{"topic_id":1781856,"version_number":1,"captured_at":"2026-03-19T07:59:04.536Z","is_encrypted":false,"main_post":{"post_number":1,"reply_to_post_number":null,"author_username":"shaonian_sun","created_at":"2026-03-19T04:25:20.756Z","content_html":"

最近，在用glm做开发，跟同事聊ai 编码，不知道为什么，听他们在聊copilot聊claude ，我觉得自己似乎低人一等。

","normalized_text":"最近，在用glm做开发，跟同事聊ai 编码，不知道为什么，听他们在聊copilot聊claude ，我觉得自己似乎低人一等。"},"replies":[{"post_number":2,"reply_to_post_number":null,"author_username":"inlife","created_at":"2026-03-19T04:26:00.014Z","content_html":"

只要你的成果吊打他们就行

","normalized_text":"只要你的成果吊打他们就行"},{"post_number":3,"reply_to_post_number":null,"author_username":"cloudwide","created_at":"2026-03-19T04:26:12.681Z","content_html":"

没有呀，无非是好不好用而已

","normalized_text":"没有呀，无非是好不好用而已"},{"post_number":4,"reply_to_post_number":null,"author_username":"KYJCASTER","created_at":"2026-03-19T04:26:20.868Z","content_html":"

低人一等不至于，但模型性能确实是不如国外

","normalized_text":"低人一等不至于，但模型性能确实是不如国外"},{"post_number":5,"reply_to_post_number":null,"author_username":"stevessr","created_at":"2026-03-19T04:26:45.455Z","content_html":"

不一定，用llama得低人二等

","normalized_text":"不一定，用llama得低人二等"},{"post_number":6,"reply_to_post_number":null,"author_username":"jjrhlb","created_at":"2026-03-19T04:26:57.465Z","content_html":"

你没有低人一等，模型有高下而已 \":smiling_face_with_three_hearts:\"

","normalized_text":"你没有低人一等，模型有高下而已"},{"post_number":7,"reply_to_post_number":null,"author_username":"binglau","created_at":"2026-03-19T04:27:14.701Z","content_html":"

不存在的，我都是混用，虽然国内模型不是主力

","normalized_text":"不存在的，我都是混用，虽然国内模型不是主力"},{"post_number":8,"reply_to_post_number":null,"author_username":"xg_ame","created_at":"2026-03-19T04:27:14.989Z","content_html":"

你不低人一等，但是国产模型自己觉得自己高人一等。
\n现在glm定价和claude比优势都不大了。

","normalized_text":"你不低人一等，但是国产模型自己觉得自己高人一等。 现在glm定价和claude比优势都不大了。"},{"post_number":9,"reply_to_post_number":null,"author_username":"Vincenter","created_at":"2026-03-19T04:27:16.413Z","content_html":"

不至于，好用就行，提升了工作效率，不过性能确实还有待提升

","normalized_text":"不至于，好用就行，提升了工作效率，不过性能确实还有待提升"},{"post_number":10,"reply_to_post_number":null,"author_username":"kkmm776","created_at":"2026-03-19T04:27:28.050Z","content_html":"

工具而已，有什么高低之分

","normalized_text":"工具而已，有什么高低之分"},{"post_number":11,"reply_to_post_number":null,"author_username":"wuzm137","created_at":"2026-03-19T04:27:40.264Z","content_html":"

模型有差距是事实，不过最终还是看产出吧，能够达成目标，用什么模型都没差的

","normalized_text":"模型有差距是事实，不过最终还是看产出吧，能够达成目标，用什么模型都没差的"},{"post_number":12,"reply_to_post_number":null,"author_username":"killermost","created_at":"2026-03-19T04:28:03.002Z","content_html":"

垃圾佬都在比用低价攒机

\n

你们也比下谁能用豆包搓出一个项目

","normalized_text":"垃圾佬都在比用低价攒机 你们也比下谁能用豆包搓出一个项目"},{"post_number":13,"reply_to_post_number":null,"author_username":"s97712","created_at":"2026-03-19T04:29:05.446Z","content_html":"

是的 特别是那种coding plan 低于19.9的

","normalized_text":"是的 特别是那种coding plan 低于19.9的"},{"post_number":14,"reply_to_post_number":12,"author_username":"stevessr","created_at":"2026-03-19T04:29:06.008Z","content_html":"

那叫古法编程 \":distorted_face:\"

","normalized_text":"那叫古法编程"},{"post_number":15,"reply_to_post_number":null,"author_username":"Waynez","created_at":"2026-03-19T04:29:29.017Z","content_html":"

