
Sharing insights on AI, No Code, Tech Tools & Prompts | AI & Tech Content Creator | 350K+ LinkedIn & Instagram Community | ✉️ [email protected]
Joined November 2023
- Tweets56.4K
- Following980
- Followers53.1K
- Likes58K
Pinned Tweet
I’m always interested in seeing how far AI can go beyond generating code.
IQuest-Q1 is a code-focused large model built for developers, with performance evaluated across different task scenarios.
From building interactive game demos and generating 3D scenes to debugging code and handling complex tasks with multimodal information and tools, it explores some interesting possibilities for AI in development.
Excited to see how models like this evolve!
Explore IQuest-Q1: iquestlab.github.io/
#coding #LLM
AI agents are great, but the real value is when they save hours of actual work.
Templafy Document Agents bring AI into PowerPoint, creating on-brand presentations in 8 minutes instead of 2 hours.
This means less time fixing slides and more time focusing on the work that actually matters.
🤝 Paid partnership
AI that knows your company’s content and branding? 👀
With Templafy Team, build and share AI Agents that create editable PowerPoints using your company's knowledge and templates.
Get started 👉 templafy.com/templafy-teams-…
Start your free 14-day trial templafy.com/templafy-teams-…
Manish Kumar Shah retweeted
AI agents are great, but the real value is when they save hours of actual work.
Templafy Document Agents bring AI into PowerPoint, creating on-brand presentations in 8 minutes instead of 2 hours.
This means less time fixing slides and more time focusing on the work that actually matters.
🤝 Paid partnership
AI that knows your company’s content and branding? 👀
With Templafy Team, build and share AI Agents that create editable PowerPoints using your company's knowledge and templates.
Get started 👉 templafy.com/templafy-teams-…
Manish Kumar Shah retweeted
When I look at an AI tool promotion, I want to know what’s included—not just how low the price is. Dreamina Basic is $0.29 for the first month, so I checked the credits that come with the plan.
🤝 Paid partnership · 🤖 Made with AI
Manish Kumar Shah retweeted
Remember the mysterious Space Bunny model everyone was curious about?
We finally know what's behind it: MiniMax M3.1 Flash!
In this video, I compare its visual outputs with DeepSeek V4.1 Flash 1017, using examples of animated scenes and creative designs. Both deliver similar results in these examples, making it interesting to see how they approach the same prompts.
What I find most interesting is the balance between output quality and cost. Beyond creative tasks, M3.1 Flash is built for coding and agentic workflows, offering 70% lower overall cost for agentic tasks compared to its predecessor, with output speeds reaching 200+ tokens per second.
Manish Kumar Shah retweeted
1/ Ever had an AI agent suggest an approach your team had already rejected?
The problem isn't always the model's ability to write code. Sometimes, it lacks the context behind your team's previous decisions.
I wanted to explore how EverOS Cloud could help solve this by giving agents access to past project experience.
🤝 Paid partnership
Manish Kumar Shah retweeted
GPU kernel optimization is entering a new phase.
ForgeKernel shows how a powerful frontier model, paired with a goal-driven approach commonly used in agent frameworks, can outperform expert-written kernel implementations by 1.08×–1.58× across the reported cases. Anyone can define a simple goal, such as “beat this number,” while a correctness gate validates the results.
The interesting part? A powerful frontier model paired with a simple prompt may be enough to outperform specialized agents designed specifically for kernel optimization.
This could help developers optimize GPU kernels without unnecessary manual tuning. Looking forward to the open-source release later this week.
🤝 Paid partnership
Do we still need specialized agents for GPU kernel optimization?
We asked the question, and ran the experiment to answer it.
New paper with @THU_NLP: Agents under Goal mode are now able to beat systems deliberately built for kernel optimization. We named it JustGoal.
On FlashAttention, Gated DeltaNet, and DeepSeek Sparse Attention, a general frontier model (in our case, using Fable 5.1) delivers 1.08× to 1.58× speedup over expert-written references. Against the specialized KDA-0.5 (1.45× and 1.28×), Fable 5.1 hits 1.58× on both.
Frontier models are really burying the talent of kernel optimization experts in the past.
Paper: github.com/OpenBMB/JustGoal-…
GPU kernel optimization is entering a new phase.
ForgeKernel shows how a powerful frontier model, paired with a goal-driven approach commonly used in agent frameworks, can outperform expert-written kernel implementations by 1.08×–1.58× across the reported cases. Anyone can define a simple goal, such as “beat this number,” while a correctness gate validates the results.
The interesting part? A powerful frontier model paired with a simple prompt may be enough to outperform specialized agents designed specifically for kernel optimization.
This could help developers optimize GPU kernels without unnecessary manual tuning. Looking forward to the open-source release later this week.
🤝 Paid partnership
Do we still need specialized agents for GPU kernel optimization?
We asked the question, and ran the experiment to answer it.
New paper with @THU_NLP: Agents under Goal mode are now able to beat systems deliberately built for kernel optimization. We named it JustGoal.
On FlashAttention, Gated DeltaNet, and DeepSeek Sparse Attention, a general frontier model (in our case, using Fable 5.1) delivers 1.08× to 1.58× speedup over expert-written references. Against the specialized KDA-0.5 (1.45× and 1.28×), Fable 5.1 hits 1.58× on both.
Frontier models are really burying the talent of kernel optimization experts in the past.
Paper: github.com/OpenBMB/JustGoal-…
Manish Kumar Shah retweeted
Do we still need specialized agents for GPU kernel optimization?
