Tag
PyTorch Conference North America in San Jose (Oct 20-21) will feature a poster by Bongwoo Bak of Moreh on KV cache block placement, covering vLLM offloading strategies between GPU, CPU, and shared storage. The post also promotes conference registration tiers and pricing.
The speaker believes machines will replace many jobs within five years, but human connection is irreplaceable. The purpose of human existence was never to work — it is a long transition from competition to cooperation, and he calls for building open-source AI for all of humanity in a neutral country.
PyTorch Conference North America (Oct 20-21, San Jose) spotlights keynote speaker Zongwei Zhou, Software Engineer at Meta, who will explore how TorchTPU and agent-driven development make complex AI workloads portable and performant across hardware.
Ahmad Osman sat down with Robert Scoble to discuss open-source AI, his take on Anthropic, and his conviction that local AI will become the default going forward.
The article questions whether the EU Kids Act or similar child-safety legislation could serve as a Trojan horse to justify restricting or targeting open-weight AI models, raising concerns about regulatory overreach into open AI development.
Cohere VP of Engineering Rachad Alao is announced as a keynote speaker at PyTorch Conference North America (Oct 20-21, San Jose), discussing how open science can help AI foundation models advance with transparency, security, speed, and scale.
Imbue's essay argues that the AI industry's data scraping, closed models, and ad-driven platforms have degraded the web, and calls for a 'Personal Computing 2.0' where ubiquitous, small, open AI models restore user agency and creative computing.
The article announces the PyTorch Ambassador Program and a presentation at PyTorchCon North America 2026, explaining how ambassadors drive PyTorch adoption and providing conference registration details.
The PyTorch Conference North America is an event bringing together practitioners to share insights on open-source AI and real-world applications, with sessions, networking, and workshops, taking place in San Jose from October 20-21.
Chamath warns that banning open-source AI in the U.S. could force companies to use costlier models, harming earnings, valuations, and the stock market.
The article promotes a talk at the PyTorch Conference North America focused on making enterprise agentic inference production-ready using PyTorch and vLLM, covering ecosystem updates and registration details.
The article proposes that open-source AI should focus on developing reusable domain-specific capabilities rather than building large general-purpose models, enabling developers to combine them into local tools similar to how Linux grew through shared components.
The article questions whether open-source AI is keeping up with closed AI, noting that Chinese labs are catching up quickly, which challenges the notion of a clear US lead in AI.
The article introduces 'Intelligence Density' as a metric for measuring cost per task in AI, and presents density-aware training methods that improve model efficiency and performance on professional tasks, demonstrated with NVIDIA Nemotron models.
The article discusses lessons from the first public disclosure of an autonomous agent cyberattack, highlighting the need for AI transparency and open-source tools to mitigate security risks.
The tweet advocates for the importance of open systems and open-source AI, arguing that openness leads to freedom and should prevail in AI development.
The article discusses the pattern where open-source AI models like Qwen match frontier capabilities about two quarters later, exemplified by GPT-5 and Qwen3.5-27B, and questions if this trend will continue with models like Astra and Fable 5.1.
In July 2023, the user built a high-end computer with 8x RTX 3090s and an Epyc 7004 CPU to run Llama 2 70B locally, expressing confidence in the future of open-source AI.
LLM2Jev is an open-source tool that adapts local HuggingFace models to perform structured decisions with Choice, Score, and Noul frameworks, offering prefill-only inference and integration with Transformers and SGLang.
The PyTorch Conference North America 2026 will feature a session by Red Hat engineers on Cross-Repository CI Relay to streamline CI integration for downstream repositories. Registration details and event highlights are provided.