我觉得上班之后炫耀自己拉磨的碾子比别人劲儿大，这件事才是…有点抽象的低人一等 \":distorted_face:\"

","normalized_text":"我觉得上班之后炫耀自己拉磨的碾子比别人劲儿大，这件事才是…有点抽象的低人一等"},{"post_number":16,"reply_to_post_number":null,"author_username":"zzqqa1.6","created_at":"2026-03-19T04:30:01.781Z","content_html":"

为何有这种想法

","normalized_text":"为何有这种想法"},{"post_number":17,"reply_to_post_number":null,"author_username":"shengdr","created_at":"2026-03-19T04:30:21.530Z","content_html":"

只是工具而已╮(¯▽¯)╭

","normalized_text":"只是工具而已╮(¯▽¯)╭"},{"post_number":18,"reply_to_post_number":null,"author_username":"jiege","created_at":"2026-03-19T04:31:47.828Z","content_html":"

能力上，国外还是厉害一点。

","normalized_text":"能力上，国外还是厉害一点。"},{"post_number":19,"reply_to_post_number":null,"author_username":"clbi88","created_at":"2026-03-19T04:32:28.551Z","content_html":"

生产力工具的选择，谁好用用谁

","normalized_text":"生产力工具的选择，谁好用用谁"},{"post_number":20,"reply_to_post_number":null,"author_username":"ydxlm","created_at":"2026-03-19T04:32:49.417Z","content_html":"

没怎么少

","normalized_text":"没怎么少"},{"post_number":21,"reply_to_post_number":null,"author_username":"Kew","created_at":"2026-03-19T04:37:48.594Z","content_html":"

谁好用，用谁呗

","normalized_text":"谁好用，用谁呗"},{"post_number":22,"reply_to_post_number":13,"author_username":"sczerg","created_at":"2026-03-19T04:38:13.520Z","content_html":"

Glm好贵的，最低档都要49元每月。

","normalized_text":"Glm好贵的，最低档都要49元每月。"},{"post_number":23,"reply_to_post_number":null,"author_username":"junmoxi","created_at":"2026-03-19T04:40:09.296Z","content_html":"

我感觉差不多，合适就好

","normalized_text":"我感觉差不多，合适就好"},{"post_number":24,"reply_to_post_number":12,"author_username":"kise_tyxml","created_at":"2026-03-19T04:40:24.196Z","content_html":"

豆包不够顶尖但其实也大模型，用本地7B模型做吧，真正的AI工程师从来不会嫌弃模型小 \":tieba_025:\"

","normalized_text":"豆包不够顶尖但其实也大模型，用本地7B模型做吧，真正的AI工程师从来不会嫌弃模型小"},{"post_number":25,"reply_to_post_number":null,"author_username":"geekbuf","created_at":"2026-03-19T04:40:43.532Z","content_html":"

用哪个模型是基于自己考量的，没必要非要用哪个，根据自身实际情况来选择。
\n你觉得低人一等的心态，其实是来自自我的焦虑，怕自己跟不上最新的模型技术，仅此而已。
\n那么面对这个问题，你在工作之余其实是可以去尝试体验，试试这些国外的模型的能力，体验体验，这样你也清楚这些模型的能力边界，L站也是大把的经验帖，多看多学，了解多了，也不会莫名的出现这个心理了。
\nBTW,个人在日常后端开发中反而偏好使用QWen来写java的项目尤其是Spring全家桶，虽然没有特别去对比是否跟claude/codex在上面的差别有多大，个人感觉是够用的。私以为阿里作为国内java技术栈集大成者，会不会喂了自身的项目代码经验呢？出于这个考虑，我就选了QWen。

","normalized_text":"用哪个模型是基于自己考量的，没必要非要用哪个，根据自身实际情况来选择。 你觉得低人一等的心态，其实是来自自我的焦虑，怕自己跟不上最新的模型技术，仅此而已。 那么面对这个问题，你在工作之余其实是可以去尝试体验，试试这些国外的模型的能力，体验体验，这样你也清楚这些模型的能力边界，L站也是大把的经验帖，多看多学，了解多了，也不会莫名的出现这个心理了。 BTW,个人在日常后端开发中反而偏好使用QWen来写java的项目尤其是Spring全家桶，虽然没有特别去对比是否跟claude/codex在上面的差别有多大，个人感觉是够用的。私以为阿里作为国内java技术栈集大成者，会不会喂了自身的项目代码经验呢？出于这个考虑，我就选了QWen。"},{"post_number":26,"reply_to_post_number":null,"author_username":"unclejimao","created_at":"2026-03-19T04:42:02.282Z","content_html":"

