In this webinar I walked through the AI developments that defined 2024 and outlined where I see 2025 heading.
Large language models continued to dominate. Despite AI being a much broader field, LLMs and ChatGPT became near-synonymous with "AI" in business contexts as companies recognized their practical utility.
OpenAI remained the most influential player, shipping GPT-4O and Sora for video generation. Notably, OpenAI does not necessarily build the best models, but it dominates the discourse through marketing reach and broad availability.
Many companies introduced AI assistants in 2024. These tools support employees on tasks from email drafting to coding, though many organizations left their staff without adequate training for the new tooling.
Costs for AI services dropped by roughly a factor of 1,000 compared to prior years. Tasks that cost 60 dollars in 2023 cost a few cents in 2024, unlocking use cases that were previously uneconomical.
The open-source community made strong progress. Models like Mixtral approached the quality of proprietary systems at a fraction of the cost, threatening the incumbents' moat.
AI systems will increasingly interact directly with computer interfaces, operating a desktop the way a human would and opening new automation paths.
A new generation of reasoning models will spend more computation on deliberation, handling more complex problems at higher per-query cost.
The industry is shifting from passive assistants to autonomous agents that pursue goals independently and escalate to humans only when needed.
Enterprises will begin redesigning processes from scratch around AI agents rather than bolting automation onto existing workflows.
Chinese AI companies will enter global competition more aggressively; their open-source models already show strong performance at lower cost.
The AGI debate will intensify as various companies claim to have reached the milestone, though the practical impact will likely be less dramatic than predicted.
The major trends in 2024 were the continued dominance of language models, OpenAI's market leadership, the rise of AI assistants, a dramatic price drop in AI services, and significant progress in the open-source community.
For 2025, expectations include improved computer interaction through AI, new reasoning models, the rise of autonomous agents, deeper integration into business processes, increased competition from China, and intensified AGI debates.
AI service costs dropped by roughly a factor of 1,000 in 2024. What cost 60 dollars in 2023 became available for a few cents, opening entirely new application possibilities.
Open-source models like Mixtral now approach the quality of proprietary systems at a fraction of the cost. This challenges the monopoly of large tech companies and democratizes access to AI technologies.
Businesses should develop a clear AI strategy, invest in employee training, start with simple assistants and plan the transition to autonomous agents, rethink business processes, and monitor open-source developments.
Enterprises are moving from simply providing AI tools to deeply integrating AI into their business processes. The trend is toward autonomous agents that function as virtual employees and independently handle tasks.
Chinese AI companies are expected to enter global competition more strongly in 2025. Their open-source models already show impressive performance at lower costs and could significantly influence the global AI landscape.
The AGI debate is expected to intensify in 2025, with various companies claiming to have reached this milestone. However, AGI will likely have less dramatic impact than originally expected.
.
Copyright 2026 - Joel P. Barmettler