最近,不少打工人在刷AI摸鱼工具时,意外发现美团推出了一款叫Tabbit的AI浏览器。
它与传统浏览器不同,内置多款主流大模型,用户只需一句话描述需求,它就能跨网页、跨软件完成一整套任务。
他们本来只想找个能偷懒的插件。结果,很多人打开Tabbit后,在首页的“妙招广场”里一逛就是一上午。
这个地方,有点像打工人的互助市集。做工程投标的,把筛全国中标记录的活儿拆成两个妙招,20分钟就跑完原本一天的工作量;搞医学研究的,用“文献速读员”把几十篇论文自动整理成综述;还有人主动帮别人优化妙招,或者挑出好用的推荐给同行。
这些人只是想解决工作中某个琐碎耗时的痛点。他们没意识到,这些看似随手的动作里,可能藏着一些新的职业可能性。
看着原本需要4个人埋头苦干一整天的活,被两个“妙招”二十分钟就给解决了,CQ感觉眼前一亮。
CQ是一名建筑工程师,主要负责公司流程优化和信息化管理,说白了就是用数字化工具帮企业减负提效。
对建筑公司来说,获取项目线索至关重要。CQ所在的公司,专门安排一位同事,定期从“全国公共资源交易平台”筛选设计类中标项目,再把关键信息整理成报表,供业务人员使用。
这份工作看似只是抄抄数据、做做Excel,但实际涉及五个省几十个核心城市,每次都要人工处理几千条记录,而且不同城市的表格格式还不一样,需要逐一调整。同事每次都要忙活好几天。
“这磨人的活儿,真的不能交给AI吗?”CQ一直想找个自动化工具来解决这个问题。
今年3月,他在网上看到美团推出了款名叫Tabbit的AI浏览器,便下载下来试了试。
与只能“看网页”的传统浏览器不同,Tabbit更像一个能听懂人话的AI助手。它内置了多款主流大模型,用户只需要一句话,它就跨网页、跨软件跑完一整套复杂任务。
Tabbit里有一个叫“妙招”的功能。用户的常用提示词、Agent自动化任务、脚本流程,都可以保存成一个妙招,一键复用。比如,你每周写周报都要重复输入“请根据本周工作内容,生成一份包含已完成、进行中和下周计划的周报”,这句话就可以存成一个“周报生成器”妙招。
下次只要打开你本周的在线工作文档,在侧边栏对话框输入/,就能调出“周报生成器”妙招,或者用@添加你这周写过的稿子、工作群的聊天记录,它就能一键总结你本周的工作内容。
用得顺手了,还能传到社区的“妙招广场”上,供别人直接使用或者改一改适配自己的场景。
CQ一开始想写一个“大而全”的妙招,光提示词就写了七八百字,想把打开网页、筛选数据、输出Excel等全套流程一次性塞进去。
结果刚跑到100步,系统直接报错。流程太复杂,把浏览器自动化的额度给用完了。
他把报错截图扔给Tabbit。Tabbit建议:把复杂任务拆成两个独立妙招。
第一个妙招专门负责打开符合条件的中标项目网页;第二个妙招再从这些网页中抓取数据、整理成Excel并输出到本地。两个妙招串联后,原来4个人干一整天的活,现在只需设置关键词就能自动完成。
如果说CQ是通过创建妙招“解救”了自己,财经撰稿人陈琳则是在妙招广场上“逛”出了工作的新解法。
5月初,陈琳刷到一篇Tabbit的推文,抱着“试试又不花钱”的心态下载了下来。她的工作需要跟海量资料打交道,写一篇人物稿,光搜集前期素材就要耗去两三天。
最近,陈琳在写一位芯片大佬的深度稿件,大佬有一段长达90分钟的视频访谈,是这篇稿子的核心素材。搁在以前,她只能耳机一点点地听,听到关键处就得暂停,对着字幕在Word里一句一句敲,一段采访就要耗费大半天。
在Tabbit首页的妙招广场上,陈琳找到了“救星”。她发现了一个叫“B站视频文案提取”的妙招,收藏之后就能从对话窗口直接调用。她把访谈链接丢进去,两三分钟后,访谈内容就被整理成了一份txt文档。
尝到甜头后,陈琳继续往下翻,又发现了一个叫“B站评论区总结”的。过去为了呈现网友反馈,她得抱着手机一条条翻,现在一键就能抓取热评。而“小红书首页关键词爬取”,则正好能帮她运营自媒体账号时抓选题。
一整个上午,陈琳在广场里逛得不可自拔,收藏夹里多了七八个妙招。
