Meris Dabhi

Building production AI agents | Guardrails, reliability & real-world plumbing | Sharing what actually ships

[Meris Dabhi @Merisdabhi]

@AlexiGlad We need to buy any shares? 😅

[Meris Dabhi @Merisdabhi]

@bloggersarvesh If you are an editor, what price do you charge to create this video?

[Meris Dabhi @Merisdabhi]

@Bencera Dude, me too

[Meris Dabhi @Merisdabhi]

The next layer of AI-native content creation:

design.md → websites
frame.md → videos

[HeyGen @HeyGen]

Introducing frame.md, a spec built for videos & motion

design.md kept your brand consistent across screens

but when applied to videos, agents translated it back into webpages and decks

frame.md teaches your agents how to make branded video

turn your design.md into frame.md ↓

[Meris Dabhi @Merisdabhi]

@supabase Who writes SQL manually in this era 👀

[Meris Dabhi @Merisdabhi]

@teslaownersSV we love @elonmusk ❤️

[Meris Dabhi @Merisdabhi]

@candyyueliu Technology is moving so fast.

[Meris Dabhi @Merisdabhi]

@HumansNoContext old video

[Meris Dabhi @Merisdabhi]

@v0 Definitely, V0 is much powerful, but people use Claude or Codex.

[Meris Dabhi @Merisdabhi]

@NanouuSymeon Distribution and marketing is main game

[Meris Dabhi @Merisdabhi]

@federicodonaton 😂

[Meris Dabhi @Merisdabhi]

AI labs in 2025:

Claude: "I'm better at reasoning."

GPT/Codex: "I'm better at coding."

Gemini: "I'm better at multimodal."

MiniMax M3: "Yes." 😅

[MiniMax (official) @MiniMax_AI]

Introducing MiniMax M3: The First Open-Weights Model to Combine Three Frontier Capabilities

- Coding & Agentic Frontier: 59.0% SWE-Bench Pro, 66.0% Terminal Bench 2.1, 34.8% SWE-fficiency, 28.8% KernelBench Hard, 74.2% MCP Atlas
- MiniMax Sparse Attention scales context to 1M
- Natively Multimodal from Step Zero

API: https://t.co/fHRdSV7BwZ
Token Plan: https://t.co/BDCycxepZw
🚀New! MiniMax Code: https://t.co/GvB4YiB6Ul

Weights & Tech Report in ~10 Days

[Meris Dabhi @Merisdabhi]

Me → Claude → Claude → Claude → Claude

[ClaudeDevs @ClaudeDevs]

We've updated /fork in Claude Code

/fork now runs a background agent with your exact context (system prompt, tools, history, model) and prompt cache. The result gets returned to your session.

/branch (the old /fork) still copies the transcript to a new session you drive.

[Meris Dabhi @Merisdabhi]

@HeyGen @HyperFrames_ 2 weeks ago I used HyperFrames because the AI video generation model did not have text accuracy, but now I use Gemini omni and it's a very good model.

[Meris Dabhi @Merisdabhi]

@OpenAI Codex can easily do my app's backend work.

[Meris Dabhi @Merisdabhi]

@federicodonaton 🔥

[Meris Dabhi @Merisdabhi]

The AI PC race is getting interesting.

Microsoft is building AI-first hardware.
NVIDIA is building AI-first silicon.

The real shift isn't faster laptops—it's moving AI agents from the cloud onto personal devices.

That changes everything from privacy to latency to cost.

[Microsoft Surface @surface]

Introducing Surface Laptop Ultra.

Built for world makers. Designed for what's next.

The most powerful Surface laptop ever. Coming Fall 2026.

Sign up to learn more: https://t.co/k8aEX2pTAy

[Meris Dabhi @Merisdabhi]

@ElevenLabs @mati Now I have many ideas in my mind.

[Meris Dabhi @Merisdabhi]

@gregisenberg oh man , I really love this guy's content.

[Meris Dabhi @Merisdabhi]

@nvidia No one can stop me

[Meris Dabhi @Merisdabhi]

@boardyai Late for me, but it's really true.