AI Educator | Helping You Make Money with AI, Tech Tools & Digital Skills📩 DM or Email for Collaborations:💬 [email protected]
Joined April 2026
- Tweets3.2K
- Following43
- Followers6.6K
- Likes2.2K
AI With Ridi retweeted
Having sex at 19+
Him: Your parents are home.
Her: Then be quiet.
Having sex at 39:
AI With Ridi retweeted
The extra year is a nice opportunity to keep developing your creative workflow. Flova 1.5’s launch offer lets you get one year free with your annual Pro purchase.
Seedance 2.5 720p starts at $0.025/sec.
flova.ai/?refCode=3BFU8P57
AI With Ridi retweeted
AI news production just got a lot more interesting! 🚀 StepFun Step 5 Preview, a 1M context window, and native text, image, and video input support. Turning trending AI news into a 2-minute broadcast is a great way to test its potential. Who's opening today's ZenMux blind box?
🤝 Paid partnership
ZenMux 万圣节盲盒开出来了：StepFun Step 5 Preview。
每天 PAYG 现金消费满 $1 就能开一次。我开到阶跃这颗旗舰预览版，原生可以输入文本 / 图 / 视频，输出文本，上下文 1M，就拿它做了一条 AI 新闻播报搞，文字稿是用Muse抓的AI HOT
成片是竖屏AI 快讯：10 月 28 日，缪斯播报，大约 2 分钟，五条新闻，Amazon 在 Anthropic IPO 里那笔约 180 亿敞口、Google Genie 4、Claude Code 带偏分析师、Gemini 接 SynthID、家长起诉 OpenAI那起未证实指控。整条片子直接配在下面。
玩法很简单：日常用 AI → 满 $1 开盲盒
16 款模型，每天 3000 份，单份最高估值 $500（是额度价值，不是现金）。活动到 10/31 22:00（新加坡时间）。
去开今天的盲盒：
zenmux.ai/halloween?invite_c…
AI With Ridi retweeted
The most realistic AI videos often come from ordinary moments, not flashy effects. 🎥
Breaking the workflow into prompt design, reference images, and video generation makes it easier to maintain character consistency and recreate that nostalgic early-2000s DV look.
A practical workflow for anyone experimenting with AI video creation!
🤝 Paid partnership
你敢信，这种看起来像 20 多年前家用 DV 拍出来的视频，其实全是 AI 生成的！😂今天分享一下具体是怎么做的。
最近一直在折腾这种写实感 AI 视频，发现越是普通的生活场景，反而越考验细节。人物一致性、年代氛围、镜头运动，稍不注意就容易穿帮。
具体的 AIGC 创作工作流，基本分成 3 步：
① 创意策划 / 提示词设计：Codex / Claude Opus 5.5
先把创意拆成具体的场景、人物动作和镜头语言，再补充年代背景、光线、道具等细节。
比如这条视频，重点就是还原 2000 年代的生活感：老式居民楼、白炽灯、旧风扇、晾晒的衣服，以及家用 DV 特有的噪点、色偏和轻微过曝。
② 角色与分镜参考图：GPT Image 2.5
先生成角色和场景参考图，锁定五官、发型、服装和房间布局，再根据剧情生成关键分镜。
这里有个经验：复杂场景尽量先确定空间结构和道具位置，后面的视频生成会更容易控制。
③ 视频生成：Seedance / Kling
根据参考图和分镜生成动态画面，重点检查人物有没有变脸、动作是否自然、道具位置有没有变化。
尤其是这种写实 DV 视频，镜头不需要太复杂。固定机位、自然光、生活化的小动作，加上适当的噪点和色偏，反而更容易做出真实随拍的感觉。
这套流程我现在经常用，虽然偶尔也会翻车，但把创意、参考图和视频生成分开处理，确实省了不少反复试错的时间。
顺便分享一个最近发现的 AI 工具订阅渠道：@GamsGo_Global。
这次重点体验的是 Codex Plus 独享套餐，页面价格 $16.49/月,大约比官方便宜好几美元。除了 Codex，平台还有 Claude、ChatGPT、Gemini 等工具的不同套餐，做 AIGC 的朋友可以按需求看看。
具体使用入口看评论区👇
AI With Ridi retweeted
Your next creative idea could be the start of something bigger. Flova 1.5 gives you an extra year of Pro to continue exploring your projects.
