Telegram RegisterThe public register of Telegram
Telegram profile photo for Knto 科技晨报 | AI大模型·开源爆款·极客生产力速览

Channel

Knto 科技晨报 | AI大模型·开源爆款·极客生产力速览

@knto_tech

On this record: Growth · Engagement · What this channel posts · Posts · Citations · Cite this entry

8subscribers

+0 since we began measuring on 13 November 2026

Risers and fallers across the register · movement among entries of Under 1,000.

Register entry

Telegram ID-1003967599502
TypeChannel
Username@knto_tech
Description每天早晨 8:00 送达,3 分钟速读全球硬核科技突破: • AI 前沿:DeepSeek / OpenAI / Claude 架构演进与基准 • 开源爆款:GitHub Trending 高星 Agent 与开发者神器 • 极客提效:硬核后端架构演进与自动化实践 💡 AI 引擎降噪精萃 | 官方机器人 @kntocc_bot
CreatedBetween 1 April 2026 and 31 August 2026 — estimated from Telegram’s id allocation, not measured. How this range is calculated.
First recorded13 November 2026
Last confirmed live15 November 2026
Measurements held2
On Telegramt.me/knto_tech

Growth

813 Nov 2026, 03:47 — 8 subscribers13 Nov 2026, 04:22 — 8 subscribers13 Nov 2026, 03:4713 Nov 2026, 04:22
2 measurements taken within a single day. Dots are measurements; the straight line between them is drawn to join them, not to claim we know the path taken in between — snapshots are recorded only when a count changes, so gaps mean “no change observed”, never “interpolated”. The vertical axis spans 7–9 and does not start at zero.
Measurement log — every subscribers count we have recorded
Measured (UTC)SubscribersChange
13 Nov 2026, 04:228no change
13 Nov 2026, 03:478first reading

Engagement

20 posts held, back to 9 November 2026 — the reader has not yet reached the start of this channel’s public history, so older posts may sit further back, unread. Read across 1 page of Telegram’s post history, 20 posts per page.

ERR · 30 days
91.7%
avg views ÷ 8 subscribers
Avg views / post
7.3
18 posts measured
Reaction rate
—
this channel exposes no reaction counts
Posts in window
20
of 20 held

ERR is average views per post over the last 30 days divided by subscribers, the definition TGStat uses, so this figure is comparable with the one you will see elsewhere. It falls structurally as a channel grows: a high ERR on a small channel and a low one on a large channel describe reach mathematics, not quality. We publish the figure and the sample it came from and pass no verdict on it.

ER is defined industry-wide as (forwards + reactions + comments) ÷ views — note the denominator is views, not subscribers. Telegram’s public web preview carries views and reactions but not forward or comment counts, so the reaction rate above is the reactions term only and is therefore a floor: the true ER for this channel is higher by an amount we have not measured and will not estimate.

What these figures were computed from
WindowRolling 30 days · latest post in window 13 November 2026
Posts held20 (9 November 2026 – 13 November 2026)
Views total132
Reactions total—
Forwards / commentsnot exposed by the public surface — not measured, not estimated
Readings taken13 Nov 2026, 04:22 UTC

Views are a single reading per post, taken at the time above. A post published in the last day or two is still accumulating views, which pulls the 30-day average down slightly. That is a property of the standard definition rather than a fault in it, so we keep the definition rather than “correcting” the number into something nobody can reproduce.

Precision. Telegram publishes view counts on its public widget in short form — 8.12K, 3.7M — so any reading at or above 1,000 reaches us rounded to three significant figures, and only counts below 1,000 are exact. Averages and rates derived from them are shown to the same precision rather than to the unit: a figure like 3,701,250 would assert digits nobody measured.

Reaction counts are published per emoji and rounded the same way, so a total below 1,000 is exact and a larger one is a sum that may carry a rounded component from each emoji above 1,000. Because it is a sum, it does not look rounded — read a large reaction total as three significant figures per contributing emoji rather than as the figure it prints.

What this channel posts

Links
10

Lifetime counters from Telegram’s own channel header, read 13 November 2026 — not the date at the top of this page, which is when the subscriber count was last read. Below Telegram’s rounding threshold, so these counts are exact.

Recent posts

Telegram’s own ‘edited’ marker: 0 of 20 posts held from the last 30 days (0%) carried it when this register last read them. The marker says a post was changed after publishing, not what it said before — this register holds the earlier wording only where an edit fell between two of its own reads (see the edit ledger and how it is measured).

13 Nov 2026, 03:39 UTC2 viewsread 13 November 2026

⚡ 开源权重有了「离线副本网」:私有 BT 存档,站点没了也能重建 • LibreAGI 是申请制私有 BitTorrent tracker,会员各存一份「最可能消失」的开放权重;只收 Apache-2.0 / MIT / Qwen Community License,逐文件对照源 commit 校验。 • 目录签名公开,列出每个模型的 hash 与来源,站点消失也能按目录重建 swarm;做种收益按稀缺度结算,单副本模型拿基准 4 倍。 • 模型下架、许可收紧已不是假设,这是「权重留在自己盘上」的一条退路。 🔗 libreagi.com(HN 135 分) 推荐:晨报 · 福利社 · 精选 · 讨论群 · 群管理助手

13 Nov 2026, 02:34 UTC2 viewsread 13 November 2026

⚡ NVIDIA Linux 驱动漏洞开始解禁:首批 2 个公开,另有 25 个排队待发 • 已公布:NvRoot 与 ForgePtr。保密期于 2026-11-12 到期,属 NVIDIA 内核图形驱动侧的安全缺陷。 • 为什么值得看:Phoronix 特别点名这两个缺陷对「在云端租用 GPU 切片」的用户风险最集中——多租户隔离一旦被穿透,同一张卡上其他租户的内存就在射程内。更关键的是这只是 25 个里的前 2 个。 • 该做什么:跑 NVIDIA 驱动的机器,尤其是容器化、虚拟化多租户场景,盯紧发行版内核与驱动更新,别等 25 个全出完再动。 🔗 Phoronix 原文 推荐:晨报 · 福利社 · 精选 · 讨论群 · 群管理助手

12 Nov 2026, 23:24 UTC3 viewsread 13 November 2026

⚡ #Knto晨报 | AI & 全球硬科技速览 (2026-11-13) 1️⃣ Gemini 4 Hydrogen 上线:日常主力档,智能指数反超 Argon Google DeepMind 发布 Gemini 4 家族快速执行模型,新增 xhigh 思考档。Artificial Analysis 智能指数 high 档 51 vs Argon 50、xhigh 档 54;Terminal-Bench 4.1 57%(3.8 Flash 18.2%),比 3.8 Flash 快 2.5 倍,HLE-Diamond 58.8%。定价 $1.50/$7.50 每百万 token、100 万上下文统一计价,app、Search、AI Studio、Antigravity 全开放;Antigravity 加 Orchestrator,同会话内可切 Carbon 与 Hydrogen。 ━━━━━━━━━━━━━━━━━━━━ …

12 Nov 2026, 11:47 UTC6 viewsread 13 November 2026

⚡ #Knto晨报 | AI & 全球硬科技速览 (2026-11-12) 1️⃣ Google 开源 TPU Sync:KV 缓存搬运从数秒降到 230ms Google Cloud 以 Apache 2.0 开源 C++ 库 TPU Sync,在 TPU HBM、host DRAM、共享存储与数据中心网络之间搬 KV cache 块和模型权重,服务长上下文、PD 分离推理与 RL 权重同步。官方数据:Qwen3 235B 分离部署下 KV 传输中位延迟由数秒降至约 230ms,低并发平均 TTFT 降 60%(235B)/90%(480B);前缀缓存 64K 上下文 TTFT 最多降 90%。提供 JAX/PyTorch wheel 与 KV cache manager、KV store、weight synchronizer 三套 API,可作 vLLM-TorchTPU 的 KV connector、SGLang-To…

