说黑话，角色扮演像傻子 代码能力下降，幻觉提升，上下文丢失 合着这回就升级了个风控是吧 1 个帖子 - 1 位参与者 阅读完整话题
如题, 我是wsl+ollama, 试了一个qwen3.5 2b 4b 9b, 都贼拉难用, 我就问他 1+1=? , 一直搁那吐字, 思考, 我记得以前2023年的ChatGLM没这么难用的啊, glm问他是谁他都正常吐字就完事了, qwen一直搁那思考自己是阿里巴巴开发的, 中文思考完了换英文, 半天不吐字, 介绍完自己, 又介绍阿里巴巴, 难用得要命了 13 个帖子 - 10 位参与者 阅读完整话题
先说说使用环境：自己的单个GPT PRO账号------CPA转接（这里主要是为了统一代理，不用每个机器装一个代理）------Axonhub（因为还有其他模型，一起管理）------opencode+OMO 在两台开发机上都发现同样的问题（不同的OC版本，一个1.3.17，一个1.4.8）：经常出现如图所示的报错，尤其感觉是写入长文本或者一条回复比较长报错概率会更大，非常影响开发。 我又分析了一下请求，似乎是思考/输出太长，openai先返回了一个空响应，但是在OC侧被错误处理了？ 请各位大师帮忙看看！ 我逐步审视： Axonhub接收到的相应如下： { id: "resp_0cc95817ee988d150169e25f7278ec8199a1176c6b7dafb37b" model: "gpt-5.4" object: "response" output[ ] status: "in_progress" created_at: 1776443251 } OC显示报错文本如下： Type validation failed: Value: {"error":{"code":"internal_server_error","message":"unexpected EOF","type":""}}. Error message: [ { "code": "invalid_union", "errors": [ [ { "code": "invalid_value", "values": [ "response.output_text.delta" ], "path": [ "type" ], "message": "Invalid input: expected \"response.output_text.delta\"" }, { "expected": "string", "code": "invalid_type", "path": [ "item_id" ], "message": "Invalid input: expected string, received undefined" }, { "expected": "string", "code": "invalid_type", "path": [ "delta" ], "message": "Invalid input: expected string, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.completed", "response.incomplete" ], "path": [ "type" ], "message": "Invalid option: expected one of \"response.completed\"|\"response.incomplete\"" }, { "expected": "object", "code": "invalid_type", "path": [ "response" ], "message": "Invalid input: expected object, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.failed" ], "path": [ "type" ], "message": "Invalid input: expected \"response.failed\"" }, { "expected": "object", "code": "invalid_type", "path": [ "response" ], "message": "Invalid input: expected object, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.created" ], "path": [ "type" ], "message": "Invalid input: expected \"response.created\"" }, { "expected": "object", "code": "invalid_type", "path": [ "response" ], "message": "Invalid input: expected object, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.output_item.added" ], "path": [ "type" ], "message": "Invalid input: expected \"response.output_item.added\"" }, { "expected": "number", "code": "invalid_type", "path": [ "output_index" ], "message": "Invalid input: expected number, received undefined" }, { "code": "invalid_type", "expected": "object", "path": [ "item" ], "message": "Invalid input: expected object, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.output_item.done" ], "path": [ "type" ], "message": "Invalid input: expected \"response.output_item.done\"" }, { "expected": "number", "code": "invalid_type", "path": [ "output_index" ], "message": "Invalid input: expected number, received undefined" }, { "code": "invalid_type", "expected": "object", "path": [ "item" ], "message": "Invalid input: expected object, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.function_call_arguments.delta" ], "path": [ "type" ], "message": "Invalid input: expected \"response.function_call_arguments.delta\"" }, { "expected": "string", "code": "invalid_type", "path": [ "item_id" ], "message": "Invalid input: expected string, received undefined" }, { "expected": "number", "code": "invalid_type", "path": [ "output_index" ], "message": "Invalid input: expected number, received undefined" }, { "expected": "string", "code": "invalid_type", "path": [ "delta" ], "message": "Invalid input: expected string, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.custom_tool_call_input.delta" ], "path": [ "type" ], "message": "Invalid input: expected \"response.custom_tool_call_input.delta\"" }, { "expected": "string", "code": "invalid_type", "path": [ "item_id" ], "message": "Invalid input: expected string, received undefined" }, { "expected": "number", "code": "invalid_type", "path": [ "output_index" ], "message": "Invalid input: expected number, received undefined" }, { "expected": "string", "code": "invalid_type", "path": [ "delta" ], "message": "Invalid input: expected string, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.image_generation_call.partial_image" ], "path": [ "type" ], "message": "Invalid input: expected \"response.image_generation_call.partial_image\"" }, { "expected": "string", "code": "invalid_type", "path": [ "item_id" ], "message": "Invalid input: expected string, received undefined" }, { "expected": "number", "code": "invalid_type", "path": [ "output_index" ], "message": "Invalid input: expected number, received undefined" }, { "expected": "string", "code": "invalid_type", "path": [ "partial_image_b64" ], "message": "Invalid input: expected string, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.code_interpreter_call_code.delta" ], "path": [ "type" ], "message": "Invalid input: expected \"response.code_interpreter_call_code.delta\"" }, { "expected": "string", "code": "invalid_type", "path": [ "item_id" ], "message": "Invalid input: expected string, received undefined" }, { "expected": "number", "code": "invalid_type", "path": [ "output_index" ], "message": "Invalid input: expected number, received undefined" }, { "expected": "string", "code": "invalid_type", "path": [ "delta" ], "message": "Invalid input: expected string, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.code_interpreter_call_code.done" ], "path": [ "type" ], "message": "Invalid input: expected \"response.code_interpreter_call_code.done\"" }, { "expected": "string", "code": "invalid_type", "path": [ "item_id" ], "message": "Invalid input: expected string, received undefined" }, { "expected": "number", "code": "invalid_type", "path": [ "output_index" ], "message": "Invalid input: expected number, received undefined" }, { "expected": "string", "code": "invalid_type", "path": [ "code" ], "message": "Invalid input: expected string, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.output_text.annotation.added" ], "path": [ "type" ], "message": "Invalid input: expected \"response.output_text.annotation.added\"" }, { "code": "invalid_type", "expected": "object", "path": [ "annotation" ], "message": "Invalid input: expected object, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.reasoning_summary_part.added" ], "path": [ "type" ], "message": "Invalid input: expected \"response.reasoning_summary_part.added\"" }, { "expected": "string", "code": "invalid_type", "path": [ "item_id" ], "message": "Invalid input: expected string, received undefined" }, { "expected": "number", "code": "invalid_type", "path": [ "summary_index" ], "message": "Invalid input: expected number, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.reasoning_summary_text.delta" ], "path": [ "type" ], "message": "Invalid input: expected \"response.reasoning_summary_text.delta\"" }, { "expected": "string", "code": "invalid_type", "path": [ "item_id" ], "message": "Invalid input: expected string, received undefined" }, { "expected": "number", "code": "invalid_type", "path": [ "summary_index" ], "message": "Invalid input: expected number, received undefined" }, { "expected": "string", "code": "invalid_type", "path": [ "delta" ], "message": "Invalid input: expected string, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.reasoning_summary_part.done" ], "path": [ "type" ], "message": "Invalid input: expected \"response.reasoning_summary_part.done\"" }, { "expected": "string", "code": "invalid_type", "path": [ "item_id" ], "message": "Invalid input: expected string, received undefined" }, { "expected": "number", "code": "invalid_type", "path": [ "summary_index" ], "message": "Invalid input: expected number, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.apply_patch_call_operation_diff.delta" ], "path": [ "type" ], "message": "Invalid input: expected \"response.apply_patch_call_operation_diff.delta\"" }, { "expected": "string", "code": "invalid_type", "path": [ "item_id" ], "message": "Invalid input: expected string, received undefined" }, { "expected": "number", "code": "invalid_type", "path": [ "output_index" ], "message": "Invalid input: expected number, received undefined" }, { "expected": "string", "code": "invalid_type", "path": [ "delta" ], "message": "Invalid input: expected string, received undefined" } ], [ { "code": "invalid_value", "values": [ "response.apply_patch_call_operation_diff.done" ], "path": [ "type" ], "message": "Invalid input: expected \"response.apply_patch_call_operation_diff.done\"" }, { "expected": "string", "code": "invalid_type", "path": [ "item_id" ], "message": "Invalid input: expected string, received undefined" }, { "expected": "number", "code": "invalid_type", "path": [ "output_index" ], "message": "Invalid input: expected number, received undefined" }, { "expected": "string", "code": "invalid_type", "path": [ "diff" ], "message": "Invalid input: expected string, received undefined" } ], [ { "code": "invalid_value", "values": [ "error" ], "path": [ "type" ], "message": "Invalid input: expected \"error\"" }, { "expected": "number", "code": "invalid_type", "path": [ "sequence_number" ], "message": "Invalid input: expected number, received undefined" } ], [ { "expected": "string", "code": "invalid_type", "path": [ "type" ], "message": "Invalid input: expected string, received undefined" } ] ], "path": [], "message": "Invalid input" } ] 1 个帖子 - 1 位参与者 阅读完整话题