在妙招广场上,像CQ和陈琳这样的用户并不少见。
学医的研究生,通常要翻阅几十篇文献,追踪最新的研究进展。这活儿得手动检索、逐篇下载、整理摘要,再写成综述汇报给导师。用户“CPU-Broken”创建了“医学影像报告生成论文日报”和“文献速读员”两个妙招。把关键词丢进去,就能自动生成带图表的文献综述。原本要从早翻到晚的活,几分钟就跑完了。
此外,他还制作了“论文源码检索”“AI大事早知道”等多个妙招。
炒股爱好者ZX.则把盯盘的苦差事扔给了Tabbit。他在广场上传了一个“股票控-v4.3.1”妙招,每天定时抓取自选股的价格、成交量和公告,自动生成盯盘日报。
还有更接地气的。有人给高考生做了大学专业推荐的妙招;有人临考前把网课链接丢给Tabbit,三分钟生成带核心概念和自测题的复习妙招;甚至有人做了“麦当劳自动领券助手”“Boss直聘一键投递”这种“懒人福音”式妙招。
它们的创造者和使用者,大多是普通上班族、学生和自由职业者。内测期间,广场已经积累了上千个妙招。
它们大致分三种类型。像CQ这样,把一整套业务拆成标准化组件的,属于“工作流架构型”;也有人专注打磨单个环节,让指令跑得更快、更稳,做出一些“技能优化型”妙招;还有针对特定小众场景的“垂直场景型”,比如跨境电商选品对比、小红书爆款标题生成器,用的人不多,但需求精准。
这些妙招之所以大量涌现,是因为它们都是从真实场景里长出来的。上班族的工作中藏着大量琐碎的内容。比如写稿人想快速提取视频文案、HR要筛几十份简历、行政每个月要拉十几张格式不同的报销表。这些需求太细碎,大厂一般不愿意专门立项。
Tabbit打破了这个困境。用户一句话说清需求,它就能生成可执行的妙招。有时候,20分钟写出来的东西,就能把一个长期痛点填平。
这种“想到就能做到”的感觉,让很多人乐在其中。CQ和CPU-Broken大概还没意识到,他们已经摸到了一个新职业的边界。
CQ和CPU-Broken都不是程序员,但都知道自己行业里哪个环节最磨人、最容易出错。他们把这套经验拆成可描述、可复用的步骤,变成了别人能直接用的工具。
很难用传统职业目录归类他们。普华永道、德勤也有专门做业务流程优化的顾问,但那是企业级的系统改造,需要代码和BPM软件,服务的是组织。
CQ们做的事颗粒度更细。他们从自己工作中的痛点出发,用自然语言把一道道工序拆成AI能执行的指令,封装成一键复用的妙招,先解决自己的问题,再顺手放到广场上。
如果非要起个名字,大概可以叫“妙招工程师”。他们不写代码,只拆问题,把让自己头大的环节,拆成一句话就能调用的轻量工具。
小曾也是这个路数。她是艺术专业的大四生,今年夏天写毕业论文时,导师要求她用FSQCA,一种从没碰过的量化研究方法。软件是全英文界面,她连按钮都认不全。以前这类活儿得自己翻教程、啃文献、手动整理案例的数据,光搞懂概念就要花大半个月。
为了迅速上手,她建了一个“学术概念解释”妙招。遇到陌生术语就丢进去,让它解释定义、逻辑和应用场景,再推荐相关文献。几十篇论文,她按主题分成三个标签组,让Tabbit提炼核心观点,快速拆解量化研究的实际应用案例。
从量化研究方法入门,到梳理数据结构,她全程借助Tabbit提升了调研效率,把时间留给论证逻辑打磨和内容撰写,最后拿到了优秀毕业论文。
“学术概念解释”妙招,后来被她传到广场上。她还做了“找工作和机票比价”的妙招,有两三百人在用。
CQ的妙招虽然没公开,但部门三个同事早就复用到了别的业务线上。
还有一类人,他们工作资历深、眼光独到,能在妙招广场里“淘金”。
陈琳就是其中一个。她已经在广场里泡了一个多月,前后试了二十几个妙招。发现好用的,就随手分享出去。
“B站视频文案提取”她第一时间丢进工作群,没过几天,有两个同事也用上了。她还翻到一个“AI日报”妙招,能每天自动抓取行业热点,转手发给了总担心跟不上时代的老爸。