12 Nov 2026, 08:20 UTC5 viewsread 13 November 2026

⚡ #Knto晨报 | AI & 全球硬科技速览 (2026-11-12) 1️⃣ Runway 开源 27B 机器人模型 Praxis-1 27B 参数 MoE 扩散 Transformer,把视频与动作放在一起预测,靠大规模视频预训练把「物体怎么动、任务怎么推进」学进权重,再迁移到真机。NVIDIA RoboLab-120 榜 1,200 局成功率 46.2%,高于 FLUX 3 Action 42.9%、Cosmos3 Nano 36.8%、π0.5 28.0%,官方称其为当前最强开源 VLA。权重、推理与微调代码(全量 + LoRA)以 Apache 2.0 放在 GitHub 与 Hugging Face,原生支持 LeRobot 与 vLLM-Omni,可商用;开箱覆盖 H100/B200、RTX 5090、RTX PRO 6000 到 Jetson Thor。在从未进过训练集的 Google RT-1 上微调后,…

12 Nov 2026, 06:58 UTC4 viewsread 13 November 2026

⚡ #Knto晨报 | AI & 全球硬科技速览 (2026-11-12) 1️⃣ 台积电董事会批 10 亿美元,分别入股 OpenAI 与 Anthropic 5 亿美元认购 OpenAI 面向大额投资者的优先股,另 5 亿美元按 Anthropic 预期 IPO 价买入普通股。台积电是 AI 芯片的实际制造方:Nvidia 数据中心芯片与 Apple 自研芯片绝大多数出自其晶圆厂,7–9 月单季利润超 230 亿美元、连续六季创纪录。对照 Nvidia 已向 OpenAI 投 300 亿、向 Anthropic 投 100 亿,这 10 亿让「供应商回投客户」的循环融资再添一环——分析师警告这类结构让真实的 AI 硬件需求更难评估。 ━━━━━━━━━━━━━━━━━━━━ 2️⃣ SpaceX「Kessel」机架:一块 84 层 PCB 干掉 NVLink 线缆 沿用 Vera Rubin NVL72 的 72 颗 …

12 Nov 2026, 03:07 UTC4 viewsread 13 November 2026

⚡ #Knto晨报 | AI & 全球硬科技速览 (2026-11-12) 1️⃣ Jane Street 唯一纽约机房,差点被一次误报烟感「淹掉」 承包商在机房隔壁切割扬尘触发烟感,而烟感联动会直接关停 CRAC 制冷,天花板洒水喷头随即加压待喷——只等室温够热就开闸。机房在 1 New York Plaza 33 层、约 10 排 × 10 柜,当年刻意贴着交易楼层,就是为了「跑过走廊拔电源」。系统组成员从楼下冲上 33 层,逐排逐柜拔线,把温度压在 143.6°F 停住(喷发线 150°F),次日正常开盘。事后他用 OCaml 写了每分钟轮询每台服务器温度取均值的告警:单机高温是常态,全机架同涨才是制冷失效。HN 427 分 / 220 评论。 ━━━━━━━━━━━━━━━━━━━━ 2️⃣ CMP 170HX 挖矿卡被电压毛刺「刷」成 A100-80G 研究者公开 GA100 裸片的电压故障注入漏洞,可任意烧写 …

11 Nov 2026, 23:18 UTC5 viewsread 13 November 2026

⚡ #Knto晨报 | AI & 全球硬科技速览 (2026-11-12) 1️⃣ AWS 发布本地模拟器,coding agent 不必再等部署 官方博客 11-11 宣布 AWS Emulator:一个跑在本机的 AWS API 模拟器,AWS CLI、各语言 SDK 与 coding agent 通过本地 HTTPS endpoint 调用与云端同一套 API,只要把环境变量指过去,业务代码一行都不用改。用 Rust 实现是为了在笔记本上压低 CPU 与内存占用;它按 AWS API models 仓库的服务模型生成,因此支持所有已有模型的服务,请求与报错形状跟云端一致。启动即自动生成凭据、配置与证书,写入 ~/.aws/emulator/9297/;Linux 装 RPM/Debian 包,macOS 与 Windows 走 Docker。典型闭环:建表、写入、读回这种单测不必再真部署一个 Region,Kiro、Cl…

