Random thoughts while gazing at the misty AI Frontier

2024-02-21 · Source: Elad Blog · Field: Business & Management — Corporate Strategy & Leadership, Entrepreneurship & Start-ups, Human Resources & Workforce Development · Depth: Intermediate, quick

Summary

AI's economic impact is rapidly expanding, with OpenAI and Anthropic alone approaching 0.1% of US GDP each, potentially reaching 1% of GDP run rate by late 2026. This growth coincides with a "distributed IPO" for top AI researchers, whose compensation has surged due to intense talent competition, making many "post-economic." However, the industry faces an "artificial short-term asymptote" on model capabilities due to compute limitations, particularly memory supply, which may reinforce an LLM oligopoly until 2028. Compute, or "tokens," is emerging as a new economic currency, influencing business models and even prompting non-tech companies like Allbirds to invest in GPU farms. AI's impact on jobs is initially affecting outsourced services in developing countries, while later-stage companies anticipate flattening or shrinking headcount through attrition, increasing productivity per employee. The current "Slop Age" is seen as a golden era of human-AI collaboration, where AI generates useful output that humans refine.

Key takeaway

For AI startup founders currently experiencing revenue growth, you should critically evaluate your exit strategy within the next 12-18 months. The current market, while buoyant, mirrors past tech booms where many companies failed despite early success. Considering an acquisition now could maximize your company's value before increased competition or market shifts potentially diminish opportunities. Focus on building defensibility through unique "harness" solutions and be mindful of compute as a core economic driver.

Key insights

Compute (tokens) is emerging as a fundamental economic currency in the AI industry, driving new business models and investment.

Principles

• AI product defensibility increasingly relies on the "harness" (UX, workflow) built around core models.
• AI automation prioritizes tasks with fast closed-loop learning systems and high economic value.
• The AI talent market has undergone a "distributed IPO" with significant compensation shifts for top researchers.

Method

Prioritize AI automation by identifying tasks with clear, testable feedback loops and high economic value, such as software engineering.

In practice

• Consider subsidizing compute as a user acquisition and usage model for AI products and services.
• Invest in GPU infrastructure, as compute scarcity may create an oligopoly and extend silicon lifetimes.

Topics

• AI Economics
• Compute Infrastructure
• AI Talent Market
• AI Business Models
• Workforce Transformation
• AI Regulation

Best for: CTO, VP of Engineering/Data, Executive, Director of AI/ML, Investor, Entrepreneur

Related on AIssential

• The Week AI Grew Up — AI is transitioning from an experimental phase to critical global infrastructure, driven by token demand and business model shifts.
• The AI market feels incredibly acquisitive. Should that change how I pick a company? — In an acquisitive AI market, prioritize founder quality, team caliber, company defensibility, and personal learning over acquisition likelihood.
• Why $1B Exits are Dead — AI is accelerating company scale and exit values, forcing venture capital to rethink traditional investment assumptions.
• Golden Ages — Golden ages in industries are finite, ending when incumbents achieve durability through network effects, scale, brand, or deep customer embedding.
• Venture Capital Is Concentrating Faster Than Ever. What Happens To Everyone Else? — Venture capital is increasingly concentrating in large rounds and top-tier companies, particularly in AI.
• The Subsidized AGI Economy — AI labs intentionally subsidize power users to acquire critical agentic workload data and secure future market position.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Elad Blog.