逛得多了,陈琳开始按照自己的想法“二次装修”。广场上有个“PPT神器”,能根据主题自动生成大纲和排版,她试了几轮,觉得模板库太单薄,而且只能网页版操作。她让Tabbit 帮她改了一下,把常用的几套企业模板嵌进去,又补了一条自动导出PDF的指令,做成了自己的“迭代版”。
这种“逛、发现、分享、顺手打磨”的路数,本质上是在替人淘金,和GitHub上给开源项目提交PR(代码改进)的贡献者很像。
如果社区生态再成熟一些,陈琳这类人有可能会成长为“妙招猎人”。
猎手和普通用户的区别在于,他们不只停留在发现和分享,还会动手改造。广场上那些通用的工具,比如“简历筛选”,猎手会拿过来,根据自己的行业经验,把它改成“互联网运营岗专用版”或者“外企财务岗专用版”。
他们会调整提示词、增加前置条件、优化输出格式,让原本粗糙的通用工具,变成贴合特定场景的专用版本。
这种改造需要对业务场景足够了解,知道哪些参数该调、哪些步骤该删。改完后,猎手把这个“适配版”分享给同行业的同事,或者打包成“某某行业工作流套件”。
等到平台激励和分成机制完善,也许有人会愿意为这种经过专业化定制改良的版本买单。
等广场上的妙招从几千个增长到几万、几十万个时,还会催生第三种角色——妙招测评师。
因为妙招生产变得极其容易,整个妙招广场上可能会出现成百上千个功能相似的“自动写周报”“一键整理Excel”。这个时候,用户很难判断哪个更好用,能写出重点突出、结构清晰的周报,哪个经常失败,哪个又藏着额外惊喜。
妙招测评师正是基于这样的需求诞生的一个机会。他们和妙招猎人不同,是替大家先吃第一口螃蟹、帮大家做决策的人。他们可能会测试同类妙招,详细记录哪个容易报错、哪个输出格式混乱、哪个在特定浏览器版本下跑不动。
他们的价值在于建立信任。面对几十个功能类似的妙招,普通用户没时间一个个试。测评师帮大家节省下那几十个小时的踩坑时间。
他们有点像数码测评博主,只不过评测的对象换成了妙招。他们可以测试妙招的稳定性和报错率,比较同类妙招的差异,出一些“本周值得关注的五个妙招”这样的推荐,或者针对考研、跨境电商等垂直领域做深度横评。
数码博主靠流量和广告变现,妙招测评师未来也许也可以通过平台的创作者激励或付费订阅养活自己。
可能会有人会问:AI不是已经能根据一句话临时生成解决方案了吗?为什么还要去广场找预制的妙招?如果每个人都能5分钟做一个妙招,凭什么你做的值得付费?
不是所有AI生成的内容都值得付费,但经过验证、能解决具体痛点的版本值得。每个人都能5分钟写一条小红书笔记,但专业的内容策划依然有人付费。差别就在于系统化和信任度。
猎手把通用简历筛选器改成“金融风控岗专用版”,里面沉淀的是他对这个岗位的长期理解;测评师告诉你“这个妙招在Mac上稳定,那个在Windows上容易出问题”,背后是几十次的真实测试。当生成变得廉价,筛选、改造和验证反而更值钱。
这三种新职业现在还没成气候,但苗头已经有了。
妙招工程师、妙招猎手、妙招测评师,这三个名字听起来有些超前。但类似的剧本,其实已经上演过。
十多年前,全球最大的代码托管平台GitHub,还只是一个程序员存代码的地方。最早那批开发者,根本没想过靠这个赚钱。他们只是把顺手写的工具传上去,方便下次换电脑时还能直接用。
后来star(给项目点赞)、issue(提问题反馈)、PR(提交代码改进)这些机制逐渐成熟,开源项目慢慢变成了职业背书。star数高的作者,比拿漂亮简历的人更好找工作。有人被猎头私信挖走,有人被企业高价请去当顾问,还有人靠GitHub Sponsors(GitHub的开发者赞助平台)每月收到几千美元赞助。
Notion也一样。最早只是用户把自己搭好的工作区模板分享出来,后来官方开了模板市场,开始有人卖钱。
2022年,美国创作者Thomas Frank靠卖Notion模板年收入突破100万美元。新加坡学生Easlo的模板生意同样跑出了规模,截至2023年3月他累计收入近50万美元,每年的年收入保持在六位数。
Tabbit妙招广场现在还在早期,但底层逻辑是一回事。
当下人人都在造skill,却很少有人追问:这东西真的好用吗?能跑通吗?真能解决问题吗?
