Models & Labs
New AI Training Methods Introduced
Lev Selectormedium confidence
Why it matters
- →These advancements in training methodologies could revolutionize how AI models are developed and deployed.
Two innovative training methods have been introduced: EGGROLL, which utilizes evolutionary training without backpropagation, and Google TurboQuant, which focuses on extreme KV cache compression. These methods are expected to significantly enhance the efficiency and performance of AI models, paving the way for more advanced applications in various fields. The adoption of these techniques could lead to faster training times and improved model accuracy.
Read originalThe story around this
Earlier coverage that leads up to this article, and what followed. Lines connect each piece to the closest one after it, converging here.
How we got here
- 1New Data Paradigm Introduced
AI Explained · December 19, 2025 · Related
- 2AlphaEvolve: New Coding Agent from DeepMind
AI Explained · December 23, 2025 · Background
- 3New Sequential Attention Method Enhances AI Efficiency
Google Research Blog · February 4, 2026 · Background
- 4AI Native Conf Announces Key Research Breakthroughs
Together AI Blog · March 5, 2026 · Background
- 5New Method Enhances AI Model Explanations
MIT News AI · March 9, 2026 · Background
- 6TurboQuant Introduces Extreme Compression for AI
Google Research Blog · March 24, 2026 · Related
What happened next
- 7AI Agents Increasingly Used in Business
Lev Selector · April 3, 2026 · Background
- 8New Method Enhances AI Model Training Efficiency
MIT News AI · April 9, 2026 · Related
- 9Thinking Machines Labs Unveils New Interaction Models
Matt Wolfe · May 16, 2026 · Background
- 10Microsoft's EvoLib Transforms AI Learning from Experience
Microsoft Research · July 30, 2026 · Background
- 11Microsoft's Echoverse Enhances AI Training Environments
Microsoft Research · July 30, 2026 · Background
More in Models & Labs
Models & Labsother
llama.cpp b12340 adds CUDA 13 and ROCm 10.0
This release quietly closes the hardware gap for local inference by adding default support for CUDA 13 and ROCm 10.0 across Linux and Windows. Users with newer NVIDIA GPUs or AMD accelerators no longer need to compile from source to get performance, while Apple Silicon builds now ship KleidiAI disabled by default, shifting the focus back to standard ARM optimizations. The inclusion of Snapdragon support for Linux and Android marks a significant step toward on-device AI beyond the x86/ARM desktop ecosystem. It is less about new features and more about ensuring llama.cpp remains the universal runtime for every available accelerator.
llama.cpp Releases
Models & Labscoding
llama.cpp b12342 adds few-row MMQ for speculative decoding
Speculative decoding just got a significant speed boost on NVIDIA GPUs. This release introduces a specialized kernel for 'few-row' matrix multiplications, targeting the small batch sizes typical of verification passes in speculative decoding workflows. By optimizing these specific operations for quantized formats like Q8_0 and IQ4 variants, llama.cpp reduces latency where it matters most during token generation. The change is CUDA-only for now but leverages hardware capabilities from Ampere onwards, making local inference faster without requiring new model architectures or complex configuration changes.
llama.cpp Releases
© TechCrunch AIModels & Labsmodels
Google deploys Gemini 4 Hydrogen for faster Search
Google is quietly shifting its search infrastructure by replacing older fast models with Gemini 4 Hydrogen, a tier designed specifically for speed and cost-efficiency rather than raw reasoning power. This model completes tasks 2.5 times faster than its predecessor while beating the more powerful Argon on specific intelligence benchmarks, proving that specialized efficiency can outperform brute force in high-volume workflows. By integrating this into Search's fan-out technique, Google now breaks down complex queries into subqueries with greater precision and speed, delivering richer visual responses without the latency of heavier models. This marks a strategic pivot toward agentic execution layers within consumer search, making AI Mode feel snappier for everyday multi-part questions.
TechCrunch AI