How Databricks’ FlashOptim cuts LLM training memory by 50 percent
Training large language models usually requires a cluster of GPUs. FlashOptim changes the math, enabling full-parameter training on fewer accelerators.
AI is writing your code, but who’s reviewing it?
As AI coding assistants go mainstream, a silent wave of technical debt is building. Here’s how the industry is fighting back.
Machine learning in space: Building intelligent systems for the harshest environments
A new breed of tiny, hyper-efficient AI is revolutionizing space, extending satellite life and unlocking the next great era of autonomous exploration.
Decoding the brain, inspiring AI: How Rahul Biswas is bridging neuroscience...
The convergence of AI and neuroscience opens exciting possibilities for understanding human cognition and driving innovation in deep learning.
Recursive Language Models: A new framework for infinite context in LLMs
Brute-forcing larger context windows is hitting a mathematical wall. Here is how MIT’s new framework solves "context rot" to process 10 million tokens and beyond.
Microsoft’s new Rho-alpha model brings tactile sensing to robotics
Microsoft’s Rho-Alpha upgrades Vision-Language-Action models with tactile data to bridge the gap between semantic reasoning and low-level motor control.
Vulnerability in Perplexity’s BrowseSafe shows why single models can’t stop prompt...
Lasso Security compromised Perplexity’s BrowseSafe guardrail model for AI browsers, proving that "out-of-the-box" tools fail to stop prompt injection attacks.
How test-time training allows models to ‘learn’ long documents instead of...
By treating language modeling as a continual learning problem, the TTT-E2E architecture achieves the accuracy of full-attention Transformers on 128k context tasks while matching the speed of linear models.
VL-JEPA is a lean, fast vision-language model that rivals the giants
Meta’s VL-JEPA outperforms massive vision-language models on world modeling tasks by learning to predict "thought vectors" instead of text tokens.
The evolution of LLM tool-use from API calls to agentic applications
A look at the evolution of LLM tool-use, from supervised fine-tuning to Reinforcement Learning (RLVR) and agentic applications in large and specialized models.
Applied ML: When ‘perfect’ becomes the enemy of ‘good’
Waiting for perfect data can stall your machine learning project and result in losing opportunities of creating good-enough models.
AI can’t replace software engineers yet, but here is how to...
AI is not a replacement for engineers, but it can be very useful to product managers testing hypotheses and product ideas.
How to turbocharge your product and market research with DeepSearch
If you think in terms of the JBTD framework, Deep Search products can save you a ton of time and effort in finding new product and market opportunities.