妙招工程师把自己的工作经验封装成能复用的工具,猎手把通用工具改成行业专用版,测评师替你踩坑排雷。他们补上的,是从“能生成”到“能落地”的最后一公里。能够解决实际问题,可能才是他们未来能够实现付费的原因。即便不能成为职业,也许有机会能成为副业、带来收益。
用户“蒜瓣脑袋”,在广场做了两个看起来八竿子打不着的妙招。
一个是“618满减差几块?算法来凑单”,内置了七项硬性数据校验和六维加权评分,帮你把购物车凑到最划算;另一个是“Tabbit一站式科研助手”。前者解决生活里的细碎烦恼,后者解决学术里的系统性难题。
如果生态再成熟一点,这两个精准解决特定痛点的工具,是不是也有机会像Notion模板一样打包成订阅服务,变成实实在在的收入。
这就是妙招工程师的路子,帮人省时间,省下来的时间就值一笔咨询费。
妙招猎手靠的是,对业务场景的理解和动手改造的能力,即便在AI时代它们依旧稀缺;测评师则是靠信任,当你的推荐每次都靠谱,关注的人就有可能愿意为这份判断付费。
从“觉得好用”到“愿意付费”,中间还隔着几步。悬赏功能、评分体系、创作者分成,GitHub和Notion花了五到十年才逐步完善。
在AI时代,这个速度可能会快很多。GitHub的门槛是会写代码,Notion的门槛是会搭模板,妙招的门槛只有一句话。
当生成工具变得前所未有的简单,所有人都能轻松上手。真正拉开差距的,就从技术能力,变成了观察力、发现力、总结能力和个人品味。过去那些满脑子想法,却因为不会写代码、不会做模板而卡住的人,现在终于有了施展的空间。他们最懂自己行业的痛点,也最知道什么样的妙招真正好用。
现在,他们只要用一句话描述需求,Tabbit就能生成可执行的工作流,花几分钟就能让普通人的创意和经验快速变成可复用的工具,解决一大群人的长期痛点。
眼下正是入场窗口期。Tabbit妙招大赛正在进行,6月30日截止,三条赛道最高奖金万元。平台后续还将推出创作者激励计划,让那些真正解决痛点的妙招,有机会转化为副业收入。
如果暂时没想好做什么,也可以先去妙招广场逛逛。
根据大赛热力榜,截至6月22日,50款上榜妙招累计添加量已突破1.3万。排在前三的分别是“搜索结果去广告”“视频总结”和“图反推提示词”。再往下翻,B站文案提取、做个人IP、课本下载、AI热点新闻采集…… 几乎覆盖了打工人的全部日常。
这些排名靠前的妙招,有一个共同点:它们都在替用户省掉重复劳动。榜单本质上是一份“打工人最烦什么”的实时调研。
这些妙招的创作者,大都是普通上班族和学生。他们和CQ、小曾一样,只是先把自己的麻烦事解决,顺手分享到广场。
小曾已经把自己做的几个妙招上传参赛。她说,创建一个妙招只花了5分钟,没指望一定拿奖,但这是个验证想法的好机会。
“每次看到使用人数涨上去,就特别有成就感。”小曾说。
CQ则看得更长远。在他看来,平台和工具的价值已经在那了,只是交易机制还没跟上。“其实可以开个悬赏功能,有人提需求,有人接单提供服务。”他建议。
GitHub和Notion的历史都证明,越早进场的人,越能积累数字资产和口碑。等到平台成熟、用户愿意为效率付费时,有作品、有数据的人,自然能接到第一单。
对普通人来说,现在最该做的也许不是等变现通道打开,而是先占个位置。
- END -
撰文|杨 梅
编辑|张 霞
往期回顾
Recently, many office workers browsing AI tools for slacking off at work stumbled upon Meituan's new AI browser called Tabbit.
Unlike traditional browsers, it comes with multiple mainstream large language models built in, so users just need to describe their needs in one sentence, and it can complete a full workflow across web pages and applications.
They originally just wanted to find a plugin that could help them slack off, but many ended up spending their entire morning browsing the "Smart Tips Plaza" on Tabbit's homepage.
This space is something like a mutual-aid marketplace for office workers. Engineering professionals split the tedious task of filtering national bid-winning records into two Smart Tips, finishing in 20 minutes what used to take a full day; medical researchers use the "Literature Quick-Read Assistant" to automatically organize dozens of papers into review articles; some even help others refine their Smart Tips or recommend the best ones to peers.
These people simply wanted to solve some trivial yet time-consuming pain points at work. Little did they realize that these seemingly casual actions might conceal entirely new career possibilities.
Office workers rescued by the Smart Tips Plaza
Watching what would have taken four people a full day of hard work get solved in just 20 minutes by two "Smart Tips," CQ felt a spark of inspiration.
