Thursday☕️

Time Is Money

Jul 25, 2024 • 5 min read

Trending:

Ethereum Spot ETF
Meta’s New AI Model

Markets:

Yesterday’s U.S. stock market:

Today’s commodity market:

Today’s crypto market:

Ethereum Spot ETF:

On Tuesday, the first-ever spot Ethereum exchange-traded funds (ETFs) began trading in the U.S. after receiving final approval from the U.S. Securities and Exchange Commission (SEC). This significant milestone provides institutional and retail investors with regulated financial products for direct exposure to Ethereum, further embedding the cryptocurrency into traditional financial markets.
The SEC's approval, following a thorough review, marks a growing acceptance of digital assets in regulated markets. Notable financial institutions like Grayscale Investments, Fidelity Investments, Invesco, VanEck, Franklin Templeton, 21Shares, Bitwise, and BlackRock's iShares have launched their Ethereum ETFs. These ETFs are now trading on major exchanges, including NYSE Arca and the Chicago Board Options Exchange (CBOE), highlighting strong institutional support.
Spot Ethereum ETFs are unique compared to futures-based ETFs as they directly track the price of Ethereum, offering a straightforward investment method. Unlike futures contracts, which involve agreements to buy or sell assets at a future date, spot ETFs provide immediate exposure to Ethereum’s price movements. This enables investors to access the cryptocurrency’s market performance without the complexities associated with futures trading.

Meta’s New AI Model:

Meta has recently launched Llama 3.1, a new series of artificial intelligence models. This release includes Llama 3.1 405B, which is the largest and most capable open-source AI in the world. Llama, which stands for Large Language Model Meta AI, is designed to perform a wide range of tasks, from language translation to complex problem-solving. The latest version introduces significant improvements over its predecessors, such as a larger context length of 128K, support for eight languages, and enhanced reasoning abilities.
One major difference between Llama 3.1 and other models is its open-source nature. Unlike closed-source models, which restrict access to their underlying code and architecture, open-source models like Llama 3.1 are publicly available. This openness allows developers worldwide to access, modify, and improve the code of the model, fostering a collaborative environment that can drive innovation and progress. Open-source models enable a broader range of applications and customization, allowing developers to tailor the AI to specific needs and integrate it into diverse systems.
Architecturally, Llama 3.1 uses a standard decoder-only transformer model with significant optimizations to handle the massive scale of the 405B parameter model. The training of this model involved over 16,000 Nvidia H100 GPUs and extensive fine-tuning procedures to ensure top-tier performance. Such an architecture not only improves the model's capabilities but also sets a new benchmark for training large-scale AI systems. By making such a powerful model accessible to the public, Meta is democratizing advanced AI technology. This move is expected to create rapid advancements in AI research and development, as it allows for greater experimentation and iteration from developers and researchers.

Statistic: