News for hackers. Stuff that matters.
Re:DeepSeek v4 by reactordev on Friday April 24, 2026 @09:46AM (#47890240)
I think you missed the bigger picture here. It’s that China has their own stack now, soon others will follow. It’s not about putting up the highest numbers, it’s about putting up the highest ROI. To them, this is it. Qwen too but being able to compete with today’s models means they are closer to competing with tomorrow’s.
Re:DeepSeek v4 by scottyah on Friday April 24, 2026 @01:39PM (#47893433)
At this scale, it's purely quality. The better the model, the faster the advancements. If using a model half as smart as the best made us half as productive, people would pretty much all be using the current quantized models that can run on a decent laptop. The difference between Opus xHigh and Gemma4 is very different (at least in my job).
Re:DeepSeek v4 by reactordev on Friday April 24, 2026 @02:37PM (#47894170)
YMMV and you underestimate their ability to run scale. They are not running on laptops.