My bets on open models, mid-2026

2023-11-24 · Source: Interconnects AI · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Intermediate, medium

Summary

An analysis of the open versus closed AI model landscape reveals a complex dynamic, challenging the notion that open models will fully catch up to closed labs in every area. Despite strong technical capabilities from open model labs in maintaining pace on established benchmarks, closed models like those from U.S. labs tend to exhibit greater robustness and general utility, particularly for knowledge worker assistance. Chinese open-weight labs, while technically proficient and focused on benchmarks, may face funding difficulties later this year, impacting their capability trajectories within 3-9 months. The article highlights that the long-term balance of capabilities is more an economic question than a technical one, with real AI revenue driving investment and continuous model improvement. Distribution and online Reinforcement Learning (RL) are emerging as key factors where closed labs can dominate, especially in real-world use cases involving user feedback.

Key takeaway

For CTOs and AI Architects evaluating model strategies, recognize that while open models offer cost efficiency for repetitive automation, closed models currently provide superior robustness for complex, user-driven tasks. Your investment decisions should factor in the economic sustainability of open-source initiatives, especially for Chinese labs, and anticipate increased regulatory pressure on open models. Prioritize solutions that align with long-term revenue generation to ensure sustained model improvement and competitive advantage.

Key insights

The open vs. closed AI model race is primarily an economic contest, not solely a technical capability one.

Principles

• Closed models often surpass open models in robustness and general utility.
• Economic staying power dictates long-term success in the open vs. closed model race.
• Regulation on open models is difficult to implement effectively across sovereign entities.

Method

The analysis distills key beliefs on open models, monitoring the capability gap intertwined with funding, distillation techniques, regulation, and user adoption dynamics.

In practice

• Open models are suitable for repetitive automation tasks and backend processes.
• Closed models excel as direct assistants for knowledge workers.
• Local agents and personal agents represent an overlooked market for open model usage.

Topics

• Open Models
• Closed Models
• AI Capability Gap
• AI Benchmarks
• AI Economics & Funding

Best for: CTO, VP of Engineering/Data, AI Architect, Director of AI/ML, AI Product Manager, Investor

Related on AIssential

• Open and closed models are on different exponentials — Open and closed AI models follow divergent economic exponentials, creating distinct market structures and value capture mechanisms.
• China catches up — The "no moat" dynamic, driven by inefficient and replicable AI, leads to price wars and unsustainable profits.
• Leadership Challenges at OpenAI Revealed — AI is achieving human-level performance in complex tasks while economic and regulatory challenges emerge.
• The Open vs. Closed AI Race — Open vs. closed AI is a strategic lever, not an ideological choice, for competitive positioning.
• 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 Sequence Radar #837: Last Week in AI: From Model Releases to Market Structure — AI is evolving into an infrastructure play, prioritizing scale, integration, and open models over raw product demos.

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by Interconnects AI.