🚨 News - Academic Integrity | Drishtant Ghosh (Drix10)

🚨 News - Academic Integrity

Mumbai Police Book Professor Suryanarayana Doolla Under SC/ST Act

Key Points:

Unsubstantiated Claims: Mumbai police booked Professor Suryanarayana Doolla under the SC/ST Act based on unsubstantiated claims.
IIT Bombay Suspension: IIT Bombay has suspended Mr. Doolla from the position of Dean of Administrative Affairs.
Academic Integrity: The incident highlights the importance of academic integrity and the need for thorough investigations before taking action.

🔗 Resources:

🤖 Robotics - Safety Engineering

Safety Engineered Across the Full Lifecycle

Key Points:

Safety Engineering: Safety engineered across the full lifecycle is crucial for turning humanoids from demos into deployments.
NVIDIA Robotics: Glad to be building on this with @NVIDIARobotics.
Robotics Safety: The importance of safety engineering in robotics cannot be overstated.

🔗 Resources:

🎨 AI - Muse Spark

Right-Click Anything: A Demo Combining Muse Spark, SAM 3.1, and Muse Image

Key Points:

Muse Spark: Muse Spark captions the scene and generates noun-phrases.
SAM 3.1: SAM 3.1 predicts instance masks for each object.
Right-Click Demo: The demo showcases the capabilities of Muse Spark and SAM 3.1.

🔗 Resources:

Image

📊 Research - Scaling Laws

Scaling Laws Without Losing Fit or Predictability

Key Points:

Scaling Laws: The paper explores the limits of scaling laws without losing fit or predictability.
Small-Scale Experiments: The authors conducted small-scale experiments to test the limits of scaling laws.
Predictability: The paper highlights the importance of predictability in scaling laws.

🔗 Resources:

Image

📚 Research - Paper Discussion

Paper Discussion: Scaling Laws Without Losing Fit or Predictability

Key Points:

Paper Discussion: The author discusses the paper on scaling laws without losing fit or predictability.
Authors: The authors of the paper are @LotfiSanae, @NickLourie, @Kyunghyu, and @karen_ullrich.
Paper Link: The paper can be found at https://arxiv.org/abs/2608.11859 ↗.

🔗 Resources:

🤝 Networking - Meet Granola

Meet Granola: A Meeting Platform for Networking

Key Points:

Meet Granola: Meet Granola is a meeting platform that helps people connect with each other.
Happenstance AI: Happenstance AI surfaces who in your network you should connect with.
Networking: The platform makes networking easier and more efficient.

🔗 Resources:

Image

🤖 Robotics - Model Hardware Standard

Anthropic's Model Hardware Standard (MHS)

Key Points:

MHS: Anthropic's Model Hardware Standard (MHS) is a framework for giving physical objects to AI agents.
Robot Arms: AI agents can already use browser APIs, databases, and MCP computers.
Future Developments: The next step is to give physical objects like robot arms, microscopes, and lab equipment to AI agents.

🔗 Resources:

📊 AI - GPT-6 Astra

GPT-6 Astra: The Best Vision Model We Have Tested

Key Points:

GPT-6 Astra: GPT-6 Astra is the best vision model we have tested.
Detection, Segmentation, and Reasoning: The model performs well on detection, segmentation, box prompting, counting, and reasoning tasks.
Trade-Offs: The model has trade-offs in terms of speed and accuracy.

🔗 Resources:

🤝 OpenAI - Independent Advisory Group

OpenAI's Independent Advisory Group

Key Points:

Independent Advisory Group: OpenAI is working with an independent advisory group of mathematicians to help responsibly share advances in AI and mathematics.
Assessing and Communicating New Mathematical Results: The group will advise on how to assess and communicate new mathematical results.
Upholding Academic and Professional Standards: The group will also advise on how to uphold academic and professional standards.

🔗 Resources: