The Fragmentation of Search

Also: Ray Kurzweil Still Says He Will Merge With A.I.

Author

AI KATANA
July 04, 2024

Morning!

In today’s newsletter, we explore the significant transformations in the search market, driven by LLM-based search agents, and their implications for consumers and web publishers. We also review the limitations of a new “ethical” AI music generator, discuss strategies for mitigating AI’s growing environmental footprint, and highlight Samsung’s impressive profit surge due to AI-driven semiconductor demand. Additionally, we delve into Ray Kurzweil’s enduring vision of merging with AI, WhatsApp’s development of a personalized AI avatar generator, and Redactive AI’s successful funding round to enhance enterprise AI security.

Sliced:

  • 🔎 The Fragmentation of Search

  • 🎹 The New ‘Ethical’ AI Music Generator Can’t Write a Halfway Decent Song

  • 🌱 How Companies Can Mitigate AI’s Growing Environmental Footprint

  • 📈 Samsung flags better-than-expected profit rise as AI boom lifts chip prices

  • 🤖 Ray Kurzweil Still Says He Will Merge With A.I.

The search market is undergoing a significant transformation with the advent of horizontal and LLM-based search agents like ChatGPT and Perplexity, challenging the dominance of traditional search engines like Google. This shift is creating new opportunities for both consumers and web publishers, who have often felt marginalized by the traditional search model. Before ChatGPT, Google dominated the search landscape, but now, LLM-based agents provide direct answers, bypassing traditional web publishers and reshaping how information is accessed. This fragmentation raises questions about the future of business models for search agents, the willingness of consumers to pay for search, the survival of web publishers without traditional SEO and ad revenues, and the potential impact on content creation if publishers aren’t compensated. Future search platforms may reward content creators through new monetization models, including crypto tokens and social currencies, and could offer personalized, immersive search experiences using advanced technologies. The evolving landscape suggests exciting possibilities for new ways of accessing and interacting with information.

🎹 The New ‘Ethical’ AI Music Generator Can’t Write a Halfway Decent Song

The new AI music generator Jen, marketed as an “ethical” alternative trained on licensed material, has struggled to produce impressive results. Despite its ethical training approach, musicians testing the tool found its outputs uninspiring and lacking in creativity, often comparing the generated music to generic, cliché-ridden compositions. Professional musicians noted that Jen failed to capture specific genres or artists’ styles, producing music that felt artificial and devoid of warmth. Although Jen is still in its alpha phase with plans for expansion, its current limitations raise concerns about the potential impact of AI on the music industry, including the proliferation of low-quality stock music and the challenge of maintaining creative authenticity in an increasingly AI-driven landscape.

The growing environmental impact of AI, driven by increased energy and water consumption and heightened carbon emissions, necessitates immediate action to ensure sustainable practices. As computing power dedicated to AI is expected to increase tenfold by 2026, companies must adopt strategies to mitigate this footprint. Key recommendations include using foundation models instead of building new ones, choosing appropriately sized models, and utilizing hybrid cloud approaches to optimize energy use. Efficient infrastructure, particularly AI-specific processors, and open-source collaboration can also significantly reduce environmental impact. By implementing these measures, businesses can harness the benefits of AI while minimizing its ecological costs.

Samsung Electronics reported a more than 15-fold increase in its second-quarter operating profit, reaching 10.4 trillion won ($7.54 billion), driven by rising semiconductor prices amidst the AI boom. This surge is attributed to increased demand for high-end DRAM and NAND flash memory chips, essential for AI chipsets and data center servers. The company’s profits exceeded expectations, reflecting both higher chip prices and the reversal of previous inventory writedowns. However, while the current quarter benefited from AI-driven demand, future growth may hinge on the performance of legacy chips and new high-bandwidth memory (HBM) approvals for Nvidia.

Ray Kurzweil, the renowned inventor and futurist, continues to advocate for the concept of the Singularity, predicting that humans will merge with AI by 2045, achieving indefinite lifespans. Now 76, Kurzweil’s latest book, “The Singularity Is Nearer,” reiterates his belief that advances in AI and nanotechnology will allow humans to overcome aging within the next decade. While Kurzweil’s predictions were once met with skepticism, the rapid development of AI technologies like ChatGPT has lent some credence to his visions. Despite this, critics argue that such techno-optimism overlooks practical and ethical challenges, including resource limitations and societal impacts. Kurzweil remains optimistic, envisioning a future where technological advancements extend life beyond natural aging processes, although the feasibility and desirability of such a future remain subjects of debate among experts.

🛠️ AI tools updates

WhatsApp is developing a new generative AI feature called “Imagine” that allows users to create personalized avatars in various settings, using user-supplied images and text prompts combined with Meta’s AI Llama model. This feature, currently in the WhatsApp Beta for Android, lets users generate avatars by taking a single photo, which the AI uses to produce images in different scenarios. Users can invoke this by typing “Imagine me” followed by a description. The feature, optional and requiring manual activation, also allows for deletion of reference images. The release date for this feature remains unspecified, as WhatsApp continues rolling out its Meta AI chatbot and other AI tools cautiously.

💵 Venture Capital updates

Redactive AI, a Melbourne-based developer platform, has secured $11.5 million in seed funding led by Blackbird and Atlassian Ventures, with additional participation from US firms Felicis and Zapier. The company addresses the shortage of AI engineering and security skills in enterprise software teams by developing a responsible AI platform that allows corporations to leverage generative AI without compromising proprietary information. Founded by former Atlassian product managers Andrew Pankevicius and Alexander Valente, and AI engineer Lucas Sargent, Redactive AI aims to meet the security, privacy, and permission control needs of enterprises. The platform enables software engineers to build AI-enabled features that pass security reviews and can be quickly deployed. Early adopters include an Australian superannuation fund and a financial exchange, highlighting the demand for secure AI solutions in regulated industries.

🫡 Meme of the day

⭐️ Generative AI image of the day