Silicon Investor (SI) -- The First Internet Community

STOCKTALK

We've detected that you're using an ad content blocking browser plug-in or feature. Ads provide a critical source of revenue to the continued operation of Silicon Investor. We ask that you disable ad blocking while on Silicon Investor in the best interests of our community. If you are not using an ad blocker but are still receiving this message, make sure your browser's tracking protection is set to the 'standard' level.

Technology Stocks : Qualcomm Moderated Thread - please read rules before posting -- Ignore unavailable to you. Want to Upgrade?

To: engineer who wrote (195765)	9/25/2025 3:34:30 PM
From: pheilman_	4 Recommendations Recommended By Dr. John Lance Bredvold RoseCampion sbfm Respond to of 196724

Engineer is pointing out that Nvidia did a very clever move a few decades ago, they made a programming language or at least a library for their GPUs. Programmers are efficient and don't want to learn the ins and outs of each new chip that comes along, so providing a set of powerful functions that hide the complexity of the chip is wise. It gets people using the chips and makes it very sticky once they have working code. There were many GPU vendors at the time and Nvidia distinguished themselves with these libraries.

To: engineer who wrote (195765)	9/25/2025 3:38:19 PM
From: waitwatchwander	2 Recommendations Recommended By Dr. John Lance Bredvold Read Replies (1) \| Respond to of 196724

Did you read what I posted at the link? I posted this article before and also got little response. I take it as the methods to be used to incorporate AI/ML methods into future architectures (wireless or otherwise). It is obviously directed towards wireless but does have connotations nuanced towards what you see in CUDA (ie updating and tailoring sensing models). It's not always what's obvious that matters. Python is the scripting language du jour. I'd place more money in that direction than in C code. fwiw I hear what you are saying but a new CUDA like system for model management and implementation requires dominance of AI/ML standards. Jensen created a defacto training standard by being first in the brute force & dense foliage field. Going forward, it isn't going to be so nescient. As always, it will be built upon what works best for the majority and derived from accepted standards. Yes, it will be nice seeing Qualcomm contributing more than their hardware to these efforts. It's a huge field and no single entity is likely to dominate as did Jensen with training.