Helping founders & CEOs build powerful online brands → Branding | LinkedIn | X Growth | AI 📩 kaesa2256@gmail.com
Auckland, New Zealand
- Tweets 2,846
- Following 42
- Followers 325
- Likes 2,625
Pinned Tweet
Kids animation is the most underrated money niche on YouTube 💸
No face.
No voice needed.
No trends.
Just stories + consistency.
One channel:
721K subs
24 videos
This is not luck.
It’s niche selection.
Want the blueprint?
Comment “KIDS” 👇
Follow me🙏
1
86
Arosh Khan retweeted
🚨 BREAKING: My wife used Claude as a digital planner in 2026.
6 months later, she had actually hit every target she set.
Use these 6 prompts:
28
77
144
15,350
Arosh Khan retweeted
🚨 GOOGLE JUST GAVE NOTEBOOKLM A MAJOR UPGRADE.
And it’s way more than a simple rebrand.
Gemini Notebook is being integrated deeper into Google’s ecosystem, bringing a secure cloud computer and more powerful AI workflows.
Yet most people will still use it as nothing more than a PDF summarizer.
Here are 8 powerful prompts to turn it into a serious AI research assistant. 🧵
31
92
292
17,958
Arosh Khan retweeted
🚨 Now Claude Code can read complete documentation without spending a single token.
You just need to connect Claude Code to Google’s NotebookLM via MCP.
Here’s how to do it 👇
18
38
90
5,842
Arosh Khan retweeted
I got early access to Tavus Griffin and tested the part of AI interaction that usually feels missing: reacting in real time.
Instead of simply asking a question and waiting for an answer, I could talk, gesture, interrupt and change what was happening during the conversation.
Griffin continuously sees, hears and responds.
That's what makes this worth watching.
@tavus @hassaanrza
tavus.io/griffin
Introducing Griffin, the first model to pass the video Turing test.
48% of people who talked to it live thought it was a real human. Previous systems have had a pass rate <3%. It is #1 on NVIDIA's benchmark for full-duplex AI video.
It’s the first Human Interaction Model (HIM).
Community note
The 48% figure and "video Turing test" claim are from Tavus's own study of 54 one-minute calls, not independently verified or using a standard protocol. Griffin-Lite leads NVIDIA's VideoFDB benchmark on their public leaderboard.
cellcog.ai/blog/tavus-gri…
research.nvidia.com/labs/amri/proj…
tech-ish.com/2026/10/02/tav…
21
22
30
4,577
Arosh Khan retweeted
Tavus gave me early access to Griffin, the first model to pass the Turing test, and I wanted to see what that actually felt like in practice.
Tavus reports that 45% of 120 participants thought Vanessa was a real person during its Turing test.
After testing Griffin Lite myself, the continuous face-to-face interaction was the part that stood out most.
@tavus @hassaanrza
tavus.io/griffin
Introducing Griffin, the first model to pass the video Turing test.
48% of people who talked to it live thought it was a real human. Previous systems have had a pass rate <3%. It is #1 on NVIDIA's benchmark for full-duplex AI video.
It’s the first Human Interaction Model (HIM).
Community note
The 48% figure and "video Turing test" claim are from Tavus's own study of 54 one-minute calls, not independently verified or using a standard protocol. Griffin-Lite leads NVIDIA's VideoFDB benchmark on their public leaderboard.
cellcog.ai/blog/tavus-gri…
research.nvidia.com/labs/amri/proj…
tech-ish.com/2026/10/02/tav…
21
21
30
3,636
Arosh Khan retweeted
🔥 HOOK
I stopped spending hours making presentations. 🤯
Gemini + PowerPoint can turn a simple idea into a professional presentation in minutes.
Here are 10 prompts you can steal: 👇
39
52
73
2,223
Arosh Khan retweeted
StepAudio 3 Music packs four creation modes into one release: full songs, instrumentals, covers, and vocal-to-song arrangement.
Not a small first step into music for StepFun.
StepFun and ACE Studio present StepAudio 3 Music, StepFun’s first music generation foundation model — built to turn a prompt and lyrics into a complete song.
Describe the sound you want:
• Genre, mood and vocal character
• Instruments, key and BPM
• Song structure and arrangement
Four workflows in one model:
🎤 Song generation
🎹 Instrumental generation
🔁 Music cover
🎙️ Vocal-to-song arrangement
Powered by ABC-COT, it plans musical structure and arrangement before synthesis. Generate a version, rewrite the prompt, edit the ABC notation, and iterate toward the sound in your head.
(ABC-COT API coming soon)
Try the interactive demo:
› static.stepfun.com/blog/step…
11
11
13
731
Arosh Khan retweeted
ChatGPT can now turn one selfie into a full professional studio style photoshoot.