Seedance 2.5 720p starts at $0.025/sec.
flova.ai/?refCode=3BFU8P57
AI With Ridi retweeted
One API for code, pixel art, and cinematic animations? 👀🔥 Solo game dev is getting seriously interesting. The workflow matters more than the hype!
🤝 Paid partnership
这两天外网最大的瓜：微软 Xbox 的一位高管在 X 上高调放话，说自己用 AI 只花 3 小时就「Vibe Coding」出了一个可玩的游戏，还嘲讽那些不用 AI 的游戏开发者早晚被淘汰——结果喜提全网围攻，不到 24 小时直接注销账号跑路了。
吃瓜归吃瓜，作为天天写代码的开发者，我反而产生了强烈的好奇心：
抛开 3A 大作不谈，现在一个人如果同时调动代码模型 + 生图模型 + 视频模型，到底能不能在极短时间内搓出一个带完整画面、带音效逻辑、甚至带开场 CG 动画的万圣节小游戏？
昨晚我计时实测了一把，发现以前卡住独立开发者的根本不是写代码，而是「割裂的工具链」——写逻辑去调 Claude/GPT，画精灵图去开生图网站，做过场动画又要去切 Kling，光倒腾素材就耗掉一半时间。
这次我换了个思路，全程只接了 ZenMux —— 一站式 AI 模型平台：一个账户、一套 API，直接接入多家厂商的文本、图片和视频模型。One API. Whole workflow.
实测跑通的「一人游戏流水线」非常丝滑：
1. 核心玩法与碰撞逻辑：直接调 Claude / GPT 生成单文件 Canvas 游戏主循环；
2. 美术资产生成：在同一个脚本里改一行 model 参数，调用图片模型批量生成暗黑像素风的南瓜怪与地图贴图；
3. 开场 CG 动画：紧接着把首帧画面传给 Kling 视频模型，自动渲染出一段 5 秒的万圣节开场动画。
🎃 正好这几天是 ZenMux 的万圣节活动「TRICK OR TOKENS? 🎃」，我边跑这套游戏脚本边拆了今天的盲盒：
给也在调 API 的朋友同步下活动规则（门槛极低）：
1️⃣ 参与步骤：每个活动日用现金 PAYG 消费满 $1 → 解锁一个模型盲盒 → 点击开盒（注：订阅用量、赠送额度和优惠券抵扣不算进 $1）。
2️⃣ 奖池阵容：共 16 款模型（涵盖 Claude、GPT、Gemini、DeepSeek、GLM、MiniMax、Qwen、Kimi、Doubao、StepFun、Kling 等），每天 3,000 份奖励，单份模型奖励最高估算使用价值 $500，大约 44% 的概率开到 Claude / GPT / Gemini！
3️⃣ 活动时间：新加坡时间 10/26 22:00 – 10/31 22:00（UTC 10/26 14:00 – 10/31 14:00），每晚 22:00（新加坡时间）刷新；每个账户每天最多开 1 次，全程最多 5 次。
🎁 今日开盒揭晓 & 游戏生成实测：
我当场把刚抽到的额度用在了这个万圣节小游戏的上（见下图实测效果）：
不到半小时，从代码逻辑到视觉动画全在同一个 API Key 下跑通。高管吹牛虽然翻车了，但一套 API 串起全模态工作流确实真香。
💡 顺手提醒：开完盒可以点页面上的「Invite & Earn $5」，好友首次充值得 5% Bonus，邀请人在好友首充 24 小时后得 $5 Credits。
👉 活动页：zenmux.ai/halloween
前往 ZenMux，查看并解锁今天的盲盒。
AI With Ridi retweeted
One template, endless creative possibilities. The red-and-black visual system makes every scene feel connected while keeping its own story. 🎨
🤝 Paid partnership
VISUAL OS 100
Two inks. One shared action. Everything else stays red.
BOOK BRIDGE — Two readers unfold one book long enough to reach each other.