CQ is a civil engineer mainly responsible for process optimization and informatization management at his company—essentially using digital tools to reduce burdens and boost efficiency for enterprises.
For construction companies, obtaining project leads is critical. CQ's company has a dedicated colleague who regularly screens design-related winning bids from the "National Public Resources Trading Platform," then organizes the key information into spreadsheets for sales staff.
This job might seem like merely copying data and filling out Excel sheets, but it actually spans five provinces and dozens of core cities. Each time, thousands of records must be processed manually, and spreadsheet formats vary by city, requiring adjustment one by one. The colleague always spends several days on it.
"Can't this tedious task really be handed over to AI?" CQ had always wanted an automation tool to solve this problem.
In March this year, he saw online that Meituan had launched an AI browser named Tabbit, so he downloaded it to try.
Unlike traditional browsers that can only "view web pages," Tabbit is more like an AI assistant that understands human language. It comes with multiple mainstream large language models built in; users simply describe a task in one sentence, and it runs the entire complex workflow across web pages and applications.
Fig|Tabbit has dozens of large language models built in, automatically helping people get things done
Inside Tabbit, there's a feature called "Smart Tips." Users can save their commonly used prompts, Agent automation tasks, or script workflows as Smart Tips for one-click reuse. For example, if you weekly write a progress report requiring you to repeatedly input "Please generate a weekly report based on this week's work content, including completed tasks, ongoing items, and next week's plans," you can save that sentence as a "Weekly Report Generator" Smart Tip.
Next time, just open your online work document for the week, type / in the sidebar chat, and you can pull up the "Weekly Report Generator" Smart Tip, or use @ to add drafts you wrote this week or chat records from work groups, and it will summarize your weekly work content with one click.
Once you're comfortable using it, you can also upload it to the community's "Smart Tips Plaza" for others to use directly or adapt for their own scenarios.
At first, CQ wanted to write a "comprehensive" Smart Tip, drafting prompt instructions that ran to hundreds of words, trying to cram the entire workflow of opening web pages, filtering data, and outputting Excel files into one go.
But after reaching 100 steps, the system threw an error. The workflow was too complex, exhausting the browser automation quota.
He sent the error screenshot to Tabbit. Tabbit suggested breaking the complex task into two independent Smart Tips.
The first Smart Tip is specifically responsible for opening web pages of qualified winning bid projects; the second Smart Tip then scrapes data from those pages, organizes it into Excel, and exports it locally. With the two Smart Tips linked, what used to take four people a full day can now be done automatically just by setting keywords.
Fig|CQ solved his work challenge using two Smart Tips
If CQ "rescued" himself by creating Smart Tips, financial writer Chen Lin found new solutions to her work by browsing the Smart Tips Plaza.
In early May, Chen Lin came across a promotional post about Tabbit and, with a "try it, it's free anyway" mindset, downloaded it. Her job involves dealing with massive amounts of material; writing a profile piece alone requires two to three days just to gather preliminary research.
Recently, Chen Lin was working on an in-depth feature about a chip industry boss. The boss had a 90-minute video interview, which served as the core material for the article. In the past, she could only listen to it bit by bit with headphones, pausing whenever something important came up to type out subtitles sentence by sentence in Word, taking half a day just for one interview segment.
On the Smart Tips Plaza of Tabbit's homepage, Chen Lin found her "savior." She discovered a Smart Tip called "Bilibili Video Transcript Extractor"; after saving it, she could invoke it directly from the chat window. She dropped the interview link in, and within two or three minutes, the interview content had been organized into a txt document.
Fig|Chen Lin used the "Bilibili Video Transcript Extractor" Smart Tip to boost efficiency
After tasting success, Chen Lin kept scrolling and found another called "Bilibili Comments Summary." In the past, to present netizens' feedback, she had to browse through comments one by one on her phone; now she could grab hot comments with one click. And the "Xiaohongshu Homepage Keyword Scraper" was perfect for helping her pick topics when managing her self-media accounts.
Chen Lin spent the entire morning endlessly browsing the plaza, adding seven or eight Smart Tips to her favorites.
On the Smart Tips Plaza, users like CQ and Chen Lin are not uncommon.
Medical graduate students typically need to read through dozens of papers and keep track of the latest research progress. This job requires manual searching, downloading each paper one by one, organizing abstracts, and then writing review reports for advisors. User "CPU-Broken" created two Smart Tips—"Medical Imaging Report Paper Daily Digest" and "Literature Quick-Read Assistant." Throw in some keywords, and you get auto-generated literature reviews complete with charts. What used to take from morning to night now runs in just minutes.