我现在都是用gpt或者Gemini，但是听大家说国产有个glm挺厉害的，我平时没用过，想请问一下有没有使用过的佬，能说一下它个gpt5.4或者Gemini pro比咋样吗？（各个方面都可以，项目代码或者debug测试啥的这些方面，到底那个强呀） 7 个帖子 - 7 位参与者 阅读完整话题
https://linux.do/t/topic/1981613/ 接上贴，已经改到140个版本了，今天凌晨逛l站遇到一位佬分享了反重力的ultra账号，用opus4.6写了几轮也不行，真难受啊 15 个帖子 - 7 位参与者 阅读完整话题
以下为我个人体验，成功率从高到底： 谷歌别名邮箱、qq邮箱我自己的邮箱都注册完了，成功率100% 163邮箱我只注册了一个，也成功了，不确定是否是个例 2925邮箱今天试了好多回，成功率在60% 70%左右。 duckduckgo邮箱，一两周之前是成功率100%，现在貌似全都会跳add phone 我之前薅羊毛20块买的.com域名，整了个域名邮箱，一个都没成功，成功率0% 没提到的都是我没用过的，比如hotmail邮箱，听说现在也不行了？ 7 个帖子 - 4 位参与者 阅读完整话题
真的是想尽了一切办法 愣是没有成功过 只能眼巴巴来眼馋你们了 有没有什么奇门技巧指导指导呀 8 个帖子 - 8 位参与者 阅读完整话题
IT之家 4 月 17 日消息，国家市场监督管理总局今日向美团下发了《行政处罚决定书》，认定平台未依法履行资质审核义务。 美团随后回应称，将按照食品安全监管要求落实整改措施，并在各级监管部门指导下，重点打击餐饮转单、利用虚假资质上线等违法违规行为及黑灰产产业链。 同日，美团发布“放心外卖”十大举措，宣布全面升级外卖食品安全治理体系，围绕“商家准入、透明经营、多方共治”三个环节推进。 在商家准入方面，美团已与全国 20 余个省市对接许可证数据接口，实现商家证照与监管部门数据库 100% 穿透式对接验真。 自 2026 年 5 月起，全国餐饮商家在平台上线 30 日内，须提交完整、连续、带定位的“一镜到底”门店视频，平台将结合线下巡检与大众点评消费者实拍进行交叉验证 。针对已上线商家，美团将重点核验门店地址真实性与现场环境卫生，对高风险商家开启自检、平台巡检、AI 巡检、第三方飞检四道核验程序。 在透明经营方面，美团继续扩大“明厨亮灶”覆盖范围， 目前已有超过 40 万商家开通后厨直播 ，并为中小商家提供专项支持，鼓励其通过“食安日记”等功能展示后厨信息。平台为商家建立“食安透明橱窗”，向消费者标注堂食、明厨亮灶等食安信息，食安相册功能覆盖全量商家。同时，美团升级食品安全 AI 大模型“星眸”，当前日均后厨巡检超过 1400 万次， 升级后虚假门头图、虚假环境图识别功能将覆盖全部商家 。 在多方共治方面，美团推出多项举措。骑手“食安随手拍”已覆盖全国，鼓励骑手在日常配送中发现食安线索，平台对提供有效线索的骑手给予专项奖励。“小美评审团 · 食安评审官”机制已覆盖全国超 90 个城市，邀请消费者参与食安标准讨论和线下门店巡检。 此外， 美团将上线全国首个“食安政企共治平台”，每日主动上报并滚动更新食安线索 ，监管部门可在线查询骑手上报、消费者评价、平台巡检等多源信息，支持在线协查和治理下线。针对伪造证照、协助违规上线、一店多开、违规转单等黑灰产行为，美团将依托智能大模型主动预警，与公安机关建立常态化线索移送和联合打击机制，并联合多家电商及内容平台协同治理。 IT 之家小伙伴记得用 最会买点外卖 ，享受折扣价的同时还可以获得返利，单单都能省钱！ 相关阅读： 《 共处罚没款 35.97 亿元！拼多多、美团、京东、淘宝闪购、抖音、淘宝、天猫涉“幽灵外卖”系列案被处罚 》 《 严禁自动化高频抢票：携程、同程、去哪儿、飞猪、美团、智行、高铁管家被约谈 》 《 市场监管总局要求美团、京东、淘宝闪购三大外卖平台 6 月 1 日前完成外卖食品安全整改 》
DeepSeek v4版本从春节前就在讨论，说等圣诞放出来狙击美国佬一波，然后又说等到春节放出来，卷国内厂商，最近又听到说4月中旬会放出来，也没等到啊。这是发生了什么，是我们的期待太高，还是DeepSeek 有自己的节奏？ 11 个帖子 - 10 位参与者 阅读完整话题
怎么做 Harness Engineering.pdf (1.8 MB) 前两篇 【万字长文】什么是 Harness Engineering 文档共建 什么是 Harness Engineering.pdf (1.6 MB) 写了有四五天，用了AI帮助创作，收集了很多佬友的观点 聊聊为什么要 Harness Engineering，分享下我的观点 文档共建 继上一篇 【万字长文】什么是 Harness Engineering 为什么要 Harness Engineering.pdf (1.3 MB) 欢迎佬友们讨论 1 个帖子 - 1 位参与者 阅读完整话题
看到评论区说是要苹果手机，这个到底是怎么搞的，有人知道吗？ 1 个帖子 - 1 位参与者 阅读完整话题
很多人说opus价格贵了一倍不说，上下文能力也差了很多。有对比codex中的gpt5.4的吗？到底哪个好 1 个帖子 - 1 位参与者 阅读完整话题
看到好多人都是从tg 渠道获取的，有点好奇到底是什么渠道，有没有关键词？还是说这算商业机密？ 4 个帖子 - 3 位参与者 阅读完整话题
今日奥特曼大胜，A/大败 搞七捻三 A/到底是啥啊？这个在哪里玩呀。。。。。 这个话题就从这里开始吧，L站是著名的AI 剧透 学习论坛，但是我作为一个三级号，竟然对于很多基础知识一窍不通，比如什么cc switch、什么凹凸曼、什么中转站是如何练成的、什么蹬完了，啥的，只要是你知道的，我来者不拒，今天，让各位佬友人人献出一点光，用你们的AI知识助我破鼎 2 个帖子 - 2 位参与者 阅读完整话题
我暂时唯一用处就是想它我签到下PT站， 有些moviepilot插件都没法签到的， 想说用hermes用视觉 + 模拟鼠标点一下，怎么就是不行啊。 求佬友们教下。 2 个帖子 - 2 位参与者 阅读完整话题
hermes也好，openclaw也好 装完了都不知道该用来干嘛，除了定时任务，说说天气，搜搜财经新闻，主动触发聊天 真不知道还能干啥？ 8 个帖子 - 5 位参与者 阅读完整话题
今天又被qwen3.6的35b-a3b刷屏了。 都在说/吹 很强，到底是不是真的的？ 能不能正确流畅的在ai agent上操作。 另外M5 max 128G能部署流畅使用吗，200k上下文 1 个帖子 - 1 位参与者 阅读完整话题
体感上来说 回答速度大幅度提升 基本上秒回答 之前 opus 4.6 需要半分钟回答的问题 4.7 基本上几秒钟就回答了 尤其这个色盲问题 4.6 思考了七分钟 4.7 就思考了一分钟就回答出来了 并且答对了 所以到底水平是高还是低？ 2 个帖子 - 2 位参与者 阅读完整话题
https://x.com/bcherny/status/2044825898613187051 4 个帖子 - 4 位参与者 阅读完整话题
今年一直都在说是AI赚钱的时代，到底用AI怎么赚钱呢，又有谁用真正用ai赚到了钱呢？ 从另一个角度来说，是抢走了我们的工作，还是给了我们新的可能？ 14 个帖子 - 14 位参与者 阅读完整话题