• Future Perfect
  • Posts
  • #123. šŸ¤– OpenAI's AI chip šŸ’ AI jewelry creation šŸ’» 25% of Google code is AI

#123. šŸ¤– OpenAI's AI chip šŸ’ AI jewelry creation šŸ’» 25% of Google code is AI

Plus, AI tools: Interview insights at survey scale; AI-powered productivity assistant for your day; Transform training into engaging, personalized learning paths.

šŸŽ™ļø Check out my new Marc Hoag AI Law Pod, a podcast with AI ā€œguestsā€ courtesy of Google's NotebookLM. Now that NotebookLM allows customization, the AI ā€œguestsā€ actually interact with me at times, and deep dive quite a bit beyond the bounds of each issueā€™s content. If nothing else, itā€™s a very cool alternative way to enjoy Future Perfect while driving.

  • Strategic Shift to Custom Chips: OpenAI is collaborating with Broadcom and TSMC to create its first custom AI chip to support its AI infrastructure, aiming to reduce reliance on third-party suppliers.

  • Scaling Back on Manufacturing Ambitions: Originally considering building a network of chip foundries, OpenAI has decided to hold off due to high costs and resource requirements, instead focusing on designing custom chips with the support of partners.

  • Broad Supply Chain Approach: Alongside its custom chip project, OpenAI is diversifying chip suppliers by adding AMD processors to its current Nvidia-dominated stack, addressing shortages and cost issues in the GPU market.

  • Broadcom Partnership in Focus: OpenAI has secured manufacturing capacity with TSMC through Broadcom, targeting a 2026 launch for its custom-designed chip, which is expected to focus on inferenceā€”a crucial component for AI operations as demand grows.

  • Strategic Talent and Technology Investments: OpenAI has assembled a specialized chip team led by former Google TPU engineers to guide its chip design efforts, carefully balancing its relationship with Nvidia, the dominant GPU supplier, to ensure access to the latest Blackwell chips.

  • AMD Chips Enter the Market: OpenAIā€™s partnership with AMD through Microsoftā€™s Azure platform aims to leverage AMDā€™s MI300X chips, challenging Nvidiaā€™s dominance and positioning AMD for an anticipated $4.5 billion in AI chip sales by 2024.

  • Financial Pressure Drives Innovation: Faced with high operational costs, OpenAI projects a $5 billion loss on $3.7 billion in revenue for 2024, spurring efforts to diversify suppliers and optimize compute costs across hardware, electricity, and cloud services.

  • Market Impact and Sector Implications: OpenAIā€™s move signals a broader trend among tech giants like Microsoft, Meta, and Google to pursue in-house or alternative chip solutions, potentially reshaping the competitive landscape in AI and tech hardware.

  • User-Driven Design Platform: Arcade AI, founded by Mariam Naficy, offers a marketplace where usersā€”dubbed "Dreamers"ā€”can design custom jewelry through generative AI, selecting from materials like gold, silver, and gemstones including diamonds and rubies.

  • How It Works: Users input design ideas or upload images, which are then translated into various jewelry concepts. Once a design is selected, itā€™s assigned to an artisan who crafts the piece. Prices vary widely, starting around $100 and potentially exceeding $1,000 based on materials and design complexity.

  • Customization and Quality: Arcadeā€™s generator uses models like Stable Diffusion and Midjourney. Users can refine designs through editing tools, though some designs, like multi-charm necklaces, can challenge the AI's capabilities.

  • Monetization for Creators: Designers can earn a 2.5% commission on sales and will soon be able to set their own prices. To become a seller, users need at least 100 followers and sales of three items or $500 in revenue.

  • Intellectual Property Concerns: Arcadeā€™s model ensures originality by limiting exact replication of uploaded designs, addressing potential concerns among jewelry designers about AI copying. Arcadeā€™s IP control measures stem from Naficyā€™s experience managing artist rights at her previous venture, Minted.

  • Future Growth and Funding: Arcade has raised $17 million from investors like Ashton Kutcher and LinkedIn co-founder Reid Hoffman. Plans include expanding into categories like apparel and accessories and introducing social features and competitions, with the goal of partnering with major brands in the future.

  • A Potential Industry Shift: While some may fear AIā€™s impact on traditional jewelry design, Arcade's artisans have embraced the platform as a way to engage directly with customers, turning AI-powered design into a collaborative opportunity.

  • AIā€™s Central Role in Code Creation: Google CEO Sundar Pichai revealed that AI now generates more than 25% of all new code at Google, which engineers then review and integrateā€”a significant milestone underscoring AI's importance to Googleā€™s operations.

  • Revenue Growth Driven by AI Products: Alphabet reported $88.3 billion in revenue for Q3 2024. Google Services, including Search, brought in $76.5 billion (up 13% year-over-year), while Google Cloud, home to Googleā€™s AI infrastructure, reached $11.4 billion (up 35%).

  • Strong Operating Incomes: Google Servicesā€™ operating income increased to $30.9 billion, up from $23.9 billion last year. Google Cloud also saw a major jump, with income rising from $270 million to $1.95 billion.

  • AI Expands Product Capabilities: Googleā€™s new AI-driven features include customizable chatbots, Gemini-powered ā€œGemsā€ for Google Meet, and YouTube tools for creators. AI tools are also integrated into the Pixel 9 smartphone lineup and Google Search, where they expand search possibilities and enhance user experience.

  • YouTube Revenue Milestone: YouTube achieved a major milestone with total ad and subscription revenue surpassing $50 billion over the past four quarters.

  • Legal Challenges Looming: Despite strong performance, Google faces legal scrutiny after an August ruling labeling it a monopolist in search and advertising. The case, led by the U.S. Department of Justice, is now in the remedies phase, where potential outcomes include a Google breakup.

āœØ PowerPromptā„¢

This is a prompt to help voters make informed decisions. While it was inspired by me ā€” an American living in California ā€” and the example below makes use of a link to download a PDF of voting information, it should work anywhere.

If youā€™d like to test this California/US-based information, head over to https://vig.cdn.sos.ca.gov/2024/general/pdf/complete-vig.pdf to download the official General Election information booklet.

Head over to your favorite chatbot ā€” I used ChatGPT-4o and look forward to trying Claude 3.5-sonnet ā€” and upload the PDF.

Based on what you know about me, help me figure out how to make an informed voting decision for each item in the attached PDF. Iā€™ve already made up my preliminary mind, but Iā€™m curious what you come up with. Ask me questions about each one to help you help me. Let's go through each item one at a time, one by one. Do not proceed until I say so.

Then, once youā€™ve finished, you can deep dive into which candidate you should (or shouldnā€™t) vote for. Really get the AI to challenge you. Itā€™s a fascinating bit of introspective study.

šŸ§° AI Tools & Resources

  • Aptitude AI: Let our AI conduct and analyze conversations with hundreds of customers, delivering the depth of an interview at the scale of a survey.

  • timeOS: An AI productivity companion that captures and summarizes your day, organizes all relevant information within the right tool, and proactively surfaces the knowledge you need, when you need it.

  • inncivio: Keep your team motivated and engaged as inncivio transforms your training material into fun, bite-sized knowledge, completed with personalized learning paths that ensure easy upskilling and retention!

šŸŽ‰ THATā€™S ALL FOR TODAY!

See you next time! šŸ‘‹

What'd you think of today's email?

This helps me improve it for you!

Login or Subscribe to participate in polls.

Looking for past newsletters? You can find them all here.

Reply

or to participate.