Additionally, he made several other Smart Tips, including "Paper Source Code Search" and "AI Big News Early Alert."
Stock trading enthusiast ZX handed over the painful task of monitoring stocks to Tabbit. He uploaded a "Stock Controller-v4.3.1" Smart Tip to the plaza, automatically fetching price, volume, and announcements for his watchlist every day and generating a monitoring daily report.
There are even more down-to-earth examples. Someone made a Smart Tip recommending university majors for gaokao (college entrance exam) takers; another tossed online course links to Tabbit before exams and got a review Smart Tip generated in three minutes, complete with key concepts and self-test questions; someone even made "McDonald's Automatic Coupon Grabber" and "Boss Zhipin One-Click Apply" — "lazy person gospel" style Smart Tips.
Fig|There are thousands of practical tools on the Smart Tips Plaza
Their creators and users are mostly ordinary office workers, students, and freelancers. During the beta testing period, the plaza had already accumulated over a thousand Smart Tips.
They roughly fall into three types. Like CQ, those who break down an entire business process into standardized components belong to the "workflow architect" category; some focus on polishing individual steps to make instructions run faster and more stably, creating "skill optimization" Smart Tips; there are also "vertical scenario" Smart Tips tailored for specific niche scenes, such as cross-border e-commerce product selection comparisons or Xiaohongshu viral title generators. Few people use them, but the demand is precise.
The reason these Smart Tips have emerged in large numbers is that they grow out of real scenarios. Office work hides massive amounts of trivial content. For example, writers want to quickly extract video transcripts, HR staff need to screen dozens of resumes, and administrative personnel pull over a dozen reimbursement forms with different formats every month. These needs are so granular that big tech companies generally aren't willing to invest in dedicated projects for them.
Tabbit breaks this dilemma. Users describe their needs in one sentence, and it generates executable Smart Tips. Sometimes, something written in 20 minutes can plug a long-standing pain point.
This sense of "if you can think it, you can do it" brings people great joy. CQ and CPU-Broken probably haven't realized yet that they've already brushed against the boundary of a new profession.
Neither CQ nor CPU-Broken are programmers, but they both know which parts of their industry are most tedious and error-prone. They broke down that experience into describable, reusable steps and turned them into tools others can use directly.
It's hard to classify them under traditional occupational directories. PwC and Deloitte also have consultants specializing in business process optimization, but those are enterprise-level system transformations requiring code and BPM software, serving organizations.
What CQ and his peers do is much finer-grained. Starting from their own work pain points, they use natural language to break down procedures into AI-executable instructions, package them as one-click reusable Smart Tips, solve their own problems first, then casually post them on the plaza.
If we must name this profession, it could be called "Smart Tip Engineer." They don't write code; they only deconstruct problems, breaking down head-scratching steps into lightweight tools callable with a single sentence.
Xiao Zeng follows a similar path. She's a senior art student. This summer, while writing her graduation thesis, her advisor required her to use FSQCA, a quantitative research method she had never encountered. The software has an all-English interface, and she couldn't even recognize most of the buttons. In the past, this kind of work would require flipping through tutorials, grinding through literature, and manually organizing case data herself, taking half a month just to grasp the concepts.
To get up to speed quickly, she built a "Academic Concept Explainer" Smart Tip. Whenever she encountered unfamiliar terminology, she would feed it in for explanations of definitions, logic, and application scenarios, along with recommended related literature. For dozens of papers, she grouped them into three topic-based label sets and asked Tabbit to extract core viewpoints, quickly breaking down practical applications of quantitative research cases.
From getting started with quantitative research methods to organizing data structures, she relied throughout on Tabbit to boost her research efficiency, saving time for polishing argument logic and writing content, ultimately earning an outstanding graduation thesis.
Fig|Xiao Zeng completed her graduation thesis with Tabbit's assistance
The "Academic Concept Explainer" Smart Tip was later uploaded to the plaza by her. She also made "Job Hunting and Flight Price Comparison" Smart Tips, which have hundreds of users.
Although CQ's Smart Tip isn't public, three colleagues in his department have already reused it across other business lines.
Then there's another type of person: those with deep work experience and sharp eyes, who can "pan for gold" in the Smart Tips Plaza.
Chen Lin is one of them. She had already been spending over a month in the plaza, trying out more than twenty Smart Tips. When she found ones that worked well, she casually shared them around.
She immediately dropped the "Bilibili Video Transcript Extractor" into her work group chat, and within a few days, two colleagues were using it too. She also found an "AI Daily Report" Smart Tip that automatically grabs industry hotspots every day, and passed it along to her dad, who worries about falling behind the times.