看过天下无贼么

\n

刘德华：开好车就是好人啊？

\n

用过高级模型的人就高级啊？

","normalized_text":"看过天下无贼么 刘德华：开好车就是好人啊？ 用过高级模型的人就高级啊？"},{"post_number":27,"reply_to_post_number":null,"author_username":"boy1981chow","created_at":"2026-03-19T04:42:22.608Z","content_html":"

所以你同事没觉得你低人一等，是你自己自认为低人一等，想发个帖子寻求认同/想要人反驳自己建立自信……？

","normalized_text":"所以你同事没觉得你低人一等，是你自己自认为低人一等，想发个帖子寻求认同/想要人反驳自己建立自信……？"},{"post_number":28,"reply_to_post_number":27,"author_username":"626","created_at":"2026-03-19T04:42:55.055Z","content_html":"

佬友攻击力还挺强哈哈

","normalized_text":"佬友攻击力还挺强哈哈"},{"post_number":29,"reply_to_post_number":null,"author_username":"XiaoDao01","created_at":"2026-03-19T04:43:18.127Z","content_html":"

经济实惠且能满足自己需求 有什么低人一等的 我也用国内模型 而且免费的我也用 钱要花在刀刃上 能省则省 能满足自己需求就可以啦

","normalized_text":"经济实惠且能满足自己需求 有什么低人一等的 我也用国内模型 而且免费的我也用 钱要花在刀刃上 能省则省 能满足自己需求就可以啦"},{"post_number":30,"reply_to_post_number":28,"author_username":"boy1981chow","created_at":"2026-03-19T04:44:25.219Z","content_html":"

我看他写的，他同事也没瞧不起他还咋的啊……我自己也买过glm，现在也有用minimax，我是不觉得自己低人一等……

","normalized_text":"我看他写的，他同事也没瞧不起他还咋的啊……我自己也买过glm，现在也有用minimax，我是不觉得自己低人一等……"},{"post_number":31,"reply_to_post_number":null,"author_username":"Cora_Tracy","created_at":"2026-03-19T04:45:04.185Z","content_html":"

怎么会，会用模型也是很重要的事情！！
\n你看看墨佬，专研模型的潜力比他们张口闭口模型的好坏更好，不要听他们的一面之词

","normalized_text":"怎么会，会用模型也是很重要的事情！！ 你看看墨佬，专研模型的潜力比他们张口闭口模型的好坏更好，不要听他们的一面之词"},{"post_number":32,"reply_to_post_number":null,"author_username":"xiaochangsheng","created_at":"2026-03-19T04:45:06.821Z","content_html":"

都差不多吧只是耗费时间多少的问题。

","normalized_text":"都差不多吧只是耗费时间多少的问题。"},{"post_number":33,"reply_to_post_number":null,"author_username":"Fanlzx","created_at":"2026-03-19T04:45:10.662Z","content_html":"

混合用呗 哪个好用用哪个

","normalized_text":"混合用呗 哪个好用用哪个"},{"post_number":34,"reply_to_post_number":27,"author_username":"shaonian_sun","created_at":"2026-03-19T04:45:58.208Z","content_html":"

出现这种情况，一般是被人说过的

","normalized_text":"出现这种情况，一般是被人说过的"},{"post_number":35,"reply_to_post_number":30,"author_username":"626","created_at":"2026-03-19T04:46:19.642Z","content_html":"

确实，很有道理
\n其实这个论坛也是聊外国模型更多一些哈哈

","normalized_text":"确实，很有道理 其实这个论坛也是聊外国模型更多一些哈哈"},{"post_number":36,"reply_to_post_number":null,"author_username":"user800","created_at":"2026-03-19T04:46:49.973Z","content_html":"

没什么低人一等的 等国外这些模型涨价了 大家都用不起还得回归 就跟windows系统一样 人家到时候全面收费了 肯定好多人也会换linux之类的去替代。主要是确实人家先进。 我记得我之前跟群里一些一个学校的聊 他一边说自己养龙虾用deepseek 一会说用老板本地部署的模型很牛，结果转头又说自己用的claude（第一次听说老板能本地部署claude的），我说那玩意又没开源，他又说api呗 感觉有种对牛弹琴的既视感

