Prathit

saw this tweet and built this within half an hour with antigravity you can just build things

Okay, @gdb is team CLI all the way. @garrytan thinks MCPs suck. So we hit the streets of SF to see if the city agreed. We posed a simple question: MCP or CLI? - Basically everyone under the age of 35 said CLI - One person said MCP was as bloated as Java - & unsurprisingly, numerous people told us to touch grass Final score- MCP: 3 vs CLI: 17 SF has spoken, and @composio listened. Our universal CLI is now live! Drop your best CLI vs MCP hot take in the comments and we'll send the best ones some very sick gear 👀 Link to try our CLI in the next thread ⬇️

Comment CLI if you want this to be printed

i'm devastated, now how do i delete my account?

Anthropic researchers found that playing the right background music improves Claude Code accuracy by 15%

🚨 New Paper Training an LLM to speak low-resource language (EACL workshop, 2026) Tulu is spoken by 2M+ people in coastal Karnataka and LLMs basically can't speak it. We got to 85% grammar accuracy without fine-tuning anything or collecting a single new training example.

@yashwanthsai29 @turilabs insane! congrats

Happy to share a milestone from our work at @turilabs. Two of our research papers have been accepted to ICLR 2026 Workshops. Turilabs started as a small independent AI research lab where the primary goal was to experiment and have fun. Over the last 1 year, a lot of interest was shown towards the same. Seeing our work appear in the ICLR ecosystem is incredibly motivating. When we started Turilabs, the goal was simple: - Build interesting things. - Test bold ideas. - Publish what we learn. Today Turilabs is a close-knit community of around 20 part time researchers, engineers, and builders who regularly discuss ideas around AI systems, agents, and research. For Q2, we will have a lot of interesting experiments. More details about the papers coming soon. This is one step forward in our journey. Excited for what’s next 🚀

I spent $2000 on coding agents last month..... got sent me the bill on slack with just '💀' so I built Mr Krabs, a realtime cost monitor for coding agents supports claude code, codex, opencode, gemini cli fully open source, link below

lmao great phishing attempt

Just Replaced 30 engineers with 30 agents to build an entire codebase in 12 days The agents handled everything: - 44K lines of TypeScript - 175 PRs opened - 1,500+ tests written - All CI failures self-corrected fully open source, link to code below

@aiob_me @interaction @openclaw Devs can't play with it the way they want to

How is @interaction / poke not as popular as @openclaw ? Never had an agent as efficient as poke

@sunilkumrdash wait till it starts applying for loans

Your agent will lose $450k and you'll be happy

@KirkMarple it's quite annoying

@DialecticPod is extremely underrated

Yes, Sarvam is deepseek. Let mr explain. You're comparing deepseek version 3 with Sarvam version 2. Technical wins (in LLM): 1. Data collection and curation of 16 trillion tokens iirc. Deepseek had 14T and they weren't doing low res language like kashmiri dogri etc. Sarvam accomplished low res language and emerging language expertise in M itself 2. ai4bharat and other labs have also done ground breaking research in linguistics language preservation and indic tokenisation. Sarvam is built on top of all the work 2.5- Indic tokenisation works better than gpt4 3. Yes GRPO DualPipe Expert Isolation MTP MLA were a big deal but so is IndicTrans- a low res language translation which inspires sarvam and faithfully translates low res languages better than FB (which has the most well known 1000 language project) 4. Sarvam is yet to release it's technical report. It's 105B compresses 16T tokens better than all models of its size, safe to say it did do some technical innovation. Wait karo thoda . Non LLM wins - Sarvam recently released Sarvam-Translate, a document-level translation model. This is essentially the "next generation" of IndicTrans. It is THE company to preserve Indian epistemology - Sarvam beat Gemini and deepseek on very technical ocr benchmarks. This is not a casual win which you can just flaunt. Doing ocr is way more important than doing LLMs - Sarvam voice is actually good at par with qwen which is the sota on size: performance ratio. Sarvam's dubs are at least as good as heygen. - Training a 16T LLM + a hardware (ai glass) + a complete AI stack with 4000 H100s is actually a huge huge win and believe it or not, no one else could have done it if not for sarvam. -Sarvam has it's business that is call centre automation not exactly a cash cow business been running in parallel to this LLM OCR Dubbing and STT/Tts engine business. . And this is just version 1

Claude Code can now Edit, Compose, Generate Entire Videos in minutes from a single prompt Introducing @vibemotion_ai > A video editor built equally for AI and Humans

Using this tweet to highlight all the llm and other launches in ai impact summit - Sovereign LLM from Sarvam - Sarvam on Nokia phones - 1 billion funding for Neysa - Kaze AI glasses by Sarvam - US India sub-sea fiber optics line - 2 Crore people will be trained by Google - OpenAI for India - Bharatgen models released - Gnani AI India voiceOS models - TCS Helios - Fujitsu "Takane" - Emergent at 100M ARR - fasted from India - Google will give scientist models to scientific research in India - Elevenlabs Bangalore office - L&T NVIDIA AI Factory - NVIDIA x Yotta (Shakti Cloud)- 20,000 blackwells - 100,000 GPUs in Chattisgarh - Sectoral AI Casebooks (Feb 17): 170 real-world AI deployments in gov - SAHI (Strategy for AI in Healthcare for India) - Google live speech to speech for Indian languages - BharatGPT in 2 months - Many small companies released models- will dissect everyone tomorrow - BODH (Benchmarking Open Data Platform for Health AI) - Paraspeak (speech accessibility), AgniSena (disaster response), and Z-TRACS (tracking systems)

@PwnFunction @julianweisser Digital ocean was easiest

any of you run claude code or codex on vps? what's the easiest way to do it?

@marvinvonhagen @morganlinton @interaction poke insulting the user is the best part tho

@morganlinton @interaction wait what happens when we'll open up the gates and let all humans use it without getting harassed and insulted first

Putting in my prediction now, Poke will become a unicorn. Note: I am not an investor, and know nobody at the company. I just have such strong conviction that what they’re working on is one of the biggest market opps today, and I don’t know anyone else this far ahead.

@ai_tulipp really cool! how long did it take?

New Video. ~ Daytona Circuit Video Model: Seedance 2.0 Sound Design: Eleven Labs