After browsing enough, Chen Lin started "renovating" according to her own ideas. The plaza had a "PPT Wizard" that could auto-generate outlines and layouts based on themes. After a few rounds of testing, she felt the template library was too thin and it could only operate on web. She asked Tabbit to help modify it, embedding several commonly used corporate templates and adding an auto-export-to-PDF instruction, creating her own "upgraded version."
This pattern of "browsing, discovering, sharing, and casually polishing" is essentially panning for gold on behalf of others, very similar to contributors on GitHub who submit PRs (code improvements) to open-source projects.
If the community ecosystem matures further, people like Chen Lin could potentially grow into "Smart Tip Hunters."
The difference between hunters and ordinary users is that they don't just stop at discovery and sharing—they actively tinker and改造. For generic tools on the plaza like "Resume Screening," hunters will take them and, drawing on their industry experience, transform them into "Internet Operations Role Specialized Version" or "Foreign Enterprise Finance Role Specialized Version."
They adjust prompts, add preconditions, optimize output formats, turning originally rough generic tools into specialized versions tailored to specific scenarios.
Such transformation requires sufficient understanding of business scenarios, knowing which parameters to adjust and which steps to remove. After finishing, hunters share these "adapted versions" with colleagues in the same industry or package them into "specific-industry workflow toolkits."
When platform incentives and revenue-sharing mechanisms mature, some people may be willing to pay for these professionally customized and improved versions.
When the number of Smart Tips on the plaza grows from thousands to tens of thousands or even hundreds of thousands, a third role will be born—Smart Tip Reviewers.
Because producing Smart Tips becomes extremely easy, the entire plaza might see hundreds or thousands of functionally similar "auto-write weekly reports" or "one-click Excel organizer." At that point, it will be hard for users to judge which ones work better, which can produce weekly reports with突出 highlights and clear structure, which frequently fail, and which harbor hidden surprises.
Smart Tip Reviewers are an opportunity born from just such needs. Unlike Smart Tip Hunters, they are those who eat the first crab on everyone's behalf and help make decisions. They might test similar Smart Tips, meticulously recording which ones error easily, which have messy output formats, and which can't run on certain browser versions.
Their value lies in building trust. Facing dozens of functionally similar Smart Tips, ordinary users don't have time to try each one. Reviewers save everyone those dozens of hours of trial-and-errorpitfalls.
They are a bit like digital product review bloggers, except the objects being reviewed are Smart Tips. They can test stability and error rates, compare differences among similar Smart Tips, publish recommendations like "Five Smart Tips Worth Watching This Week," or conduct in-depth horizontal reviews targeting vertical fields like graduate school exams or cross-border e-commerce.
Digital bloggers monetize through traffic and ads; Smart Tip Reviewers might future sustain themselves through platform creator incentive programs or paid subscriptions.
Some might ask: Since AI can already generate solutions on the fly from a single sentence, why go to the plaza to find pre-made Smart Tips? If everyone can make a Smart Tip in five minutes, why should yours be worth paying for?
Not all AI-generated content is worth paying for, but verified versions that solve specific pain points are. Anyone can write a Xiaohongshu note in five minutes, but professional content planners still command fees. The difference lies in systematization and trust.
When a hunter transforms a generic resume screener into a "Financial Risk Control Role Specialized Version," it embeds their long-term understanding of that role; when a reviewer tells you "this Smart Tip is stable on Mac, but that one tends to have issues on Windows," behind that is dozens of real-world tests. When generation becomes cheap, screening, modifying, and verifying become more valuable.
These three new professions haven't fully taken shape yet, but the seeds are already sprouting.
Ordinary folks who dare to try new things are the first to reap the dividends
Smart Tip Engineer, Smart Tip Hunter, and Smart Tip Reviewer—these three names sound somewhat futuristic. But similar scripts have already played out.
Over a decade ago, GitHub, the world's largest code hosting platform, was just a place for programmers to store code. The earliest developers never thought of making money from it. They simply uploaded tools they casually wrote so they could use them directly when switching computers next time.
Later, mechanisms like stars (liking projects), issues (submitting feedback), and PRs (submitting code improvements) gradually matured, and open-source projects slowly became career endorsements. Authors with high star counts found jobs more easily than those with fancy resumes. Some were headhunted via private messages, others hired as consultants at high fees, and still others received thousands of dollars monthly through GitHub Sponsors.
Notion was the same. Initially, users simply shared the workspace templates they had built, until the official launched a template marketplace, and some started making money.
