OpenAI Furious DeepSeek Might Have Stolen All the Data OpenAI Stole From Us

[ourkaos@lemmy.today]

"Now" is always a good time to rewatch it & get more out of it!

[sgt_choke_n_stroke@lemmy.world]

[asafum@feddit.nl]

If you're running it on your own system it isn't connected to their server or sharing any data. You download the model and run it on your own hardware.

From the thread I was reading people tracked packets outgoing and it seemed to just be coming from the chatbot program as analytics, not anything going to deepseek.

[just_another_person@lemmy.world]

[teft@lemmy.world]

I hear tulip bulbs are a good investment...

[just_another_person@lemmy.world]

It's a true comedy that still holds up. I honestly thought for years that Mel Brooks had something to do with it, but he didn't. It's so well crafted that there are many layers to it that you can't even grasp when watching as a child. Seeing it as an adult just open your eyes to how amazingly well done it was.

I could do without the whole Billy Crystalizing of large portions of it though.

[humble_pete_digger@lemm.ee]

Thank you China.
No for real - it's either EU or frigging china that helps us with these oligarch overlords

[chemicalwonka@discuss.tchncs.de]

[ulrich@feddit.org]

That's why "value" is in quotes. It's not that it didn't exist, is just that it's purely speculative.

Hell Nvidia's stock plummeted as well, which makes no sense at all, considering Deepseek needs the same hardware as ChatGPT.

Stock investing is just gambling on whatever is public opinion, which is notoriously difficult because people are largely dumb and irrational.

[nightwatch_admin@feddit.nl]

That is true, and running locally is better in that respect. My point was more that privacy was hardly ever an issue until suddenly now.

[asafum@feddit.nl]

Absolutely! I was just expanding on what you said for others who come across the thread

[ramblingpanda@lemmynsfw.com]

How much for two thousands?

[nielsbohron@lemmy.world]

I'm an AI/comp-sci novice, so forgive me if this is a dumb question, but does running the program locally allow you to better control the information that it trains on? I'm a college chemistry instructor that has to write lots of curriculum, assingments and lab protocols; if i ran deepseeks locally and fed it all my chemistry textbooks and previous syllabi and assignments, would I get better results when asking it to write a lab procedure? And could I then train it to cite specific sources when it does so?

[a_a@lemmy.world]

[cygnus@lemmy.ca]

Hell Nvidia’s stock plummeted as well, which makes no sense at all, considering Deepseek needs the same hardware as ChatGPT.

Common wisdom said that these models need CUDA to run properly, and DeepSeek doesn't.

[petter1@lemm.ee]

EU is in best way to become a group of dictators having billionaires in their asses as well..

https://www.politico.eu/article/mapped-europe-far-right-government-power-politics-eu-italy-finalnd-hungary-parties-elections-polling/

[0x0@programming.dev]

Good that 404 are unafraid of tackling issues, but tbh i find the "hahaha" unprofessional and dispense with the informal tone in news.

[ulrich@feddit.org]

Sure but Nvidia still makes the GPUs needed to run them. And AMD is not really competitive in the commercial GPU market.

[ryper@lemmy.ca]

AMD apparently has the 7900 XTX outperforming the 4090 in Deepseek.

[ulrich@feddit.org]

Those aren't commercial GPUs though. These are:

https://developer.nvidia.com/blog/introducing-hgx-a100-most-powerful-accelerated-server-platform-for-ai-hpc/