Pi School

🚀 𝐏𝐢 𝐀𝐈 𝐖𝐞𝐞𝐤𝐥𝐲 𝐓𝐫𝐞𝐧𝐝𝐬 #𝟗𝟎 𝐢𝐬 𝐡𝐞𝐫𝐞
It’s Friday! Get ready to stay ahead with the latest AI breakthroughs, handpicked by our Deep Learning Scientist, Mubashir Shah.
This week’s highlights:

⚡ 𝐌𝐢𝐧𝐢𝐌𝐚𝐱 𝐒𝐩𝐚𝐫𝐬𝐞 𝐀𝐭𝐭𝐞𝐧𝐭𝐢𝐨𝐧: 𝟐𝟖𝐱 𝐂𝐨𝐦𝐩𝐮𝐭𝐞 𝐂𝐮𝐭 𝐚𝐭 𝟏𝐌 𝐂𝐨𝐧𝐭𝐞𝐱𝐭
MiniMax introduces MSA, a blockwise sparse attention mechanism built on Grouped Query Attention that scores and selects only the most relevant KV blocks per query group. The production model it powers, MiniMax-M3 (109B params), matches full GQA quality while cutting attention compute by 28.4x at 1M-token context, with 14.2x prefill and 7.6x decoding speedups on H800s. Model and kernels are open on Hugging Face and GitHub.
🌐 https://pischool.link/e53a40

🧪 𝐀𝐠𝐞𝐧𝐭𝐬' 𝐋𝐚𝐬𝐭 𝐄𝐱𝐚𝐦: 𝐀 𝐑𝐞𝐚𝐥𝐢𝐭𝐲 𝐂𝐡𝐞𝐜𝐤 𝐟𝐨𝐫 𝐀𝐈 𝐀𝐠𝐞𝐧𝐭𝐬
Built with 250+ industry experts, ALE benchmarks AI agents on 1,000+ economically valuable real-world tasks spanning 55 subfields across 13 industry clusters. Despite rapid progress in agent systems, the hardest tier remains largely unsolved, with average full pass rates below 1% across mainstream configurations. The benchmark is designed to track whether AI progress translates into meaningful real-world economic impact.
🌐 https://pischool.link/999139

🤖 𝐖𝐨𝐫𝐥𝐝 𝐀𝐜𝐭𝐢𝐨𝐧 𝐌𝐨𝐝𝐞𝐥𝐬: 𝐓𝐡𝐞 𝐍𝐞𝐱𝐭 𝐅𝐫𝐨𝐧𝐭𝐢𝐞𝐫 𝐢𝐧 𝐄𝐦𝐛𝐨𝐝𝐢𝐞𝐝 𝐀𝐈
OpenMOSS introduces World Action Models (WAMs), a framework for embodied foundation models that jointly model future world states and actions. By unifying world modelling and policy learning, WAMs aim to provide a shared representation of environment dynamics and decision making. The survey organises existing approaches into Cascaded and Joint WAM taxonomies and maintains an open-source tracker of developments across the field.
🌐 https://pischool.link/6742db

Was this helpful? Let us know by liking and sharing!

🚀 𝐏𝐢 𝐀𝐈 𝐖𝐞𝐞𝐤𝐥𝐲 𝐓𝐫𝐞𝐧𝐝𝐬 #𝟖𝟖
3 things our Deep Learning Scientist Giuseppe Tanzi is paying attention to this week.

🚀 𝐍𝐀𝐒𝐀 𝐇𝐏𝐒𝐂: 𝐀𝐈-𝐑𝐞𝐚𝐝𝐲 𝐒𝐩𝐚𝐜𝐞𝐜𝐫𝐚𝐟𝐭 𝐏𝐫𝐨𝐜𝐞𝐬𝐬𝐨𝐫 𝐚𝐭 𝟓𝟎𝟎× 𝐂𝐮𝐫𝐫𝐞𝐧𝐭 𝐏𝐞𝐫𝐟𝐨𝐫𝐦𝐚𝐧𝐜𝐞
NASA and Microchip Technology are testing the High Performance Spaceflight Computing (HPSC) processor, a radiation-hardened chip that delivers performance hundreds of times that of current spaceflight computers while surviving tests designed to mimic the harsh conditions of space. The technology will enable autonomous spacecraft to use artificial intelligence to respond in real time to complex situations and environments where human input isn't possible.
🌐 https://pischool.link/nasa

⚡𝐄𝐱𝐜𝐢𝐭𝐨𝐧-𝐏𝐨𝐥𝐚𝐫𝐢𝐭𝐨𝐧𝐬: 𝐋𝐢𝐠𝐡𝐭-𝐌𝐚𝐭𝐭𝐞𝐫 𝐏𝐚𝐫𝐭𝐢𝐜𝐥𝐞𝐬 𝐟𝐨𝐫 𝐀𝐈 𝐂𝐨𝐦𝐩𝐮𝐭𝐢𝐧𝐠
Researchers at the University of Pennsylvania demonstrated all-optical signal switching using exciton-polaritons, using only about 4 quadrillionths of a joule of energy, far below the energy needed to power a tiny LED briefly. If scaled, the technology could lead to photonic chips capable of processing information directly from cameras without repeated conversions between light and electricity, lowering the massive energy demands of large AI systems and potentially supporting basic quantum computing functions.
🌐 https://pischool.link/Ectnplrtn

🧠 𝐓𝐋𝐓: 𝐓𝐚𝐦𝐢𝐧𝐠 𝐭𝐡𝐞 𝐋𝐨𝐧𝐠 𝐓𝐚𝐢𝐥 𝐢𝐧 𝐑𝐞𝐚𝐬𝐨𝐧𝐢𝐧𝐠 𝐋𝐋𝐌 𝐓𝐫𝐚𝐢𝐧𝐢𝐧𝐠
MIT researchers identified that the rollout phase consumes a disproportionately large fraction (~85%) of total RL training step time, creating a major bottleneck for reasoning LLMs. Their solution, TLT, uses idle processor downtime to continuously train a lightweight drafter model on the fly, keeping it aligned with the target model at zero extra cost. Tested across multiple reasoning LLMs, TLT accelerated training between 70 and 210 per cent while preserving the accuracy of each model.
🌐 https://pischool.link/MIT

𝐖𝐡𝐢𝐜𝐡 𝐨𝐟 𝐭𝐡𝐞𝐬𝐞 𝐰𝐨𝐮𝐥𝐝 𝐜𝐡𝐚𝐧𝐠𝐞 𝐡𝐨𝐰 𝐲𝐨𝐮 𝐰𝐨𝐫𝐤? 𝐃𝐫𝐨𝐩 𝐢𝐭 𝐢𝐧 𝐭𝐡𝐞 𝐜𝐨𝐦𝐦𝐞𝐧𝐭𝐬.