Johan Sosa (@johansosas) 's Twitter Profile
Johan Sosa

@johansosas

Sic itur ad singularitatem 🚀

ID: 241903818

calendar_today23-01-2011 12:37:14

523 Tweet

124 Takipçi

778 Takip Edilen

Juan Stoppa (@juanstoppa) 's Twitter Profile Photo

🚀 Happy to introduce Daily Vibe Casting: a daily AI-generated podcast covering the most viral 𝕏 posts from the previous 24 hours Drops daily on Daily Vibe Casting If you spend time on 𝕏, I think you’ll like it! Would love feedback, podcast link in comments Why I built this?

Qwen (@alibaba_qwen) 's Twitter Profile Photo

>>> Qwen3-Coder is here! ✅ We’re releasing Qwen3-Coder-480B-A35B-Instruct, our most powerful open agentic code model to date. This 480B-parameter Mixture-of-Experts model (35B active) natively supports 256K context and scales to 1M context with extrapolation. It achieves

>>> Qwen3-Coder is here! ✅

We’re releasing Qwen3-Coder-480B-A35B-Instruct, our most powerful open agentic code model to date. This 480B-parameter Mixture-of-Experts model (35B active) natively supports 256K context and scales to 1M context with extrapolation. It achieves
Hunyuan (@tencenthunyuan) 's Twitter Profile Photo

We're thrilled to release & open-source Hunyuan3D World Model 1.0! This model enables you to generate immersive, explorable, and interactive 3D worlds from just a sentence or an image. It's the industry's first open-source 3D world generation model, compatible with CG pipelines

alphaXiv (@askalphaxiv) 's Twitter Profile Photo

Introducing NotebookLM for arXiv papers 🚀 Transform dense AI research into an engaging conversation With context across thousands of related papers, it captures motivations, draws connections to SOTA, and explains key insights like a professor who's read the entire field

Rohan Paul (@rohanpaul_ai) 's Twitter Profile Photo

The paper builds a self-improving coding agent that chooses changes by lineage potential, not one-off scores. 2.38x fewer CPU-hours than prior methods, and human-level results on SWE-bench Lite. They call the gap between scores and growth the metaproductivity performance

The paper builds a self-improving coding agent that chooses changes by lineage potential, not one-off scores. 

2.38x fewer CPU-hours than prior methods, and human-level results on SWE-bench Lite.

They call the gap between scores and growth the metaproductivity performance
Krishna Agrawal (@krishnasagrawal) 's Twitter Profile Photo

💥 Top 50 LLM Interview Questions & Answers Get ready for your next AI / ML / LLM interview with this power-packed Q&A guide covering: ✅ Prompt Engineering ✅ Fine-tuning & RAG ✅ Transformer Architecture ✅ Tokenization & Attention ✅ Real-world LLM Scenarios Perfect for

💥 Top 50 LLM Interview Questions & Answers

Get ready for your next AI / ML / LLM interview with this power-packed Q&A guide covering:
✅ Prompt Engineering
✅ Fine-tuning & RAG
✅ Transformer Architecture
✅ Tokenization & Attention
✅ Real-world LLM Scenarios

Perfect for
Sam Rodriques (@sgrodriques) 's Twitter Profile Photo

Today, we’re announcing Kosmos, our newest AI Scientist, available to use now. Users estimate Kosmos does 6 months of work in a single day. One run can read 1,500 papers and write 42,000 lines of code. At least 79% of its findings are reproducible. Kosmos has made 7 discoveries

Chao Huang (@huang_chao4969) 's Twitter Profile Photo

🚀 Our lab just open-sourced DeepTutor - AI-Powered Personalized Learning Assistant We're actively exploring how agentic AI can assist students and researchers throughout their learning process. DeepTutor brings together research, problem-solving, practice generation, and

🚀 Our lab just open-sourced DeepTutor - AI-Powered Personalized Learning Assistant

We're actively exploring how agentic AI can assist students and researchers throughout their learning process. DeepTutor brings together research, problem-solving, practice generation, and
McKinsey Global Institute (@mckinsey_mgi) 's Twitter Profile Photo

At current levels of capability, AI agents could perform tasks that occupy 44% of US work hours today, and robots 13%. That means more than half of current work hours could be automated with today’s technologies. But work that draws on social and emotional skills remains

At current levels of capability, AI agents could perform tasks that occupy 44% of US work hours today, and robots 13%.

That means more than half of current work hours could be automated with today’s technologies.

But work that draws on social and emotional skills remains
Michael Truell (@mntruell) 's Twitter Profile Photo

We built a browser with GPT-5.2 in Cursor. It ran uninterrupted for one week. It's 3M+ lines of code across thousands of files. The rendering engine is from-scratch in Rust with HTML parsing, CSS cascade, layout, text shaping, paint, and a custom JS VM. It *kind of* works! It

We built a browser with GPT-5.2 in Cursor. It ran uninterrupted for one week.

It's 3M+ lines of code across thousands of files. The rendering engine is from-scratch in Rust with HTML parsing, CSS cascade, layout, text shaping, paint, and a custom JS VM.

It *kind of* works! It
MIT Sloan School of Management (@mitsloan) 's Twitter Profile Photo

With the rapid adoption of AI technologies, the MIT Center for Information Systems Research created a business model framework for the AI era that shows businesses evolving to become increasingly outcome oriented and enabled by autonomous AI. Learn more: bit.ly/4aUDML5

With the rapid adoption of AI technologies, the MIT Center for Information Systems Research created a business model framework for the AI era that shows businesses evolving to become increasingly outcome oriented and enabled by autonomous AI.
 
Learn more: bit.ly/4aUDML5