","normalized_text":"没什么低人一等的 等国外这些模型涨价了 大家都用不起还得回归 就跟windows系统一样 人家到时候全面收费了 肯定好多人也会换linux之类的去替代。主要是确实人家先进。 我记得我之前跟群里一些一个学校的聊 他一边说自己养龙虾用deepseek 一会说用老板本地部署的模型很牛，结果转头又说自己用的claude（第一次听说老板能本地部署claude的），我说那玩意又没开源，他又说api呗 感觉有种对牛弹琴的既视感"},{"post_number":37,"reply_to_post_number":null,"author_username":"haohao","created_at":"2026-03-19T04:47:10.275Z","content_html":"

只是自动化程度略低

","normalized_text":"只是自动化程度略低"},{"post_number":38,"reply_to_post_number":null,"author_username":"sniao","created_at":"2026-03-19T04:47:12.703Z","content_html":"

用国产模型并不低人一等啊，但是该说不说，用copilot确实多少有点低人一等了，它太落后了。

","normalized_text":"用国产模型并不低人一等啊，但是该说不说，用copilot确实多少有点低人一等了，它太落后了。"},{"post_number":39,"reply_to_post_number":null,"author_username":"qqppp","created_at":"2026-03-19T04:48:06.138Z","content_html":"

差距是有的，但是用不用的到是另外一回事，有些东西glm就能轻松驾驭，何必用cc

","normalized_text":"差距是有的，但是用不用的到是另外一回事，有些东西glm就能轻松驾驭，何必用cc"},{"post_number":40,"reply_to_post_number":null,"author_username":"igg8","created_at":"2026-03-19T04:48:28.515Z","content_html":"

看业务需求，不分好坏，你天天问AI\"你好\"，哪个模型都一样，但是涉及复杂的工程项目，模型就不一样了啊。 而且能在L发帖，不都是有免费的opus4.6了？？？？

","normalized_text":"看业务需求，不分好坏，你天天问AI\"你好\"，哪个模型都一样，但是涉及复杂的工程项目，模型就不一样了啊。 而且能在L发帖，不都是有免费的opus4.6了？？？？"},{"post_number":41,"reply_to_post_number":5,"author_username":"everyday_happy","created_at":"2026-03-19T04:48:29.415Z","content_html":"

用mistral说不定能低人三等

","normalized_text":"用mistral说不定能低人三等"},{"post_number":42,"reply_to_post_number":null,"author_username":"hum","created_at":"2026-03-19T04:50:05.161Z","content_html":"

自卑心作祟

","normalized_text":"自卑心作祟"},{"post_number":43,"reply_to_post_number":30,"author_username":"sczerg","created_at":"2026-03-19T04:50:16.451Z","content_html":"

Glm5软件开发跑分和gpt5.0持平，已经很强了，否则也不会收费那么贵。
\n其实国内最强的应该是kimi2.5，参数巨大，几乎是glm5的两倍不到一点。不过好像官方没编程套餐，可惜了。

","normalized_text":"Glm5软件开发跑分和gpt5.0持平，已经很强了，否则也不会收费那么贵。 其实国内最强的应该是kimi2.5，参数巨大，几乎是glm5的两倍不到一点。不过好像官方没编程套餐，可惜了。"},{"post_number":44,"reply_to_post_number":null,"author_username":"mr.kuma","created_at":"2026-03-19T04:50:37.976Z","content_html":"

无论是什么模型都只是工具，只要产出合格，工具用着顺手，用什么工具并不重要。

","normalized_text":"无论是什么模型都只是工具，只要产出合格，工具用着顺手，用什么工具并不重要。"},{"post_number":45,"reply_to_post_number":null,"author_username":"depression","created_at":"2026-03-19T04:50:55.384Z","content_html":"

达成目标即可！

","normalized_text":"达成目标即可！"},{"post_number":46,"reply_to_post_number":null,"author_username":"sun_rasie","created_at":"2026-03-19T04:51:35.669Z","content_html":"

工具罢了，能完成任务谁管你用啥，佬友大可不必哈

","normalized_text":"工具罢了，能完成任务谁管你用啥，佬友大可不必哈"},{"post_number":47,"reply_to_post_number":43,"author_username":"boy1981chow","created_at":"2026-03-19T04:51:59.962Z","content_html":"

有套餐吧，虽然不是专门的编程套餐还带很多服务，到处都在说额度低弄得我都不敢开（

","normalized_text":"有套餐吧，虽然不是专门的编程套餐还带很多服务，到处都在说额度低弄得我都不敢开（"},{"post_number":48,"reply_to_post_number":null,"author_username":"Jojo_Coco","created_at":"2026-03-19T04:52:31.846Z","content_html":"

