How Far Can Classical NLP Go? From Bag-of-Words to Stacking on Spooky Author Identification

Unlocking the Power of the TPU Stack: Introducing our new Developer Hub

Google Cloud vừa giới thiệu TPU Developer Hub, một nền tảng giáo dục tập trung dành cho nhà phát triển ML sử dụng TPU, bao gồm kiến trúc phần cứng, stack phần mềm (XLA, Pallas kernels), công cụ gỡ lỗi XProf, chiến lược tối ưu hóa (như offloading KV cache) cùng networking và bảo mật. Nội dung đa dạng từ Colabs tương tác, mã nguồn mở đến tài liệu chuyên sâu, hỗ trợ tích hợp AI-assisted development.

Lập trình viên ML nên đọc để hiểu cách tối ưu hóa hiệu suất và chi phí của mô hình trên TPU với các công cụ mới như XLA, Pallas và các chiến lược parallelism, từ đó tiết kiệm thời gian và nguồn lực trong triển khai sản phẩm AI.

Ex-Tesla Optimus engineer settles trade secret lawsuit and raises $11M to build robot hands

Proception, a robotics startup founded by ex-Tesla Optimus engineer Jay Li, has settled a year-long trade secret lawsuit with Tesla and raised an $11M seed round led by First Round Capital. The company is now shipping its first dexterous robotic hands — featuring 22 degrees of freedom and a sensor-laden glove for scalable training data collection — to researchers and robotics companies. Tesla had sued Li in June 2025, alleging he downloaded confidential files before leaving to found Proception. The dexterous hand market is heating up globally, with competitors like China's Linkerbot dominating market share and European and Chinese startups raising hundreds of millions. Proception is betting that humanoid robot makers will outsource hand manufacturing rather than build in-house, positioning itself as a key supplier in the emerging humanoid robot supply chain.

Twisted Tethers Make Tidal Energy Cheaper and Cleaner

Researchers at Princeton have developed AI-driven methods to automate the design of radio-frequency integrated circuits (RFICs), a process traditionally considered a 'dark art' requiring years of expert experience. Using reinforcement learning to determine circuit architecture and topology from scratch, combined with an AI emulator that predicts electromagnetic behavior without running full simulations, the team produced a 5G millimeter-wave power amplifier that outperformed state-of-the-art human designs. They also applied diffusion models to make AI-generated circuit structures more interpretable by engineers. The approach eliminates reliance on human-designed templates, dramatically reduces design time from months to minutes, and has been validated for multiport circuits and sub-terahertz amplifiers. Key remaining challenges include reducing AI hallucinations, improving generalizability, and building open datasets — since most simulation data is locked behind NDAs.

DiScoFormer: One transformer for density and score, across distributions

DiScoFormer (Density and Score Transformer) is a new model from AI2 that estimates both the probability density and score of a distribution from a finite sample in a single forward pass, without retraining per distribution. Built on transformer architecture with cross-attention, it analytically generalizes kernel density estimation (KDE) and outperforms it significantly in high dimensions — cutting score error by ~6.5x and density error by ~37x in 100 dimensions. Trained on Gaussian Mixture Models (which have closed-form densities), it generalizes to unseen distributions including Laplace and Student-t. A consistency loss between the density and score heads enables test-time adaptation to out-of-distribution inputs. The model targets shared infrastructure for generative modeling, Bayesian inference, and scientific computing.

3 Questions: Beyond data-driven aesthetics

MIT Architecture alumnus Alexandros Haridis discusses his exhibition 'Beyond Data-Driven Aesthetics' at MIT's Keller Gallery, which traces the history of computational aesthetics from the 1930s through today. The exhibition translates dense algorithmic research — covering topics like George Birkhoff's mathematical aesthetic measure and the AICAN machine learning system — into physical installations and interactive visualizations. Haridis argues that many questions now framed as novel AI challenges around creativity and aesthetic judgment have deep roots in 20th-century computing research. The work is organized around five themes: Aesthetic Measure, Aesthetic Guidelines, Algorithmic Aesthetics, Aesthetic Appropriation, and Aesthetic Novelty, using design itself as an interpretive method to make opaque computational systems legible and experiential.

WiMi Explores Neural Networks for Twin-Field Quantum Key Distribution Optimization

WiMi Hologram Cloud is researching neural network models to optimize parameter selection in twin-field quantum key distribution (TF-QKD) systems. Three models were evaluated: BPNN, RBFNN, and GRNN. RBFNN and GRNN showed higher prediction accuracy in high-dimensional parameter spaces, while BPNN offered the fastest computation speed. Compared to traditional local search algorithms (LSA), neural network-based prediction reduced computation time by multiple orders of magnitude. Future work will explore deep learning and reinforcement learning approaches, with plans to integrate the technology into quantum communication hardware for practical deployment.

Towards Understandable Software

Programming is fundamentally broken — too complex, too boilerplate-heavy, and too inaccessible — and LLMs are not the fix. They automate bad code rather than abstracting it away. The real solution is to raise the level of abstraction through three complementary approaches: literate programming (writing documentation first, code second, using tools like Entangled), visual programming (GUI-based IDEs that make software creation accessible without code), and deterministic natural language programming (NLP-based compilers that translate human language into machine code predictably, unlike LLMs). Projects like Eve and Inform have explored this space before. The author is building ReTangled, a Rust-based literate programming tangler, and calls for a community effort to create accessible visual or natural language programming environments.