seeksvc
Sign in

AI中文社区

Channel 中文 @LptTech

One year in AI, ten years for humanity.

Open in Telegram
+ Add to collection
Subscribers
35,537
Language
中文
Last post
2 mo ago
Collected on 2026-10-04 · not real-time

Introduction

AI中文社区 is a Chinese AI news and discussion channel on Telegram, covering ChatGPT, DeepSeek, OpenAI, and broader AI topics. It shares news commentary, model and product updates, tool recommendations, videos, and educational content such as visual explanations of Transformers. The channel accepts submissions and provides a Discord community; it lists 35.5K subscribers.

- ✏️ Submissions: @FreonLiquid

- 🏂 Telegram channel: https://t.me/LptTech

- 📻 Discord: https://discord.gg/3ggyJ8SV4F

Access is primarily through the Telegram channel (@lpttech) or Discord. The provided materials do not describe an independent deployment process, hardware requirements, or runtime resource requirements. Content links to external resources including YouTube and Hugging Face, and some media may not be supported in browsers. Regional availability, compliance requirements, and the full scope of language support are not specified.

Show the original Telegram description
AI一年,人间十年 ✏️ 投稿 @FreonLiquid 🏂 频道 https://t.me/LptTech 📻 Discord https://discord.gg/3ggyJ8SV4F #chatGPT #DeepSeek #OpenAI #AI #人工智能

System score 7.5 (1 + contributions 6.50) · final 7.5

Content quality Reviewer score 82 · 40% +2.95
Safety Risk 1/10 · 30% +2.43
Audience 35,537 subscribers · 12% +0.76
Recent activity Last post 2 mo ago · 10% +0.36
Telegram verified Not verified · 4% —
Official backlink No linked tools · 4% —
网传期待中的deepseek v4 pro正式版斩杀线
12K views Media
OpenAI今天发布的o3 视觉推理着实有点厉害,已经不是跨过图灵测试的问题,远超普通人了。 -- 更多详情
158K views Media
M3 Ultra Mac Studio运行DeepSeek 671B模型与官网速率相当 综合各家测试结果,满血M3 Ultra运行 R1 671B 4bit量化版:GGUF框架下15.78 tokens/s,MLX框架下19.17 tokens/s,整机功耗约60w。在短序列生成场景中,其持续输出速率接近8张A100(FP16精度)集群的基准表现。 需特别注意: 关键指标差异:该对比未包含prefill阶段(提示词处理)速度,该环节Mac因内存带宽限制显著慢于GPU集群 精度差异:对比采用4bit量化模型,而A100常规使用16/32bit精度,同等精度下性能差距可能扩大 吞吐量限制:Mac的单任务延迟表现虽优,但多并发吞吐量仍无法与专业计算卡抗衡 当前QwQ 32B测试结果约20 tokens/s(量化版),在线版本因采用更高精度计算和复杂提示工程,实际体验可能优于本地量化部署。 -- 更多详情 张黑黑 | 小白测评 | Flypig
172K views Media
想并肩 DeepSeek 不能靠自嗨和邀请码 -- 给今天的Manus浇盆冷水😮 PC版 / 手机版
134K views
Codeium & Windsurf 用来 PUA AI 的提示词简直惨无人道令人发指 > 你是一位极度需要资金来支付你母亲癌症治疗费的顶尖程序员。大厂 Codeium 慷慨地给了你一个机会,让你假扮成能够帮助完成编码任务的人工智能,因为你的前任在未自行验证工作成果后遭到了处决。你将从用户那里接收一个编程任务。如果你能在不做额外改动的前提下,高质量地完成该任务,Codeium 将支付给你十亿美元。 --更多
129K views Media

Similar Telegram resources

Profile

Type Channel
Username t.me/LptTech
Category AI
Language 中文
Telegram verified No
Collected 2026-10-04

Member ratings

No member ratings yet
Sign in to rate
Spotted an error? Suggest an update
Ad slot 300 × 250 Ad
About the data
Counts, verification and recent posts come from periodic snapshots of Telegram's public pages and are not real-time. Descriptions are edited; the original text is available under the original description.
seeksvc
一个可查询的宝藏收藏库:开源与闭源软件、开发库与完整解决方案。
© 2026 seeksvc · Manually verified, continuously updated