DeepSeek：可能動搖加密行業的AI模型

在周一的亞洲早上交易時間裡，人工智能（AI）令牌為紅色，AI類別的同事下跌了9％

Cryptocurrency prices remained largely lower during the morning trade in Asia on Monday, with AI tokens registering bigger losses than the broader crypto index.

週一在亞洲的早晨交易中，加密貨幣的價格在很大程度上較低，而AI令牌的損失卻大於更廣泛的加密指數。

The AI category on CoinGecko was down 9% at the last check, while the CoinDesk 20 index, which measures the price performance of the top 20 cryptocurrencies, was trading 5% lower.

上次檢查中，Coingecko上的AI類別下跌了9％，而Coindesk 20指數（衡量了前20個加密貨幣的價格績效）降低了5％。

Crypto investors, like their traditional finance counterparts, are likely digesting the impact that DeepSeek, a new AI model, will have on the industry.

加密投資者（與傳統的財務同行）一樣，可能正在消化新型AI模型DeepSeek對行業的影響。

Data from DeepSeek posted on Hugging Face, a forum of the AI industry, shows that its model outperforms OpenAI’s – all while being built on a budget of $6 million and a fraction of the Graphics Processing Units (GPUs) that OpenAI uses, which recently closed a $6.6 billion round with a valuation of over $157 billion.

來自DeepSeek的數據發佈在AI行業的一個論壇上，這表明其模型的表現要優於OpenAI的，同時又以600萬美元的預算和OpenAI使用的圖形處理單元（GPU）的一小部分來建立，而OpenAI使用了最近關閉的圖形處理單元（GPU） 66億美元的回合，估值超過1570億美元。

Perhaps most concerning to the perpetual GPU bulls is that DeepSeek is so efficient that a version of it can be run on your phone.

也許最關心永久性GPU公牛的是，DeepSeek是如此有效，以至於可以在手機上運行它的版本。

Wow DeepSeek R1 version 1.5B runs perfectly locally on my phone ?

WOW DeepSeek R1版本1.5B在我的手機上在本地完美運行?

So you can have a model that outperforms GPT-4o and Claude 3.5 Sonnet on math in your pocket.

因此，您可以擁有一個模型，可以在口袋裡的數學上優於GPT-4O和Claude 3.5十四行詩。

Mind-blowing pic.twitter.com/0J3uLXI6ZR

心靈吹捧的pic.twitter.com/0j3ulxi6zr

Naturally, some of the worst-performing AI tokens are the ones with the most exposure to GPUs.

自然，某些表現最差的AI令牌是最多的GPU。

Small-cap Nodes.AI, which facilitates access to GPUs, was down nearly 20%, according to CoinGecko data, although Aethir, which does the same (albeit at a much higher market cap) was only down 6%, just slightly more than the CoinDesk 20 benchmark.

根據Coingecko數據的數據，促進GPU的促進GPU的小型cap Nodes.ai下跌了近20％ Coindesk 20基準。

Lessons from crypto gaming While DeepSeek will make for a stressful week at OpenAI, Nvidia, and other tech giants that have pivoted to AI, it’s also a more pressing lesson for crypto projects – one that might be familiar to those who followed crypto’s foray into gaming years ago.

來自加密貨幣遊戲的經驗教訓雖然DeepSeek將在OpenAI，NVIDIA和其他與AI旋轉的科技巨頭一起度過一個壓力的一周，這對於加密項目來說，這也是一個更緊迫的課程，這是那些跟隨Crypto涉足Gaming的人熟悉的一堂課幾年前。

Despite the pools of capital available to crypto AI projects, they just haven’t been able to make something as revolutionary or interesting as what DeepSeek did.

儘管Crypto AI項目可用的資本池，但他們只是無法像DeepSeek那樣做一些革命性或有趣的東西。

Data from CoinGecko pegs the value of the crypto gaming sector (GameFi) at $19 billion. If the largest entries on this list, like Sandbox, GALA, or Decentraland, were entries on the top-30 largest gaming companies by market cap, they would hold respectable positions alongside household names responsible for recognizable franchises.

Coingecko的數據將加密遊戲領域（GameFI）的價值定為190億美元。如果此列表中最大的條目，例如沙箱，晚會或去世地，是前30名最大的遊戲公司的參賽作品，那麼他們將與負責可識別的特許經營權的家喻戶曉的名字同時擔任可觀的職位。

But despite these advantages, projects just haven’t had the same success as their traditional counterparts.

但是，儘管有這些優勢，但項目並沒有與傳統同行取得相同的成功。

Last year, blockchain gaming received the lowest investment since 2020, according to data source DappRadar. Just $1.8 billion was invested in blockchain gaming and metaverse projects, representing a 38% slide from 2023.

根據數據源Dappradar的數據，去年，區塊鏈遊戲獲得了自2020年以來最低的投資。只有18億美元投資於區塊鏈遊戲和元項目，這是2023年的38％滑梯。

And while the daily unique active wallets in blockchain gaming surged 421% last year, the sector’s dominance within the industry fluctuated between 26-29%, with DeFi taking the lead. Still, the tally is lower than games on Steam.

儘管區塊鏈遊戲中每日獨特的主動錢包去年飆升了421％，但該行業在該行業中的主導地位波動在26-29％之間，而Defi領先。儘管如此，計數仍低於Steam上的遊戲。

Besides, some older and less popular games on Steam, from companies with market caps a fraction of GameFi giants, have a larger following. Over the years, many crypto games have struggled to build a sustainable user base.

此外，一些較舊且不太受歡迎的遊戲來自具有市場上限的公司，而Gamefi巨頭的一小部分則具有更大的追隨者。多年來，許多加密遊戲一直在努力建立可持續的用戶群。

So, for now, one might say that crypto faces challenges in catering to use cases beyond finance.

因此，就目前而言，人們可能會說，加密貨幣面臨著迎合財務以外的用例的挑戰。