# Anthropic goes public, ChatGPT gets a wallet & AI designs new drugs

## Executive summary

The discussion covers three major AI trends: the financial and technological implications of Anthropic's IPO and the AI chip market; the rise of agentic commerce via tools like ChatGPT Wallet; and the advancement of drug discovery, highlighted by Isomorphic Labs' $5 billion funding round. A key theme is the increasing reliance on sophisticated AI models to optimize complex systems, whether in hardware design or biological simulation.

## Key takeaways

- AI Optimization and Hardware Diversity: AI is proving highly effective at optimizing complex systems, such as managing diverse hardware types. The market is moving toward diversification, with companies like Amazon, Google, and Anthropic utilizing different chip flavors (Nvidia, AMD, custom silicon) to meet compute demands.
- Agentic Commerce and Security: OpenAI introduced ChatGPT Wallet, allowing AI agents (like Claude) to complete purchases. Security relies on cryptographic components, specifically the use of one-time card numbers and requiring human approval (e.g., Face ID) for each transaction.
- AI Accelerates Drug Discovery: Isomorphic Labs, a spin-out of Google DeepMind, raised $5 billion to advance AI drug design. AI accelerates the process by pruning vast search spaces and improving the confidence of candidates, which reduces the risk in later human trials.

## Technical details

- AI Model Capabilities: AI is used not just for language inference but for simulating complex environments and creating 'digital twins' of biological systems, which helps predict properties like solubility and toxicity before wet lab testing.
- Security Protocols in Agentic Commerce: The ChatGPT Wallet architecture utilizes a 'one-time card number' for each purchase, a security best practice conceptually similar to just-in-time credentialing, limiting the scope of potential damage.
- Biological Simulation and Control: IBM Research mentioned a model called MALformer, which focuses on 'controlled generation' in drug discovery. This means generating molecules that are not just effective, but also soluble and non-toxic, based on physical and chemical properties.

## Practical implications

- The integration of AI into the supply chain and hardware design requires greater diversity in chip solutions (e.g., Nvidia, AMD, custom silicon) to manage massive compute demands.
- Agentic commerce introduces new consumer trust issues, requiring robust security measures like limited-scope, one-time credentials.
- AI is significantly reducing the time and cost of drug development by shifting the bottleneck from the wet lab to the simulation phase, leading to higher success rates in clinical trials.

## Topics

AI, Artificial Intelligence, Drug Discovery, Agentic Commerce, Hardware/Chip Market, Mixture of Experts, Anthropic, Isomorphic Labs

Source: https://www.youtube.com/watch?v=Y6npte_lEKU
