期的东欧老乡给气着之后落了个病根儿,一生气就胃疼。哪怕稍微烦躁一下、血压升高了都会疼,反应极其灵敏,真是逼着我学会修身养性,凡事儿别生气。 花了一天时间,已经和出版社battle得差不多了,他们已经把电子书撤了,我发了我自己的,等审核通过就可以给大家发链接了,真是不容易。也证明了很多时候生气什么的情绪其实没什么用,先把事情沟通做好最重要。
Use this simple prompt to generate more poses In a 3x3 grid, show this character in different angles, keep the scene the same, random poses https://t.co/CW9g1VEXyM
🌿🍂 Capture the romance of the seasons with this healing botanical leaf art style—nature is the best artist. 🎨 🍌 nano banana pro prompt --- Prompt Template --- Creative botanical art collage depicting [Scene/Activity]. The image is constructed entirely from exquisite cut natural leaves, flower petals, plant stems, and wild berries. The silhouettes of figures and objects are formed by the intricate arrangement of these plant elements. The plant surfaces feature a slightly wet, glossy texture adorned with crystal-clear morning dew droplets. Rich, vibrant colors with distinct visual layering. The background is a soft, natural bokeh that harmonizes with the mood of the scene. Macro photography style, hyper-realistic organic textures, 8K resolution, masterpiece. --- Scene:a girl riding a bicycle through a park
最近推友推荐的这两个 App,都是 AI 让小应用变得非常便捷的案例 比大公司产品好用得多 https://t.co/m6JI6Epvok
几秒钟内将你的 localhost 分享给任何人。 通过一个在全球可访问的 HTTPS URL! $ brew install cloudflared $ cloudflared tunnel --url <你的本地域名> 这是 ngrok 的一种替代方案,无需注册、完全免费。 https://t.co/I0XqN47TD3
This JSON lets you turn any image into a stylish, retro magazine cover poster with one click. Just replace {{TITLE_TEXT}} with the title you want, and the big headline on the poster will update automatically. { "poster_structure": { "format": { "aspect_ratio": "2:3", "orientation": "vertical", "safe_margins": "medium", "grid": "classic magazine cover grid with generous top header area and bottom utility area" }, "layers": { "background_layer": { "type": "single-photo-or-flat-color", "treatment": \[ "subtle paper grain", "light halftone/print texture", "slightly faded print tones" \] }, "title_layer": { "type": "oversized headline behind subject", "text": "GOOD NIGHT", "placement": "upper third, spanning wide", "z_index": "behind_subject", "style": { "case": "title case or sentence case (match exactly)", "font_feel": "bold retro editorial display", "color_rule": "match dominant background color for embedded look", "effects": \["no outline", "no glow", "clean fill"\] } }, "subject_layer": { "type": "foreground cutout", "masking": "clean edge cutout, natural silhouette", "z_index": "top", "rules": \[ "subject must overlap and partially cover the title text", "no extra cutout artifacts", "no duplicated subject" \] }, "frame_layer": { "type": "border + inner line + ornaments", "outer_border": { "material": "off-white textured paper", "shape": "rounded corners", "rule": "NO outer black frame" }, "inner_frame_line": { "style": "thin line near edges", "color": "subtle warm gray/beige", "opacity": "low" }, "corner_ornaments": { "style": "small classic corner marks", "quantity": "4", "subtlety": "minimal" } }, "header_microtext_layer": { "type": "small header metadata text", "placement": "top margin area", "alignment": "left to center, magazine-like", "text_blocks": \[ { "text": "GOOD NIGHT", "role": "micro_title" }, { "text": "Designed by @underwoodxie96", "role": "credit_line" }, { "text": "Date 16-12-2025", "role": "date_line" }, { "text": "Day Thirty {two}", "role": "day_counter" }, { "text": "2/5", "role": "issue_fraction" } \], "style": { "font_feel": "clean editorial sans or simple serif", "size": "very small", "tracking": "slightly increased", "color": "muted dark gray" } }, "barcode_layer": { "type": "barcode graphic", "placement": "bottom right", "size": "small to medium", "style": "clean black/gray on light background", "rule": "must remain visible" } }, "print_finish": { "global_texture": \["paper grain", "subtle halftone"\], "contrast": "medium-low", "saturation": "slightly faded", "clarity": "high but not HDR" } }, "constraints": { "must_keep": \[ "off-white textured border with rounded corners (NO outer black frame)", "thin inner frame line near the edges", "small corner ornaments", "oversized title text placed behind the subject and partially covered by the subject cutout", "paper/halftone print texture across the whole design", "barcode at bottom right", "header microtext blocks present at the top" \], "text_lock": \[ "Designed by @underwoodxie96", "Date 16-12-2025", "Day Thirty {two}", "2/5", "GOOD NIGHT" \], "avoid": \[ "outer black border/frame", "extra random text", "misspelled or garbled typography", "neon colors", "heavy HDR", "watermarks or extra logos" \] }, "negative_prompt": [ "watermark", "extra logos", "misspelled text", "garbled typography", "random letters", "outer black frame", "overprocessed", "heavy HDR", "neon colors", "low resolution" ] }
每当我抛出一个观点 评论区都会有人问 那该咋变现赚钱? 信息差赚钱方法不要太多 赚国内的就是卖课卖社群 搞咨询公司卖咨询服务 我说一个靠信息海外赚钱的思路吧 就是做参考阅读的海外版 直接搞聚合然后做精选的深度阅读 然后用Dan koe的工作流newsletter转长推做增长 重点解决俩问题: 信息过载的精选-有限时间,不错过重要内容 深度信息的洞察压缩-3小时访谈精选出核心观点(甚至带着时间戳) 就围绕一个人群干垂直 肯定可以变现
现在有一种流派玩法 前期靠运营n8n验证 验证成功后工程师vibe coding+hitp去反转n8n workflow成代码产品化 不过n8n的workflow有很都不太能直接投产使用 比如就拿youtube视频转中文来举例吧 这种东西卡点很多 ytb视频下载就是一个非常大的卡点 之前还可以无限下 现在一个cookie只能下一次 然后时间戳校对也是个卡点 这些卡点都拼在一起后,多数情况下光靠工作流想做到比较好的效果还是非常困难的 这种需求还是建议找靠谱的开源项目,比如videolingo 简单总结一下我对n8n的看法: 1、产品/运营专家能用来验证流程,想量产还是要代码化,如果会vibecoding,直接写很多时候比n8n快 2、个体户/机构可以用来唬smb,没啥使用规模,业务几近定制,n8n切这里是有很多市场空间的,一方面可以自己外包赚钱,另外一方面可以靠这个验证需求,后面再想办法产品化 3、做培训的/搞自媒体的可以用来搭玩具,做一些看起来有用的情绪价值很大的东西,很容易传播 4、不要对n8n抱有太高预期,也不应该太低估它,很适合个体户/小业务做运营验证和初步自动化,也适合搞社群卖课,不适合产品化规模化,上体量追求性能还是需要换换

宝玉
https://t.co/K2b5BeYshn 扎克伯格讲了一个特别精彩的观点:在打造AI团队时,人并不是越多越好。好比在厨房里煮一道菜,厨师再多,炖汤的时间也不会更快。 最近一次访谈里,扎克伯格揭开了Meta打造AI顶尖团队背后的独特秘诀。听上去,这更像是在组织一支精锐的“突击小分队”,而不是管理庞大的研发部门。 一场“小而精”的科学实验 “你真正想做的,”扎克伯格开门见山,“是把它看作一场小而精的集体科学实验。也就是说,你需要的团队越小越好,小到团队里的每个人,都能把整个项目的全貌清晰地装进自己脑子里。” 扎克伯格也强调,并非所有项目都适合这种模式。比如,Meta赖以生存的动态信息流(Feed)和广告推荐系统,就是典型的“大兵团作战”:通过不断增加人员,整体产出就能持续提高。即便每个人带来的边际贡献会逐渐减少,但总量依旧可观。 “在那种大团队里,即便你身边同事的实验失败了,也不会明显拖慢你的节奏。”他说。 但大语言模型的开发,则截然不同。 “它更像是一艘小船,需要团队成员紧密协作,”扎克伯格生动地比喻道,“你希望船上的人尽可能少,每个人都能完全掌握全局,拿出最顶尖的工作状态。这意味着,船上每个座位都无比珍贵,一票难求。” 扁平化管理,拒绝“技术退化” 为了维持这支小团队的顶尖战斗力,扎克伯格明确指出,要尽量避免团队管理的层级过多。 “我们不想让团队内部形成复杂的管理层次,”他说,“因为一旦某个人转入管理角色,即使几个月前他还是一线科研大牛,技术能力也会迅速衰退。” 因此,Meta推崇的是一种“人数精简、人才密度极高、管理结构极度扁平”的团队模式。在扎克伯格看来,这样的团队才是真正能够打赢硬仗的精锐之师。 CEO亲自下场,但只做两件事 当被问到他本人在团队中的角色时,扎克伯格谦虚地承认,自己并不是AI科学家。他的作用只有两个,却至关重要——招揽世界顶尖人才,提供最优质的资源保障。 “我的首要任务,就是吸引全球最厉害的AI研究者加入Meta,”他说。为了达成这个目标,他花了大量时间与世界范围内的顶尖科学家和行业专家面对面交谈,去了解谁最适合Meta,以及Meta是否能在他们职业生涯的重要节点上,提供最好的机会。 扎克伯格补充道:“第二件我非常关注的事情,是确保我们的人均算力(compute per researcher)远远超过其他任何实验室。”他自信满满地说,“目前来看,在人均算力方面,我们确实遥遥领先。” 扎克伯格也坦言,这种大手笔投入的底气,来自Meta强劲的商业模式和稳健的盈利能力。 采访者笑着调侃道:“你们利润还不错吧?” 扎克伯格也笑了:“嗯,还算可观。” “所以你可以直接打电话给老黄(英伟达CEO黄仁勋),跟他说:‘请再来点GPU?’” “哈哈,倒也没那么简单,”扎克伯格开了个玩笑,“不过通常我都是用智能眼镜给他发短信的。”
Est. 600 views for your reply
