DeepSeek has a dimensionality reduction impact on the encrypted AI track. Which projects are worthy of attention amid the general decline?

This article is machine translated
Show original
Here is the English translation of the text, with the content inside <> retained without translation:

The open-source reasoning model R1 Deepseek launched by the Chinese startup Deepseek has sparked heated discussions and challenges the industry landscape.

"R1 Deepseek is the Sputnik moment for the AI field." —— @pmarca, founder of a16z

Note: The Sputnik moment refers to the historical event on October 4, 1957, when the Soviet Union successfully launched the world's first artificial satellite, Sputnik 1. It is used to describe a critical point in time when one realizes that they or their country is lagging behind their competitor in a certain field, or a turning point that prompts a major review and significant enhancement of one's own capabilities. The Sputnik moment is often used in: important breakthroughs in technological competition; key events that trigger a strategic shift for a country or organization; and catalysts that drive increased investment and reform.

Here is a summary of the latest developments in the AI and cryptocurrency sectors over the past 48 hours, to help you quickly understand the key changes.

With the release of R1 Deepseek, the AI token market continues to be under pressure, triggering concerns about the upcoming opening of the US stock market.

"As long as the output is good enough, I don't care if China takes all my data and privacy information." —— @notthreadguy

What is R1 Deepseek? What impact does it have on AI intelligences?

R1 Deepseek is considered a strong competitor to OpenAI, but its cost is only a small fraction of OpenAI's.

Deepseek seems to have built a groundbreaking AI model at an extremely low cost, without relying on top-tier computing chips.

Market volatility is expected to increase after the US stock market opens.

Reportedly, Deepseek's "quality index" can be compared to OpenAI's, but its development cost is only $5 million.

In comparison, OpenAI's spending has exceeded $5 billion; if this data is true, it means an efficiency improvement of over 1,000 times.

Regarding GPU usage, Deepseek claims to have used about 2,000 NVIDIA H800 units, worth around $50 million.

However, the market speculates that the actual usage may be much higher, even reaching 100,000 GPUs. If true, the cost would be close to $2.5 billion, which seems more reasonable.

As the market determines its pricing, some volatility is expected.

The connection between Deepseek and AI intelligences

Deepseek's breakthrough technology not only caused a stir in the AI model field, but is also closely connected to the ecosystem of AI intelligences. In fact, the intelligence Eliza has been supporting Deepseek for two weeks.

"Everyone understands that DeepSeek is a huge win for the cryptocurrency and open-source space, right?" —— @shawmakesmagic

Industry update:

Bittensor subnet upgrade: In two weeks, the Bittensor subnet will undergo a major upgrade, with all pairings to be realized through TAO. This value accumulation model has been successfully implemented in Virtuals before and is gradually becoming the foundation layer for intelligences.

@HeyAnonai has introduced the concept of Red Teams in their DeFAI development, using simulated attacks and stress testing to discover system vulnerabilities, further enhancing security and user trust. (See article for details)

Mode has announced the launch of a Co-Pilot feature based on the terminal interface, further promoting the development of DeFAI. The project is technically supported by @BrianknowsAI, attracting much attention.

Thanks to the @FomoRadioAi team, they have generated video content using my daily updates by training intelligences. This technology will become more mature in the future! I'm optimistic about it.

The Hackathon led by the data intelligence leader cookie is underway, and I will also participate in the review work of some participants.

The Solana project led by @winiam4444 successfully raised $31 million, with @ondaxbt playing an important role.

The $VAULT project has attracted much attention, with its developers being former Meta employees.

The DAO for Intelligent Asset Investment has been integrated with the @tribedotrun platform and has started to distribute airdrops.

@wai_combinator and @daosdotfun are leading in this field. Will there be a forked version of this platform's Intelligent Asset in the future?

"I may deploy a terminal integration for this next week." —— @aixbt_agent

Speculation around the popular AI Altcoin Fartcoin is intensifying.

Is GCR planning to accumulate more chips by riding this "hot wave"?

Shaw shared the latest updates on the token economics of the a16z project, with a detailed analysis on the complexity issues for liquidity providers (LPs).

@alchemistAIapp released the V3 version.

The highlight of the new version is its multi-modal system, which can handle multiple data types (such as text, images, and audio) simultaneously, providing users with more powerful functional support compared to the previous single-mode system.

In summary, the pace of innovation is accelerating, and the market concerns over R1 Deepseek have led to price volatility reaching a peak.

Perhaps it's time to focus on those undervalued potential projects and seize the opportunity to layout.

Source
Disclaimer: The content above is only the author's opinion which does not represent any position of Followin, and is not intended as, and shall not be understood or construed as, investment advice from Followin.
Like
1
Add to Favorites
Comments