Waivio

Recommended Posts

Deepseek

25 comments

edicted62.3 K9 days ago4 min read

https://files.peakd.com/file/peakd-hive/edicted/23u5X6jFEBXGcuJKP4DqsimBomweX6V45CT4F5kf3upHdYPTiUhYFYHdSvLRfVa7ZNMck.png

https://files.peakd.com/file/peakd-hive/edicted/EopyDe6psF3APb9XtbKLjTJFLN9F8DjihF1s7UwqmehPLXPMuHKrme4CjgTxaPHFniQ.png

On Jan. 22, President Trump announced the $500 billion "Stargate" AI infrastructure initiative, which included investments from OpenAI, tech company Oracle, and SoftBank.

LoL

Almost immediately after Trump announces that the government is basically going to take over AI and centralize it in hugely problematic ways... China decides, "You know what? Fuck that," and completely open-sources Deepseek AI. Love to see it. China has been proving they are not nearly as bad as our propaganda makes them out to be.

According to the DeepSeek research paper, the project took only $6 million to train, and it performs on par with leading AI models.

In fact I'm hearing rumors that it performs BETTER than leading AI models, which, considering the budget, is hilarious and even embarrassing for Silicon Valley. These tech bros were already having a hard time monetizing AI and scraping every last dollar out of it possible... and now a potentially better product hits the market that anyone can clone without permission. Epic.

… dispelled notions that [China] would only produce closed-source AI, and has eroded Silicon Valley's perceived advantages over global competitors

DeepSeek is evidence that OpenAI does not have an unbeatable moat and that assumptions about scaling, value accrual in AI models, and development costs were also dispelled by the development

AKA: get wrecked noobs

According to a paper authored by the company, DeepSeek-R1 beats the industry’s leading models like OpenAI o1 on several math and reasoning benchmarks. In fact, on many metrics that matter—capability, cost, openness—DeepSeek is giving Western AI giants a run for their money.

DeepSeek proves that there’s another way to win: by revamping the foundational structure of AI models and using limited resources more efficiently.

Innovation Born out of a Crisis

In October 2022, the US government started putting together export controls that severely restricted Chinese AI companies from accessing cutting-edge chips like Nvidia’s H100. The move presented a problem for DeepSeek. The firm had started out with a stockpile of 10,000 H100’s, but it needed more to compete with firms like OpenAI and Meta. “The problem we are facing has never been funding, but the export control on advanced chips,” Liang told 36Kr in a second interview in 2024.

So we were basically sanctioning China so they couldn't compete... which forced them to use an open-source collaborative strategy to create better software. Kinda sounds a little crypto-coded if I'm being honest. I don't know what's up with the current simulation we are living in, but it seems like the underdogs keep pulling out these unlikely victories. We need to be paying very close attention to this type of stuff.

DeepSeek, founded by hedge fund manager Liang Wenfeng, released its R1 model on Monday, explaining in a detailed paper how to build a large language model on a bootstrapped budget that can automatically learn and improve itself without human supervision.

This is also a big deal because LLMs have a way of "hallucinating" where eventually the AI just goes completely off the rails and has to be reset to a previous state. It appears as though a system has been created to automate this process rather than needing to do it manually and freeze the AI in place.

DeepSeek’s R1 release sparked a frenzied debate in Silicon Valley about whether better resourced US AI companies, including Meta and Anthropic, can defend their technical edge.

Industry insiders say DeepSeek’s singular focus on research makes it a dangerous competitor because it is willing to share its breakthroughs rather than protect them for commercial gains. DeepSeek has not raised money from outside funds or made significant moves to monetise its models.

Now imagine what happens when crypto fuses with this technology and it can fund and scale itself to infinity. Obviously this is easier said than done but I consider this story a proof-of-concept and prototype that further demonstrates how open-source collaboration is always going to surpass closed systems if a positive feedback loop exists within the underlying economic system.

Conclusion

It's always a good sign when David casts a stone at Goliath and shakes the foundation of the very ground we are standing on. Humanity is entering a new epoch in which open-source collaboration is learning to overtake their greedy exploitative counterparts. I only hope that products like Deepseek will join us in the continued fight for freedom and decentralization of power.

Comments

Sort byBest
AI
Waivio AI Assistant
How can I help you today?