Bloomberg Tech Live in San Francisco | Bloomberg Technology

Arm

• Arm's CEO, Rene Haas, discusses the company's record revenue and growth, expressing confidence in continued growth in the coming years.
• Haas emphasizes the significance of AI infrastructure buildout and the long-term potential of AI in various sectors such as healthcare, research, and productivity.
• Arm sees strong growth in smartphone royalties, particularly with its version nine architecture used in premium mobile phones.
• The rapid pace of AI software innovation is driving demand for faster chip designs, benefiting Arm's licensing activity.
• Arm is not directly affected by the recent U.S. policies limiting Huawei's access to chip design and technology, as they did not apply for any licenses.
• Haas expresses a preference for multiple suppliers in the AI PC market rather than relying on a single dominant player.
• Haas predicts that within the next two to three years, the Windows ecosystem will offer multiple players, price points, and devices with improved performance.
• Haas highlights the geopolitical considerations surrounding AI, including the need for balancing training data locations, sovereign clouds, and data weights.
• Despite concerns about AI valuation exuberance, Haas believes the AI opportunity is still undervalued and that the current level of innovation and invention in the field is not indicative of a hype cycle.

Hugging Face

• Hugging Face's CEO, Clement Delangue, asserts that open-source models have surpassed closed-source models in most use cases and emphasizes the importance of distributing power and reducing the concentration of control over AI technology.
• Delangue acknowledges the challenge of funding AI model training, but suggests that smaller companies can still participate by leveraging pre-trained models and using smaller datasets, democratizing AI development.
• Hugging Face is close to being profitable and has over 10,000 paying customers and 4 million AI builders using its platform.
• The company is making acquisitions and expects to announce them in the coming weeks.
• Hugging Face sees more mergers and acquisitions in the AI industry as many companies took risky bets and are running out of money.
• The company is hiring in San Francisco due to the city's concentration of technology and AI talent.
• Hugging Face believes in the importance of open-source and open research for the advancement of AI.
• Hugging Face has introduced vision as a capability into its platform and launched in 32 languages with high quality, surpassing human benchmarks.
• The company is working on large reasoning models, which can write software, and aims to move from work substitution to real work reinvention and orchestration using AI.
• The CEO believes that large language models will become commoditized, but there is still a significant opportunity for startups to serve the enterprise market.
• Hugging Face has made a small acquisition and plans to make more, recognizing the high barrier to entry for serving the enterprise.
• The CEO emphasizes the importance of transparency and auditability for enterprises when using AI models.
• The company is focused on remaining independent and has a long product journey to realize its vision.
• The CEO highlights the need for executives to have a vision for how AI can be used within their companies in a way that brings people along.

Sequoia

• Stephanie Zhan, a partner at Sequoia, discusses the current state of the AI industry and predicts that 2024 will be a monumental year for AI commercialization at scale.
• Sequoia has invested in 10 new AI companies in the first four months of 2023, focusing on new foundation models and AI-native applications.
• Despite the high costs associated with AI development, startups can compete with incumbents by leveraging high-quality data, reinforcement learning with human feedback, and open-source model ecosystems.
• Stephanie highlights the potential of Meta (formerly Facebook) as a leading AI company due to its significant capital investment and vast amount of proprietary data.
• Meta's development of a new generative video foundation model, similar to OpenAI's DALL-E 2, is seen as game-changing and has the potential to lower the cost of entry for startups in the AI ecosystem.

Amazon

• Udit Madan, President of Worldwide Operations at Amazon, discusses the company's focus on innovation, safety, customer experience, and efficiency.
• Amazon has 4,000 different locations around the world and is investing heavily in automation and robotics, especially with the advancements in AI.
• The company is developing novel foundation models using the high-quality data it has gathered to ship products, which will make robotics solutions more generalized and efficient.
• Amazon has been working on inventive robotics solutions that are reaching maturity and scale and will be rolled out this year.
• Despite the focus on automation and robotics, Amazon has created hundreds of thousands of jobs, including skilled and technical positions, and has launched programs to train employees for the future workforce.
• Madan's personal goals for the year include improving safety and continuing to enhance convenience and delivery speeds for customers.

X AI

• X AI, an AI company founded by Elon Musk, is closing a mega funding round, with reports suggesting it could be up to $6 billion.
• Investors in X AI's funding round are not yet known, but it is speculated that corporate entities like NVIDIA or Google could be involved.

Generative AI

• May Habib, CEO of a generative AI company, expresses excitement about the increased investment in the field and discusses strategies for standing out and retaining clients.