If you're building with agents, or planning to, this is the protocol to watch.
In our deep dive into A2A you'll learn how it works and how to start with it, whether MCP and A2A are competitors, and if Google might use it to index every agent on the internet 👇
Enjoy and leave your feedback!
• Specialist agents working together like modular teams
• Easy cross-enterprise workflows
• Standardized human-in-the-loop collaboration between AI and people
• And even a searchable, internet-scale agents directory
Why is A2A important?
Most AI agents today live in silos. @Google’s A2A protocol aims to be the “common language” that lets them to collaborate.
A2A could unlock many possibilities:
People want to understand agentic infrastructure protocols better. The strong response to our MCP article shows there’s real demand for clarity around standardization of AI ecosystems
Since so many people asked, we are making our article on Agent2Agent (A2A) free to read on @hf.co
🧵
Hyena Edge is an experimental convolutional multi-hybrid model, It runs efficiently on smaller devices like your phone. At its core, it replaces 2/3 of attention with fast convolutions and gating
And Liquid AI are working on something even more interesting
How can their models beat Transformers? 👇
The most important features of LFMs (Liquid Foundation Models) from Liquid AI?
Memory-efficiency, inference speed, without compromising model quality.
LFMs have been benchmarked on real hardware, proving that they can beat Transformers.
Liquid AI have also just released Hyena Edge👇
Podcasts:
YouTube www.youtube.com/@RealTuringP...
Spotify open.spotify.com/show/2SQxCUR...
Apple podcasts.apple.com/us/podcast/i...
If you like it, please, subscribe to the Turing Post's YouTube channel – Inference, and listen to our podcasts on all major platforms.
YouTube channel - 'Inference'-> www.youtube.com/@RealTuringP...
What happens when the biggest advocate for coding literacy starts telling people not to learn to code?
In the new Inference episode, I sat down with Amjad Masad, CEO and co-founder at Replit, to discuss the evolution in coding.
Are we entering a post-coding world?
www.youtube.com/watch?v=PlDe...
See the full list of the freshest research and other important news of the week in our free newsletter: www.turingpost.com/p/fod98
7. Values in the Wild: Discovering and Analyzing Values in Real-World Language Model Interactions, @anthropicai.bsky.social
Maps AI value expressions across real-world interactions to inform grounded AI value alignment
arxiv.org/abs/2504.15236
6. Roll the dice & look before you leap: Going beyond the creative limits of next-token prediction
Highlights limitations of next-token prediction and proposes noise-injection strategies for open-ended creativity
arxiv.org/abs/2504.15266
GitHub: github.com/chenwu98/alg...
5. The Sparse Frontier: Sparse Attention Trade-offs in Transformer LLMs
Investigates sparse attention trade-offs and proposes scaling laws for long-context LLMs
arxiv.org/abs/2504.17768
4. Efficient Pretraining Length Scaling
Presents PHD-Transformer to enable efficient long-context pretraining without inflating memory costs
arxiv.org/abs/2504.14992
3. Paper2Code
Automates end-to-end ML paper-to-code translation with a multi-agent framework
arxiv.org/abs/2504.17192
Code: github.com/going-doer/P...
2. LLMs are Greedy Agents: Effects of RL Fine-tuning on Decision-Making Abilities
Analyzes how RL fine-tuning improves exploration and decision-making abilities of LLMs
arxiv.org/abs/2504.16078
1. TTRL: Test-Time Reinforcement Learning
Introduces a method for self-evolving LLMs at test-time using reward signals without labeled data
arxiv.org/abs/2504.16084
GitHub: github.com/PRIME-RL/TTRL
Top 7 research papers of the week:
▪️ Test-Time Reinforcement Learning
▪️ LLMs are Greedy Agents
▪️ Paper2Code
▪️ Efficient Pretraining Length Scaling
▪️ The Sparse Frontier
▪️ Roll the dice & look before you leap
▪️ Discovering and Analyzing Values in Real-World Language Model Interactions
🧵
9. Skywork R1V2: Multimodal Hybrid Reinforcement Learning for Reasoning
Advances multimodal reasoning with a hybrid RL paradigm balancing reward guidance and rule-based strategies.
arxiv.org/abs/2504.16656
Model: huggingface.co/Skywork/Skyw...
8. Process Reward Models That Think introduces ThinkPRM
It's a generative verifier that scales step-wise reward modeling with minimal supervision.
arxiv.org/abs/2504.16828
GitHub: github.com/mukhal/think...
7. Surya OCR:
Release an open-source, high-speed OCR model supporting 90+ languages with LaTeX formatting and structured output for real-world document processing.
x.com/VikParuchuri...
6. Trillion-7B:
Develops a highly token-efficient multilingual LLM using specialized cross-lingual techniques for Korean, Japanese, and more.
huggingface.co/papers/2504....
Model:
huggingface.co/trillionlabs...
5. Eagle 2.5 by NVIDIA
Expands vision-language models to handle long-context video and image comprehension with specialized training tricks and efficient scaling.
arxiv.org/abs/2504.15271
Project page: nvlabs.github.io/EAGLE/
4. Aimo-2 winning solution by Nvidia
Builds state-of-the-art mathematical reasoning models with OpenMathReasoning dataset.
arxiv.org/abs/2504.16891
3. Kimi-Audio
Builds a universal audio foundation model for understanding, generating, and conversing in audio and text, achieving SOTA across diverse benchmarks.
arxiv.org/abs/2504.18425
Codes, model checkpoints, the evaluation toolkits: github.com/MoonshotAI/K...
2. Tina: Tiny Reasoning Models via LoRA:
Achieve strong reasoning capabilities with tiny models by applying cost-efficient low-rank adaptation and reinforcement learning.
arxiv.org/abs/2504.15777
1. Hyena Edge by @LiquidAI_
Design a convolution-based hybrid architecture to outperform Transformer models in speed, memory, and quality on smartphones and other edge devices.
www.liquid.ai/research/con...
9 notable AI models of the week:
▪️ Hyena Edge
▪️ Tina: Tiny Reasoning Models via LoRA
▪️ Kimi-Audio
▪️ Aimo-2 winning solution
▪️ Eagle 2.5
▪️ Trillion-7B
▪️ Surya OCR
▪️ ThinkPRM
▪️ Skywork R1V2
🧵