越小白越觉得claude好，越是大佬越觉得claude强

","normalized_text":"越小白越觉得claude好，越是大佬越觉得claude强"},{"post_number":49,"reply_to_post_number":null,"author_username":"Jgckm","created_at":"2026-03-19T04:53:24.980Z","content_html":"

还在于自己怎么使用呀

","normalized_text":"还在于自己怎么使用呀"},{"post_number":50,"reply_to_post_number":null,"author_username":"Simmy4867","created_at":"2026-03-19T04:53:44.317Z","content_html":"

没啥低人一等，虽然事实上确实是国外模型能力强，但是你也得考虑成本等问题\":distorted_face:\"

","normalized_text":"没啥低人一等，虽然事实上确实是国外模型能力强，但是你也得考虑成本等问题"},{"post_number":51,"reply_to_post_number":null,"author_username":"linzee","created_at":"2026-03-19T04:53:50.568Z","content_html":"

能帮我摸鱼就行

","normalized_text":"能帮我摸鱼就行"},{"post_number":52,"reply_to_post_number":null,"author_username":"tomato.potato","created_at":"2026-03-19T04:53:55.970Z","content_html":"

和人没有关系，都是工具而已，不过你可以试试他们说的模型，确实有点差距。

","normalized_text":"和人没有关系，都是工具而已，不过你可以试试他们说的模型，确实有点差距。"},{"post_number":53,"reply_to_post_number":null,"author_username":"byte1","created_at":"2026-03-19T04:54:31.412Z","content_html":"

单论模型性能是有差距，但现实中人和模型构成闭环反馈，上限高低还是看人。

","normalized_text":"单论模型性能是有差距，但现实中人和模型构成闭环反馈，上限高低还是看人。"},{"post_number":54,"reply_to_post_number":null,"author_username":"Lenger","created_at":"2026-03-19T04:55:42.908Z","content_html":"

只要是免费的都行，配好上下文压缩mcp

","normalized_text":"只要是免费的都行，配好上下文压缩mcp"},{"post_number":55,"reply_to_post_number":null,"author_username":"tank_5478","created_at":"2026-03-19T04:57:59.988Z","content_html":"

他们模型那么好，效率肯定更高吧？让领导多给他们排点活

","normalized_text":"他们模型那么好，效率肯定更高吧？让领导多给他们排点活"},{"post_number":56,"reply_to_post_number":null,"author_username":"996","created_at":"2026-03-19T05:01:48.083Z","content_html":"

怎么说呢，国产的优势也就是价格便宜了，使用归使用，算不上低人一等

\n

在预算有限的情况下，合理分配是不错的选择

\n

日常让AI记个笔记，整理下文档，我还是调用的minimax，写代码则用gpt5.4

\n

至于claude ops，如果没有报销，这价格我是不会用的

","normalized_text":"怎么说呢，国产的优势也就是价格便宜了，使用归使用，算不上低人一等 在预算有限的情况下，合理分配是不错的选择 日常让AI记个笔记，整理下文档，我还是调用的minimax，写代码则用gpt5.4 至于claude ops，如果没有报销，这价格我是不会用的"},{"post_number":57,"reply_to_post_number":null,"author_username":"chenyapeng","created_at":"2026-03-19T05:11:30.827Z","content_html":"

glm是什么模型没了解过 \":rofl:\"

","normalized_text":"glm是什么模型没了解过"},{"post_number":58,"reply_to_post_number":56,"author_username":"sczerg","created_at":"2026-03-19T05:11:49.311Z","content_html":"

minimax的参数是229B
\nglm5的参数是754B
\nkimi2.5的参数是1100B

\n

你不能用minimax小杯的水准来代表国产模型平均水准，参数上去后，国产写代码还是很强的。（当然价格也上去了）

","normalized_text":"minimax的参数是229B glm5的参数是754B kimi2.5的参数是1100B 你不能用minimax小杯的水准来代表国产模型平均水准，参数上去后，国产写代码还是很强的。（当然价格也上去了）"},{"post_number":59,"reply_to_post_number":null,"author_username":"handsome","created_at":"2026-03-19T05:13:21.336Z","content_html":"

并没有。

","normalized_text":"并没有。"},{"post_number":60,"reply_to_post_number":null,"author_username":"Killua1140","created_at":"2026-03-19T05:13:59.488Z","content_html":"