CLAY DAY — Two pairs of hands pinch one giant vase into shape.
PUPPET CLUB — The puppets are bigger than the people working them.
RAIN PARADE — Three neighbors, one umbrella, room for one more.
One template below. Swap the scene, keep the system.
If you want to run them this week, ZenMux has a Halloween event, Trick or Tokens?
Spend $1 a day on pay-as-you-go from your cash balance and you unlock one mystery box.
16 models, 3,000 boxes a day. Rewards include image credits like 500 Kling V3 images. One reward can be worth up to $500 in model usage. That's usage, not cash, and not every box.
Go to ZenMux and unlock today's box → zenmux.ai/halloween
Resets daily at 14:00 UTC. Ends Oct 31.
#ZenMux #Halloween2026
Template below ↓
AI With Ridi retweeted
Getting married at 20:
Him: We’re soulmates
Her: Nothing can break us
Getting married at 35:
AI With Ridi retweeted
8 Foods That Turn Average Men Into Sexual Beasts.
(No school will teach you this)
1. Garlic
AI With Ridi retweeted
More time to experiment can help you discover your own creative style. Flova 1.5 offers a second year of Pro free during its launch promotion.
Seedance 2.5 720p starts at $0.025/sec.
flova.ai/?refCode=3BFU8P57
🤝 Paid partnership
AI With Ridi retweeted
Having sex as a wife in the 90s:
Wife: The kids are asleep. Come to bed, darling.
Husband: Give me two minutes. Let me finish watching the match.
Having sex as a wife in 2026:
AI With Ridi retweeted
The launch had the numbers. the BTS has the personality. you need both to understand what @tavus built
AI With Ridi retweeted
A subscription bonus that leaves the choice up to you. Flova 1.5 offers a free second year of Pro, which you can keep for future projects or gift.
Seedance 2.5 720p from $0.025/sec.
flova.ai/?refCode=3BFU8P57
AI With Ridi retweeted
OCR is only useful when you can trust what it extracts.
MiniCPM-OCR is tackling that challenge with a 1B model for document parsing and a separate model to evaluate the results.
From tables and formulas to complex documents, this open-source suite is built to make document processing faster and more reliable.
Your OCR says it’s done—but did it get everything right? 🤔 In high-stakes workflows, even a small error can be costly.
🚀 Meet the MiniCPM-OCR suite: an open-source solution to parse fast and guard the results.
Two new models:
🔹 MiniCPM-OCR 1.0 — SOTA OCR performance and speed at 1B scale. Compact, accurate, and fast.
🔹 MiniCPM-OCR Jev — A 2B decision model that helps assess parsing accuracy, with performance approaching that of Qwen3.5-397B-A17B on JevOCRBench.
⚡ What’s new?
Small model, SOTA results: MiniCPM-OCR 1.0 scores 93.68 on Real5-OmniDocBench and 54.7 on OCRBench v2.
A sharp goalkeeper 🛡️: MiniCPM-OCR Jev is model-agnostic, checking text, formulas, and tables from different OCR parsers. It achieves 93.6 on JevOCRBench, outperforming JEV-35B-A3B (90.1) and approaching Qwen3.5-397B-A17B (93.8).
Up to 128 pages per minute 🤯: With DSpark, MiniCPM-OCR runs at 1.49× the speed of PaddleOCR-VL-1.6 and 3.62× that of MinerU2.5 Pro. The SDK’s concurrency optimizations further boost throughput to as much as 128 pages per minute.
Deploy where you work: Day-0 support for Arm, Houmo AI, Intel, and NVIDIA, plus 10 more chip vendors through FlagOS.
Ready for your stack: vLLM, SGLang, llama.cpp, and Ollama, with quantized models, fine-tuning support, and Agent Skills.
🔗 GitHub: github.com/OpenBMB/MiniCPM-O…
🤗 Hugging Face:
MiniCPM-OCR 1.0：huggingface.co/openbmb/MiniC…
MiniCPM-OCR Jev：huggingface.co/openbmb/MiniC…
👉 Try the Demo：platform.modelbest.ai/consol…
Use the API：platform.modelbest.ai/consol…
