Last Week in AI #333 - ChatGPT Ads, Zhipu+Huawei, Drama at Thinking Machines

2024-03-11 · Source: Last Week in AI · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Novice, medium

Summary

OpenAI is introducing labeled banner ads in ChatGPT for free-tier and $8/month ChatGPT Go users in the U.S. and other markets. These ads will appear as distinct sections at the bottom of relevant answers, with sensitive categories like health and politics excluded. OpenAI assures that ads will not influence model outputs and user conversations will not be shared with advertisers, offering privacy controls such as ad dismissal and personalization toggles. Concurrently, the AI industry sees significant developments: Thinking Machines Lab faces turmoil with key founders and employees departing for OpenAI and Meta, while Zhipu AI successfully trained its GLM-Image model entirely on a Huawei stack, reducing reliance on US chips. Sequoia Capital is reportedly investing in Anthropic's new funding round, breaking traditional VC norms by backing a direct competitor to its existing portfolio companies like OpenAI and xAI.

Key takeaway

For entrepreneurs developing AI products, you should consider diverse monetization strategies beyond subscriptions, such as ad-supported tiers, while prioritizing robust privacy controls. The intense competition for AI talent, as seen with Thinking Machines Lab, underscores the need for strong retention strategies and a clear vision. Additionally, exploring non-US hardware stacks for model training could offer strategic advantages and reduce supply chain risks.

Key insights

The AI industry is rapidly evolving with new monetization strategies, intense talent competition, and efforts to reduce hardware dependency.

Principles

• Monetization through advertising is a viable strategy for free-tier AI services.
• Talent acquisition and retention are critical challenges for AI startups.
• Diversifying hardware supply chains enhances national AI autonomy.

Method

Zhipu AI trained its GLM-Image model using a full Huawei stack, including Ascend Atlas 800T A2 servers and the MindSpore ML framework, demonstrating a domestic training pipeline.

In practice

• Implement clear ad labeling and user privacy controls in AI products.
• Monitor competitor talent movements for strategic hiring opportunities.
• Explore alternative hardware ecosystems for AI model training.

Topics

• OpenAI Monetization
• AI Geopolitics
• AI Startup Ecosystem
• Large Language Models
• AI Infrastructure

Best for: Entrepreneur, Tech Journalist, Investor, AI Product Manager

Related on AIssential

• This Week in AI: Your Recap — AI platforms are expanding monetization via ads and enhancing enterprise capabilities for agent deployment and cross-platform access.
• Can open-source beat OpenAI? — Open-source AI models foster global collaboration and efficiency, challenging closed-source dominance through accessible innovation and diverse monetization.
• The Startups That Win Will Ship Faster and Smarter — Accelerated product development and market shifts necessitate a renewed focus on core business fundamentals.
• AI App Crisis, OpenAI Does Math, Big Nvidia Deal — AI apps monetize quickly but struggle with long-term retention due to over-promising and intense competition.
• ChinAI #358: Around the Horn (25th episode) — China's AI ecosystem is dynamic, marked by rapid startup valuations, evolving business models, and significant talent and regulatory challenges.
• It's War: Ads vs No Ads in AI — AI companies are exploring ad-based monetization, potentially altering user experience and competitive dynamics.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Last Week in AI.