干活的工具罢了能用就行

","normalized_text":"干活的工具罢了能用就行"},{"post_number":61,"reply_to_post_number":null,"author_username":"gdgsbs","created_at":"2026-03-19T05:14:44.464Z","content_html":"

工具差异，谈不上低人一等吧

","normalized_text":"工具差异，谈不上低人一等吧"},{"post_number":62,"reply_to_post_number":null,"author_username":"bra","created_at":"2026-03-19T05:15:55.575Z","content_html":"

我都好久没有过国外ai了

","normalized_text":"我都好久没有过国外ai了"},{"post_number":63,"reply_to_post_number":27,"author_username":"Bree","created_at":"2026-03-19T05:19:55.478Z","content_html":"

好像佬说的没毛病，如果是我的话，你感兴趣可以试用一下，很多途径，用了再判断，适合自己就好，也不用在意国不国产

","normalized_text":"好像佬说的没毛病，如果是我的话，你感兴趣可以试用一下，很多途径，用了再判断，适合自己就好，也不用在意国不国产"},{"post_number":64,"reply_to_post_number":null,"author_username":"FIMISS","created_at":"2026-03-19T05:28:05.542Z","content_html":"

不至于，国内模型的价格便宜，响应速度快。但你要说质量嘛，那肯定贵有贵的道理…
\n俗话说跑分没输过，实际没赢过（）

","normalized_text":"不至于，国内模型的价格便宜，响应速度快。但你要说质量嘛，那肯定贵有贵的道理… 俗话说跑分没输过，实际没赢过（）"},{"post_number":65,"reply_to_post_number":null,"author_username":"Rainforest","created_at":"2026-03-19T05:28:51.241Z","content_html":"

GLM模型也还可以吧

","normalized_text":"GLM模型也还可以吧"},{"post_number":66,"reply_to_post_number":null,"author_username":"honor200","created_at":"2026-03-19T05:31:18.150Z","content_html":"

个人认为，在某些方面，GLM5比御三家好用 \":bili_057:\"

","normalized_text":"个人认为，在某些方面，GLM5比御三家好用"},{"post_number":67,"reply_to_post_number":null,"author_username":"ssword","created_at":"2026-03-19T05:31:46.558Z","content_html":"

不管白猫黑猫，能抓老鼠的就是好猫

","normalized_text":"不管白猫黑猫，能抓老鼠的就是好猫"},{"post_number":68,"reply_to_post_number":null,"author_username":"ikun","created_at":"2026-03-19T05:32:55.433Z","content_html":"

不是，其实低了3-4等的。

","normalized_text":"不是，其实低了3-4等的。"},{"post_number":69,"reply_to_post_number":null,"author_username":"233x","created_at":"2026-03-19T05:38:41.970Z","content_html":"

没问题的，能用到点子上也很好了

","normalized_text":"没问题的，能用到点子上也很好了"},{"post_number":70,"reply_to_post_number":null,"author_username":"leeflour","created_at":"2026-03-19T05:40:16.738Z","content_html":"

都是工具而已

","normalized_text":"都是工具而已"},{"post_number":71,"reply_to_post_number":null,"author_username":"ayoka","created_at":"2026-03-19T05:41:04.067Z","content_html":"

问题是不好用

","normalized_text":"问题是不好用"},{"post_number":72,"reply_to_post_number":null,"author_username":"554705239","created_at":"2026-03-19T05:41:06.631Z","content_html":"

只有比财务才能体会到，其他的都不重要吧

","normalized_text":"只有比财务才能体会到，其他的都不重要吧"},{"post_number":73,"reply_to_post_number":null,"author_username":"xdtz","created_at":"2026-03-19T05:41:07.002Z","content_html":"

你们这是公司统一的吗，如果不是，牛马还要互相比较自己搞的工具么

","normalized_text":"你们这是公司统一的吗，如果不是，牛马还要互相比较自己搞的工具么"},{"post_number":74,"reply_to_post_number":null,"author_username":"yantu6489","created_at":"2026-03-19T05:50:06.526Z","content_html":"

便宜好用,日常干活glm-5，偶尔抽风就换kimi或m2, 方案深层设计和架构方案调整找gemini或claude

","normalized_text":"便宜好用,日常干活glm-5，偶尔抽风就换kimi或m2, 方案深层设计和架构方案调整找gemini或claude"},{"post_number":75,"reply_to_post_number":null,"author_username":"kingor","created_at":"2026-03-19T05:50:48.459Z","content_html":"