Here are 7 prompts to try 👇
(Save now, Edit later)
30
70
129
6,094
Arosh Khan retweeted
🚨 BREAKING: Claude has a feature called Resume Mode.
It pulls your CV out of the pile and into an interview, like a recruiter rewrote it for you.
Here are 6 prompts to start:
📌 Save this
32
131
648
55,408
Arosh Khan retweeted
Your iPhone has a hidden setting that can make your Wi-Fi work better.
And most people never touch it.
Here are 7 Wi-Fi settings you should check tonight:👇
26
45
55
1,976
Arosh Khan retweeted
🚨 BREAKING: Claude can now BUILD a YouTube channel from scratch—and hit monetization in just 90 days. 100% free.
Here are 8 prompts to make it happen:
34
37
102
8,267
Arosh Khan retweeted
📂 Drive Self-Study E-Book:
drive.google.com/file/d/1xl7……
📂 Printed Book Version: s.shopee.co.id/2Vqoobxy9r
📂 Drive Vocab:
drive.google.com/drive/folde……
📂 Drive Grammar, TOEFL, IELTS, & TOEIC:
drive.google.com/drive/folde……
Follow : @nesu_ai_x
15
107
266
9,615
Arosh Khan retweeted
The interesting number here isn’t 309B. It’s how little of it needs to be active at once.
Introducing Naive-N0.5-Flash: Building Frontier AI with AI
🔹 309B MoE, 15.5B active: top-tier in coding, leading in AI R&D.
🔹 Native 1M context, no full-attention layers (hybrid SWA + DSA).
🔹 Inference runtime built by AI: up to 2,000 tok/s in Ultrafast mode.
Weights are open today under MIT license.
🔗 Tech blog: naive.ai/en/research/
🤗 Hugging Face: huggingface.co/NaiveAI/Naive…
💻 GitHub: github.com/NaiveAI-Labs/Naiv…
🌐 naive.ai/en/
5
11
12
974
Arosh Khan retweeted
🚨 Stop spending hours creating classroom materials from scratch.
Claude can build worksheets, quizzes, reading tasks, discussion questions, graphic organizers, and revision resources in minutes.
Here are Claude prompts every teacher should save 👇
(Save this for your next lesson-planning session)
38
86
125
3,848
Arosh Khan retweeted
309B sounds massive, but ~16B active is the number I’d pay closer attention to.
Introducing Naive-N0.5-Flash: Building Frontier AI with AI
🔹 309B MoE, 15.5B active: top-tier in coding, leading in AI R&D.
🔹 Native 1M context, no full-attention layers (hybrid SWA + DSA).
🔹 Inference runtime built by AI: up to 2,000 tok/s in Ultrafast mode.
Weights are open today under MIT license.
🔗 Tech blog: naive.ai/en/research/
🤗 Hugging Face: huggingface.co/NaiveAI/Naive…
💻 GitHub: github.com/NaiveAI-Labs/Naiv…
🌐 naive.ai/en/
5
12
14
1,271
Arosh Khan retweeted
🚨 Powerpoint is finished!
ChatGPT can now build full professional presentations in seconds.
Here are the prompts to use:
35
42
72
6,813
Arosh Khan retweeted
🚨𝗕𝗥𝗘𝗔𝗞𝗜𝗡𝗚: GOODBYE CHATGPT forever.
Claude just collapsed every tool you were using into one single 120-second system completely free.
7 prompts go from completely across 10 tools to completely unstoppable with just one: 👇
37
63
104
3,662
Arosh Khan retweeted
A 309B model with ~16B active parameters says a lot about where model efficiency is heading
Introducing Naive-N0.5-Flash: Building Frontier AI with AI
🔹 309B MoE, 15.5B active: top-tier in coding, leading in AI R&D.
🔹 Native 1M context, no full-attention layers (hybrid SWA + DSA).
🔹 Inference runtime built by AI: up to 2,000 tok/s in Ultrafast mode.
Weights are open today under MIT license.
🔗 Tech blog: naive.ai/en/research/
🤗 Hugging Face: huggingface.co/NaiveAI/Naive…
💻 GitHub: github.com/NaiveAI-Labs/Naiv…
🌐 naive.ai/en/
5
10
11
1,001
Arosh Khan retweeted
The next billion voice AI users are not all speaking American English into studio microphones. Datasets need to reflect that reality.
Introducing Monsoon ASR ⚡️ @voicearena_ai
Speech recognition does not have a model problem anymore. It has a data problem.
The best ASR systems are approaching human-level performance in English. But move into the long tail of the world's languages, especially real, conversational speech and error rates can still be 5-10X higher.
Today, we're releasing Monsoon ASR: a new generation of training data built specifically to close that gap.
50 languages. 100,000+ hours. Dense spontaneous speech.
And one goal: Single-digit WER across the world's languages.🧵
10
11
13
5,447