In 2022, American creator Thomas Frank earned over $1 million annually selling Notion templates. Singaporean student Easlo's template business also scaled up; by March 2023, his cumulative income neared $500,000, maintaining six-figure annual revenue.
Tabbit's Smart Tips Plaza is still in its early stages, but the underlying logic is the same.
Right now, everyone is building skills, yet few pause to ask: Is this thing truly useful? Does it actually work? Can it really solve problems?
Smart Tip Engineers package their work experience into reusable tools; Hunters transform generic tools into industry-specific versions; Reviewers help踩坑排雷 on your behalf. What they fill in is the last kilometer from "can generate" to "can implement." Being able to solve practical problems may well be the reason they can achieve monetization in the future. Even if they can't become full-time careers, they might become side hustles that generate income.
User "Garlic Head" created two seemingly unrelated Smart Tips on the plaza.
One is "How Many Yuan Short for 618 Discounts? Algorithm Helps Bundle," embedding seven hard data validations and a six-dimensional weighted scoring system to help you optimize your shopping cart for maximum deals; the other is "Tabbit One-Stop Research Assistant." The former solves trivial daily annoyances, the latter tackles systematic academic challenges.
If the ecosystem matures a bit more, could these two precisely-targeted pain-point-solving tools also be packaged into subscription services like Notion templates, turning into tangible income?
That's the path of Smart Tip Engineers: helping people save time, and the time saved is worth a consultation fee.
Smart Tip Hunters rely on their understanding of business scenarios and hands-on改造 ability, which remain scarce even in the AI era; Reviewers rely on trust—when your recommendations are consistently reliable, followers may be willing to pay for that judgment.
From "finding it useful" to "willing to pay" still lies several steps away. Bounty features, rating systems, creator revenue-sharing—GitHub and Notion took five to ten years to gradually perfect these.
In the AI era, this speed might be much faster. GitHub's threshold was knowing how to code; Notion's was knowing how to build templates; Tabbit's threshold is simply one sentence.
When tool generation becomes unprecedentedly simple, everyone can get started easily. What will truly widen the gap shifts from technical ability to observation, discovery, summarization, and personal taste. Those who once had plenty of ideas but were stuck because they couldn't write code or build templates now finally have room to shine. They understand their industry's pain points best and know exactly which Smart Tips truly work.
Right now, they only need to describe their needs in one sentence; Tabbit can generate executable workflows, and in just minutes, ordinary people's creativity and experience can quickly become reusable tools solving long-standing pain points for many.
Fig|Describe needs in one sentence, Tabbit can make mini-games
Now is the window period for entry. The Tabbit Smart Tips Competition is underway, with a deadline of June 30 and top prizes of up to 10,000 yuan across three tracks. The platform will subsequently launch a creator incentive program, giving those truly pain-point-solving Smart Tips the chance to convert into side income.
If you haven't decided what to make yet, you can also just go browse the Smart Tips Plaza.
According to the competition's popularity ranking, as of June 22, the cumulative additions of the top 50 Smart Tips had surpassed 13,000. The top three are "Search Result Ad Blocker," "Video Summarizer," and "Image-to-Prompt Reverse Engineer." Scrolling further down, you'll find Bilibili transcript extraction, personal IP building, textbook downloading, AI hotspot collection… almost covering all aspects of office workers' daily lives.
Fig|Top three on the Smart Tips Competition popularity leaderboard
These top-ranked Smart Tips share one common trait: they all save users from repetitive labor. The leaderboard is essentially a real-time survey of "what office workers hate most."
The creators of these Smart Tips are mostly ordinary office workers and students. Like CQ and Xiao Zeng, they simply solved their own troubles first and casually shared them on the plaza.
Xiao Zeng has already uploaded several of her Smart Tips to the competition. She said creating a Smart Tip took only 5 minutes; she didn't necessarily expect to win, but it's a good opportunity to validate her ideas.
"Every time I see the usage count go up, I feel especially accomplished," Xiao Zeng said.
CQ looks further ahead. In his view, the platform and tools' value is already there; it's just that the transaction mechanism hasn't caught up. "Actually, we could add a bounty feature—someone posts a need, someone else takes the job to provide the service," he suggested.
History from both GitHub and Notion proves that those who enter earliest accumulate digital assets and reputation first. By the time the platform matures and users are willing to pay for efficiency, those with works and data will naturally land their first orders.
For ordinary people, what they should do most right now might not be waiting for monetization channels to open, but rather to stake out a position first.
- END -
撰文|杨 梅
编辑|张 霞
往期回顾




