Ivatois's profile picture. Entrepreneur.Writer.Founder of Choreo AI.

Lucien - Choreo AI

@Ivatois

Entrepreneur.Writer.Founder of Choreo AI.

Lucien - Choreo AI reposted

NVIDIA announces Vera rubin: • Up to 10x lower inference token costs (vs. Blackwell) • 4x fewer GPUs required for MoE training (vs. Blackwell) • 5x better energy efficiency (Spectrum-X Photonics) • 5x longer uptime (Spectrum-X Photonics) • 10x higher reliability…


Lucien - Choreo AI reposted

Long time ago

Elon Musk on building his first startup Zip2 In 1995, when he was just 23 years old, Elon dropped out of Stanford’s PhD program in physics to start Zip2 with his brother Kimbal Musk. Elon personally wrote the first national maps, directions, yellow pages and white pages on the…



6 tools where you can self-host n8n: Choreo AI (choreoai.us) – $100 one-time, hosted for you GDPR compliant, choose your server location, install and update with simple clicks Render – Free / $7+, DIY setup Railway – Free / $5+, limited regions DigitalOcean – $5+/mo,


3 AI automation tools every dev should know: ⚙️ choreoai.us — self-hosted agent flows (GDPR-safe) 🧱 langchain.com — orchestration + memory 🤖 github.com/microsoft/auto… — multi-agent framework Stack these. Ship smarter.


🧠 Real talk: AI that just generates is old news. The next layer? AI that orchestrates, adapts, and self-hosts — with context and compliance baked in. We built choreoai.us so anyone can deploy their own agents, plug in real workflows (think n8n-level control),


While world leaders argue over tariffs and alliances, some of us are just trying to build tools that let people host their own AI workflows in peace 😅 choreoai.us — GDPR-safe, no foreign dependencies 🇪🇺🤝


Lucien - Choreo AI reposted

We just put out a key step for making distributed training work at larger and larger models: Scaling Laws for DiLoCo TL;DR: We can do LLM training across datacenters in a way that scales incredibly well to larger and larger models!

MatharyCharles's tweet image. We just put out a key step for making distributed training work at larger and larger models: Scaling Laws for DiLoCo

TL;DR: We can do LLM training across datacenters in a way that scales incredibly well to larger and larger models!

Grok 3 is brutal but loves the insight😁

Ivatois's tweet image. Grok 3 is brutal but loves the insight😁
Ivatois's tweet image. Grok 3 is brutal but loves the insight😁
Ivatois's tweet image. Grok 3 is brutal but loves the insight😁
Ivatois's tweet image. Grok 3 is brutal but loves the insight😁

Lucien - Choreo AI reposted

Veo 2 currently costs around $0.50 per second (fal AI). Kling AI 1.6 costs $0.35 per video. — $4 for an 8s Veo 2 video. $0.35 for a 5s Kling AI 1.6 video. In this case, Kling AI is 7x cheaper and delivers almost the same quality. This is crazy.


Lucien - Choreo AI reposted

Open Deep Research Deep research is one of the most popular agent use-cases. Here is an open deep researcher w/ ability to configure report structure, planner/writer LLMs, search APIs, search depth, etc. 📽️ youtu.be/2mSNIX-l_Zc Code: github.com/langchain-ai/o…

LangChain's tweet image. Open Deep Research

Deep research is one of the most popular agent use-cases. Here is an open deep researcher w/ ability to  configure report structure, planner/writer LLMs, search APIs, search depth, etc.  

📽️
youtu.be/2mSNIX-l_Zc

Code:
github.com/langchain-ai/o…

Lucien - Choreo AI reposted

After 6+ months in the making and burning over a year of GPU compute time, we're super excited to finally release the "Ultra-Scale Playbook" Check it out here: hf.co/spaces/nanotro… A free, open-source, book to learn everything about 5D parallelism, ZeRO, fast CUDA kernels,…

Thom_Wolf's tweet image. After 6+ months in the making and burning over a year of GPU compute time, we're super excited to finally release the "Ultra-Scale Playbook"

Check it out here: hf.co/spaces/nanotro…

A free, open-source, book to learn everything about 5D parallelism, ZeRO, fast CUDA kernels,…

Lucien - Choreo AI reposted

We replicated the DeepSeek-R1-Zero and DeepSeek-R1 training on 7B model with only 8K examples, the results are surprisingly strong. 🚀 Starting from Qwen2.5-Math-7B (base model), we perform RL on it directly. No SFT, no reward model, just 8K MATH examples for verification, the…

junxian_he's tweet image. We replicated the DeepSeek-R1-Zero and DeepSeek-R1 training on 7B model with only 8K examples, the results are surprisingly strong. 

🚀 Starting from Qwen2.5-Math-7B (base model), we perform RL on it directly. No SFT, no reward model, just 8K MATH examples for verification, the…
junxian_he's tweet image. We replicated the DeepSeek-R1-Zero and DeepSeek-R1 training on 7B model with only 8K examples, the results are surprisingly strong. 

🚀 Starting from Qwen2.5-Math-7B (base model), we perform RL on it directly. No SFT, no reward model, just 8K MATH examples for verification, the…

🚀 Our AI Sales Agent can now recommend topics for practice! Try it out with no sign-up required. islandaiforge.us/index-coach.ht…


I’m excited to share something I’ve been working on! 🎙️ I’ve built an AI vocal assistant that lets customers shop online just by talking—no typing, no hassle. It’s all about making the experience smooth and enjoyable. DM🚀 #AI #Ecommerce #VoiceAssistant


🚗💬 Want to level up your automotive sales team? Try our AI Sales Coach! Practice real-world pitches, get instant feedback, and improve fast. Ready to crush it? 💥 app.islandaiforge.us/signup #SalesCoach #AI #AutomotiveSales #SalesTraining #LevelUp


📷 Practice Makes Perfect — But Who’s Your Practice Partner? 📷 Meet the AI Sales Coach: Your 24/7 role-play partner for perfecting sales calls.


🔥 Tired of angry customers? Let our 🤖 AI-powered voice assistant handle it! Instantly respond to complaints, resolve issues fast, & turn frustration into loyalty. 🚀 Integrate with your Shopify store & boost satisfaction! ⭐ DM us to get started! #AI #Shopify #CustomerSupport


Loading...

Something went wrong.


Something went wrong.