Rio was amazing, I had such a fun time at @iclr_conf @liquidai presented many papers, managed a busy booth, announced a partnership with Mercedes, and organized a (delicious) dinner event 10/10 would do it again!
@maximelabonne
-
Tokenizer: The Real Business Model Behind AI Models
By
–
The model might be the product, but the tokenizer is the business plan
-
Anthropic’s Hidden 30% Cost Increase via Tokenizer Change
By
–
Turns out Anthropic basically snuck in a 30% cost increase by changing the tokenizer
-
Synthetic Data Generalization Limitations Beyond Benchmarks
By
–
Definitely a big limitation. Overall, the problem with synthetic data is generalization beyond the benchmarks that they target in the first place. This is where the most interesting results can be found.
-
Synthetic Pretraining Improves Reasoning in Sub-1B Models
By
–
Synthetic pretraining for sub-1B reasoning models Cool write-up from Tufa Labs (Matteo Saponati) on whether synthetic data augmentation actually helps very small (<1B) models reason better. They pretrain a 0.8B model with the Qwen3 architecture from scratch on 12B tokens of
-
In-Car AI Assistant Translates Speech into Personalized Function Calls
By
–
For example, an in-car assistant that translates speech (intents) into function calls with extra personalization
-
Liquid’s Approach to Small Models and Agentic Reinforcement Learning
By
–
Quick chat about what Liquid does with small models, unique challenges, and smol agentic RL 👀
— Maxime Labonne (@maximelabonne) 28 avril 2026
Thanks for the invitation @josephinePqt1! https://t.co/CDpKefwLPLQuick chat about what Liquid does with small models, unique challenges, and smol agentic RL Thanks for the invitation @josephinePqt1
! -
Real-World AI Use Cases: Function Calling and Data Extraction
By
–
Some real-word use cases from past year: function calling, data extraction, query rewriting, personal assistants, intent classification, complexity routing, data rephrasing, image captioning, object detection, live translation, etc. We actively work in this field at @liquidai
! -
RL Datasets Addition Planned for Next Update
By
–
I'm planning to (finally) add RL datasets in the next update