Technology

The Download: how to run an LLM, and a history of “three-parent babies”

Source

MIT Technology Review

Published

Jul 18, 2025

TL;DR

AI Generated

The article discusses how advancements in technology have made it possible for individuals to run large language models (LLMs) on their laptops or smartphones without the need for expensive GPUs. It also provides a guide on how to set up and run a local model for those interested in privacy or independence from big LLM companies. Additionally, the article explores the history and controversy surrounding "three-parent babies," with eight babies recently born in the UK using an experimental IVF technique involving DNA from three people to prevent genetic diseases. The piece also highlights other tech news, including OpenAI's ChatGPT Agent, the White House's stance on "woke AI," and Elon Musk's plans for SpaceX rockets.

Read Full Article

Microsoft will allow users to indefinitely pause updates in Windows 11 — first change in over a decade to the mandatory update policy

Microsoft has announced a significant change in its update policy for Windows 11, allowing users to indefinitely pause updates. This marks the first time since Windows 10's launch in 2015 that users can pause updates indefinitely. The change aims to address user complaints about the mandatory update policy and lack of control over update timing. The new update options include delaying updates on new devices, scheduling updates, restoring normal shutdown and restart options, and providing clearer update information. Microsoft emphasizes that while users can now control updates more effectively, updates remain crucial for security and system stability.

Tom's Hardware•

6 days ago

Disaggregating LLM Inference: Inside the SambaNova Intel Heterogeneous Compute Blueprint

SambaNova Systems and Intel have introduced a blueprint for heterogeneous inference that optimizes modern large language model (LLM) workloads by utilizing specialized hardware for different phases of inference: GPUs for prefill, SambaNova RDUs for decode, and Intel Xeon 6 CPUs for agentic tools and orchestration. This approach addresses the complexity of agentic AI systems with varying compute demands. By isolating tasks onto specific hardware, the architecture improves efficiency, scalability, and cost-effectiveness. The design reflects a shift towards specialized compute fabrics and better supports the evolving landscape of AI reasoning systems.

SemiWiki•

1 week ago

Tested: Microsoft may finally end forced Windows 11 updates with a new pause feature that gives you full control

Microsoft is testing a new feature in Windows 11 that allows users to select a specific date to pause updates until then, potentially ending forced updates. This new feature replaces the existing dropdown menu for pausing updates for up to five weeks. Users may now have greater control over Windows Updates, with the ability to pause updates for as long as they want. The new pause controls are still under development, but users can now pick a date in the calendar to pause updates. This change marks a significant shift in Microsoft's approach to Windows updates.

Windows Latest•

2 weeks ago

SemiEngineering

Silent Data Corruption: A Major Reliability Challenge in Large-Scale LLM Training (TU Berlin)

Researchers at Technische Universitat Berlin published a technical paper on the challenges of Silent Data Corruption (SDC) in Large Language Model (LLM) training. As LLMs grow in size, hardware-induced faults like SDC can bypass detection mechanisms, leading to severe consequences during training. The study explores how intermittent SDC impacts LLM pretraining, highlighting the sensitivity of different factors like bit positions and kernel functions. The research proposes a lightweight detection method to identify harmful parameter updates and demonstrates the effectiveness of recomputing training steps upon detection in mitigating corruption.

SemiEngineering•