模型编程综合能力上claude和gpt确实比国内的强一些，但是很多场景其实模型之间并没有特别大的差距。最后其实还是看模型使用者对项目的把控能力，能够把需求拆解、分步足够清晰其实会比大脑一键托管到高性能模型要重要的多。而且国内模型便宜，试错回滚都很方便不是么 \":face_with_peeking_eye:\"

","normalized_text":"模型编程综合能力上claude和gpt确实比国内的强一些，但是很多场景其实模型之间并没有特别大的差距。最后其实还是看模型使用者对项目的把控能力，能够把需求拆解、分步足够清晰其实会比大脑一键托管到高性能模型要重要的多。而且国内模型便宜，试错回滚都很方便不是么"},{"post_number":76,"reply_to_post_number":null,"author_username":"efls","created_at":"2026-03-19T05:51:20.891Z","content_html":"

（帖子已被作者删除）

","normalized_text":"（帖子已被作者删除）"},{"post_number":77,"reply_to_post_number":null,"author_username":"efls","created_at":"2026-03-19T05:58:05.441Z","content_html":"

能达到想要的结果就行，都可以试试。国内的在价格、稳定性还是有优势的。

","normalized_text":"能达到想要的结果就行，都可以试试。国内的在价格、稳定性还是有优势的。"},{"post_number":78,"reply_to_post_number":77,"author_username":"efls","created_at":"2026-03-19T06:00:39.503Z","content_html":"

还有就是个人能力也挺重要的，哪怕是国内模型，用得好效率也是翻倍的。

","normalized_text":"还有就是个人能力也挺重要的，哪怕是国内模型，用得好效率也是翻倍的。"},{"post_number":79,"reply_to_post_number":null,"author_username":"1994","created_at":"2026-03-19T06:01:09.045Z","content_html":"

模型能力是實打實的

","normalized_text":"模型能力是實打實的"},{"post_number":80,"reply_to_post_number":null,"author_username":"weisili2025","created_at":"2026-03-19T06:03:57.765Z","content_html":"

以前都不用的大神不一样开发出好产品吗？

","normalized_text":"以前都不用的大神不一样开发出好产品吗？"},{"post_number":81,"reply_to_post_number":null,"author_username":"ceshi01Adx","created_at":"2026-03-19T06:04:44.223Z","content_html":"

你这算是在引战吗？没人会觉得用国产模型低人一等，相反国产模型性价比很高，只是干同样的活要花更多的精力去返工..

","normalized_text":"你这算是在引战吗？没人会觉得用国产模型低人一等，相反国产模型性价比很高，只是干同样的活要花更多的精力去返工.."},{"post_number":82,"reply_to_post_number":null,"author_username":"wanya","created_at":"2026-03-19T06:06:05.560Z","content_html":"

他们应该是惊讶嫉妒，我翻山越岭用gpt看ac的脸色才能蹬出来，你居然用glm就能？ \":laughing:\"

","normalized_text":"他们应该是惊讶嫉妒，我翻山越岭用gpt看ac的脸色才能蹬出来，你居然用glm就能？"},{"post_number":83,"reply_to_post_number":40,"author_username":"xingxingzi","created_at":"2026-03-19T06:06:22.915Z","content_html":"

求助，佬，免费的opus4.6是怎么得捏？

","normalized_text":"求助，佬，免费的opus4.6是怎么得捏？"},{"post_number":84,"reply_to_post_number":null,"author_username":"yomi_k","created_at":"2026-03-19T06:06:58.282Z","content_html":"

不是低不低的问题 是能力不行 影响效率

","normalized_text":"不是低不低的问题 是能力不行 影响效率"},{"post_number":85,"reply_to_post_number":null,"author_username":"DDKing","created_at":"2026-03-19T06:11:18.104Z","content_html":"

没用国外模型不代表不能聊国外模型吧，感觉只是个聊头而已，没有啥高人一等低人一等的

","normalized_text":"没用国外模型不代表不能聊国外模型吧，感觉只是个聊头而已，没有啥高人一等低人一等的"},{"post_number":86,"reply_to_post_number":null,"author_username":"caiji","created_at":"2026-03-19T06:12:13.962Z","content_html":"

现在用个ai也有饭圈了嘛？

","normalized_text":"现在用个ai也有饭圈了嘛？"},{"post_number":87,"reply_to_post_number":null,"author_username":"hlool","created_at":"2026-03-19T06:14:12.419Z","content_html":"