We asked the question, and ran the experiment to answer it.
New paper with @THU_NLP: Agents under Goal mode are now able to beat systems deliberately built for kernel optimization. We named it JustGoal.
On FlashAttention, Gated DeltaNet, and DeepSeek Sparse Attention, a general frontier model (in our case, using Fable 5.1) delivers 1.08× to 1.58× speedup over expert-written references. Against the specialized KDA-0.5 (1.45× and 1.28×), Fable 5.1 hits 1.58× on both.
Frontier models are really burying the talent of kernel optimization experts in the past.
Paper: github.com/OpenBMB/JustGoal-…
Manish Kumar Shah retweeted
The CapCut plugin is coming soon to ChatGPT. Your favorite idol moments are about to get a new editing workflow. 🎤
At CapCut World, leading creators and OpenAI shared ideas on AI-powered video creation. Now, here's a sneak peek at using ChatGPT to control CapCut Web for video editing.
Imagine combining your idol's best stage expressions, dance breaks, and iconic close-ups into one fan edit. Describe the result you want, and ChatGPT can help carry out supported edits in CapCut.
Setup: Open CapCut Web → Video Editor → top-right plugin icon → copy the prompt → paste and send it in ChatGPT desktop on your PC to install the plugin.
Desktop early access only for now, with mobile and web support expected in mid-November.
#CapCutWeb #CapCutDidThat #ChatGPT #CapCutWorld
🤝 Paid partnership · 🤖 Made with AI
Manish Kumar Shah retweeted
We’ve raised a $2.5M pre-seed to fix hardware hiring.
The most reliable way to know if an engineer can do the job is to watch them work.
In hardware, that’s been almost impossible to do at scale.
Colare runs simulation-based assessments that show how candidates actually work, in the tools they’d use on the job, before anyone books an interview.
We assess across:
- Mechanical (CAD, DFM, FEA)
- Electrical (PCB, schematics)
- Systems
- Manufacturing, & more.
Candidates solve problems from first principles.
Hiring teams see that across their whole pool, so engineers only interview people who’ve already cleared the bar.
We're already working with everyone from seed-stage startups to multi-billion-dollar aerospace manufacturers.
If you’re hiring hardware engineers, let’s talk:
colare.com/
When I look at an AI tool promotion, I want to know what’s included—not just how low the price is. Dreamina Basic is $0.29 for the first month, so I checked the credits that come with the plan.
🤝 Paid partnership · 🤖 Made with AI
The plan includes 1,575 credits, but that doesn’t translate into a guaranteed number of finished videos. Generation settings affect credit use, and you may want to make several versions before choosing a final result. Checking real examples and Dreamina’s official credit rates can help you estimate what the plan offers.
#DreaminaPartner #DreaminaOnly29Cents
Seedance is one of the AI video models available on Dreamina’s creative platform. Both are part of ByteDance’s broader creative ecosystem.
dreamina.capcut.com/pricing/…
That puts the tools in context, but it doesn’t tell us whether each discounted subscription costs the company more than it earns. The actual cost structure hasn’t been publicly established.
#Dreamina #Seedance
The CapCut plugin is coming soon to ChatGPT. Your favorite idol moments are about to get a new editing workflow. 🎤
At CapCut World, leading creators and OpenAI shared ideas on AI-powered video creation. Now, here's a sneak peek at using ChatGPT to control CapCut Web for video editing.
Imagine combining your idol's best stage expressions, dance breaks, and iconic close-ups into one fan edit. Describe the result you want, and ChatGPT can help carry out supported edits in CapCut.
Setup: Open CapCut Web → Video Editor → top-right plugin icon → copy the prompt → paste and send it in ChatGPT desktop on your PC to install the plugin.
Desktop early access only for now, with mobile and web support expected in mid-November.
#CapCutWeb #CapCutDidThat #ChatGPT #CapCutWorld
🤝 Paid partnership · 🤖 Made with AI
Manish Kumar Shah retweeted
Good breakdown of the workarounds before getting to the fix. I've used all three.
🤝 Paid partnership
Manish Kumar Shah retweeted
Ever realized how much time goes into handling the small things while building a product?
File uploads, image optimization, storage + delivery are essential, but they can take time to build from scratch.
That’s where @Filestack helps.
It provides APIs to simplify file management, image processing, security + content delivery.
Here’s what you can do with it:
- Upload files from different sources
- Resize, crop + optimize images
- Deliver files through a global CDN
- Manage file security
- Use AI tools for OCR, image captioning + content analysis
Whether you're building a web app, SaaS product, or AI application, Filestack helps you focus more on building and less on file management.
Worth checking out if you're a developer 👇
filestack.com/?utm_source=Ma…
#WebDevelopment #AI #DeveloperTools #APIs
🤝 Paid partnership
Ever realized how much time goes into handling the small things while building a product?
File uploads, image optimization, storage + delivery are essential, but they can take time to build from scratch.
That’s where @Filestack helps.
It provides APIs to simplify file management, image processing, security + content delivery.
Here’s what you can do with it:
- Upload files from different sources
- Resize, crop + optimize images
- Deliver files through a global CDN
- Manage file security
- Use AI tools for OCR, image captioning + content analysis
Whether you're building a web app, SaaS product, or AI application, Filestack helps you focus more on building and less on file management.
Worth checking out if you're a developer 👇
filestack.com/?utm_source=Ma…
#WebDevelopment #AI #DeveloperTools #APIs
🤝 Paid partnership






