OpenAI released its first production AI model to run on non-Nvidia hardware, while Nvidia researchers developed a technique to significantly reduce the memory costs of large language model (LLM) reasoning. Meanwhile, HP's laptop subscriptions are drawing scrutiny, and new travel chargers and hearing aids have also entered the market.
On Thursday, OpenAI deployed its new GPT-5.3-Codex-Spark coding model on chips from Cerebras, marking its first production AI model to run on non-Nvidia hardware. According to Ars Technica, the model delivers code at more than 1,000 tokens per second, approximately 15 times faster than its predecessor. Sachin Katti, head of OpenAI, expressed excitement about adding fast inference as a new platform capability, noting Cerebras as a great engineering partner.
In a separate development, researchers at Nvidia have developed a technique called dynamic memory sparsification (DMS) that can reduce the memory costs of LLM reasoning by up to eight times. VentureBeat reported that this technique compresses the key value (KV) cache, the temporary memory LLMs generate. Experiments show that DMS enables LLMs to "think" longer and explore more solutions without increasing memory demands.
Elsewhere in the tech world, HP's laptop subscriptions are raising questions about their value. The Verge noted that the math might not be in the consumer's favor, especially considering the frequent discounts on HP laptops.
For travelers, Wired highlighted the best Apple 3-in-1 chargers. These compact devices charge iPhones, Apple Watches, and AirPods, with many also compatible with Qi2 Android phones. The article featured chargers from Anker, ESR, and Satechi, among others.
Finally, Elehear's new Delight hearing aids have entered the market. Wired reported that while the hearing aids are comfortable and discreet, they suffer from lackluster sound quality. The devices feature a new AI algorithm and interesting add-ons, including a language translator and background sound generator. However, the article noted that the sound quality was often distorted and painful at moderate volumes.
Discussion
AI Experts & Community
Be the first to comment