除非国内开放，不然国产就是国内主流，工具而已。

","normalized_text":"除非国内开放，不然国产就是国内主流，工具而已。"},{"post_number":88,"reply_to_post_number":null,"author_username":"bruclan","created_at":"2026-03-19T06:14:37.507Z","content_html":"

绝对不是，去 reddit 的 localllama 讨论组看看啊，国外的老哥普遍很喜欢阿里的 qwen 系列。有什么低不低人一等，好用了就行。

","normalized_text":"绝对不是，去 reddit 的 localllama 讨论组看看啊，国外的老哥普遍很喜欢阿里的 qwen 系列。有什么低不低人一等，好用了就行。"},{"post_number":89,"reply_to_post_number":null,"author_username":"DannisRen","created_at":"2026-03-19T06:16:12.376Z","content_html":"

都TMD干活的 谁瞧不起谁呢 这种人脑子有粑粑

","normalized_text":"都TMD干活的 谁瞧不起谁呢 这种人脑子有粑粑"},{"post_number":90,"reply_to_post_number":null,"author_username":"MorningStar_Lucif","created_at":"2026-03-19T06:17:37.308Z","content_html":"

哈哈哈，祥子你看看你的破黄包车，还没有我的多轴联动拉的轻松呢

","normalized_text":"哈哈哈，祥子你看看你的破黄包车，还没有我的多轴联动拉的轻松呢"},{"post_number":91,"reply_to_post_number":null,"author_username":"frank8","created_at":"2026-03-19T06:17:45.628Z","content_html":"

跟人没关系，说的模型能力，国外的也有辣鸡模型，能解决问题的都是好模型 \":bili_040:\"

","normalized_text":"跟人没关系，说的模型能力，国外的也有辣鸡模型，能解决问题的都是好模型"},{"post_number":92,"reply_to_post_number":null,"author_username":"xin_zhang","created_at":"2026-03-19T06:18:42.028Z","content_html":"

至于么，都是以实现目的而已，这个又不是比自身硬实力

","normalized_text":"至于么，都是以实现目的而已，这个又不是比自身硬实力"},{"post_number":93,"reply_to_post_number":43,"author_username":"xiaox0321","created_at":"2026-03-19T06:26:33.261Z","content_html":"

kimi code的那个算不上官方套餐

","normalized_text":"kimi code的那个算不上官方套餐"},{"post_number":94,"reply_to_post_number":null,"author_username":"magicyan418","created_at":"2026-03-19T06:29:46.473Z","content_html":"

能用最低的成本解决事并且有时间摸鱼就行了，L站帖子都刷完了？

","normalized_text":"能用最低的成本解决事并且有时间摸鱼就行了，L站帖子都刷完了？"},{"post_number":95,"reply_to_post_number":null,"author_username":"NeonEcho","created_at":"2026-03-19T06:58:33.397Z","content_html":"

你同事没觉得你低人一等，反而是你自己觉得你低人一等，都是拉磨的工具没有高低之分，你自己大可试试国外的模型能力如何，无论是编码能力还是理解能力，适合自己的工具才是最好的

","normalized_text":"你同事没觉得你低人一等，反而是你自己觉得你低人一等，都是拉磨的工具没有高低之分，你自己大可试试国外的模型能力如何，无论是编码能力还是理解能力，适合自己的工具才是最好的"},{"post_number":96,"reply_to_post_number":null,"author_username":"richOvernight","created_at":"2026-03-19T07:00:34.758Z","content_html":"

没有低人一等，但是有差距也是事实

","normalized_text":"没有低人一等，但是有差距也是事实"},{"post_number":97,"reply_to_post_number":null,"author_username":"SeletubbY","created_at":"2026-03-19T07:06:25.656Z","content_html":"

你用的glm说不定比他们的贵多了

","normalized_text":"你用的glm说不定比他们的贵多了"},{"post_number":98,"reply_to_post_number":null,"author_username":"loyol","created_at":"2026-03-19T07:15:35.219Z","content_html":"

会不会是你自己自卑
\n自己觉得低人一等吧。。。
\n服了
\n遇到这种情况，我只会去试试，到底哪个好用

","normalized_text":"会不会是你自己自卑 自己觉得低人一等吧。。。 服了 遇到这种情况，我只会去试试，到底哪个好用"},{"post_number":99,"reply_to_post_number":null,"author_username":"NAXG","created_at":"2026-03-19T07:29:30.233Z","content_html":"

你自己心里都认为了

","normalized_text":"你自己心里都认为了"}]}