11 Nov 2026, 15:38 UTC9 viewsread 13 November 2026

⚡ #Knto晨报 | AI & 全球硬科技速览 (2026-11-11) 1️⃣ Google DeepMind 发布 AlphaGenome 2,变异识别数量近乎翻倍 基于 68 类组织、超 1.07 亿个单细胞图谱训练,官方称相比前代能识别出近两倍的可改变细胞内基因表达的变异。架构从「并行预测多组织」改为条件式模型:单次只对一种细胞类型或状态做预测,并把生物条件注入每一层,同时新增 DNA 甲基化预测能力。已通过 AlphaGenome API 向非商业研究开放,商业研究走私密预览。 ━━━━━━━━━━━━━━━━━━━━ 2️⃣ 把 SSH 私钥放进 macOS Secure Enclave,每次签名按 Touch ID 系统自带 sc_auth 就能做,不必装第三方 App:sc_auth create-ctk-identity -l "ssh-MacBook" -k p-256-ne -t bio。私钥在芯片…

11 Nov 2026, 14:28 UTC8 viewsread 13 November 2026

⚡ #Knto晨报 | AI & 全球硬科技速览 (2026-11-11) 1️⃣ 48 个 AI Agent 读完 1,200 km² 激光雷达,找出未登记的玛雅遗址 作者把公开 LiDAR 条带切成图,48 个 agent 逐张筛金字塔、台地、广场,无模型无分类器,纯按指南看图;72 小时内从念头到墨西哥 INAH 正式回函「该坐标无登记遗址」。图上金字塔高 8–12 m,另有广场与疑似蓄水凹地。原文 ━━━━━━━━━━━━━━━━━━━━ 2️⃣ .NET 11 正式 GA STS 支持两年;C# 15 增 union types 与 closed hierarchies(switch 穷尽检查);Native AOT CLI host 加常驻 MSBuild server 压掉每次命令开销;ASP.NET Core 补异步验证与 OpenAPI 3.2,MAUI 默认换 CoreCLR。发布说明

Showing the 12 most recent of 20 posts we hold for @knto_tech. View and reaction counts are the latest single reading for each post, not a live figure, and a recent post is still accumulating both. A view count marked ≈ was rounded by Telegram before we ever saw it — t.me prints views in full below 1,000 and to three significant figures above, so ≈1,200,000 means somewhere between 1,150,000 and 1,249,999. Unmarked counts are exact. Text is reproduced from the public post preview and truncated for length.

Mentions

Named by 2 registered channels — every channel on the register whose own posts have named this one, by its current username or any other username it currently holds, merged from two separately captured readings of the same fact so a namer caught by only one of them is not missed and a namer both caught is not counted twice. A username this channel has since dropped is not matched — that handle may belong to someone else now, and crediting today’s namer to yesterday’s owner would misattribute it.

A mention is a weaker signal than a forward and is counted separately for that reason — naming a channel is not republishing it, and a handle in a post body is easy to place deliberately. The post counts beside each row below are distinct posts in which the handle appeared, from posts we have read on both sides — the “Named by N registered channels” figure above is a different count, of distinct NAMING CHANNELS rather than posts, and is not the sum of the rows under it.

Cite this entry

A live page changes as we take new readings, so a citation should name the measurement it is based on, not just the URL. The line below cites the subscriber count as measured 13 November 2026 — this entry's latest reading, not the date you are reading this.

“Knto 科技晨报 | AI大模型·开源爆款·极客生产力速览” (@knto_tech), 8 subscribers as measured 13 November 2026. Telegram Register, tgregister.com/channel/knto_tech.

Full measurement history, CC BY 4.0. Every reading this register holds for this entry, not just the latest one, as a dated, downloadable record: CSV · JSON. Free to use with attribution to tgregister.com. Each file carries its own generation timestamp, which is the figure to cite for exactly when the data was retrieved.