As the corporate fad of ‘tokenmaxxing’ reaches its limits, workplaces are seeking cheaper artificial intelligence options. The trend, which involved maximizing the use of AI-generated work, has led to rising costs without a corresponding increase in productivity.
Tokenmaxxing Fades
The term ‘tokenmaxxing’ refers to the practice of maximizing the use of tokens, the building blocks of generative AI. However, this approach has proven to be expensive, with companies realizing that the costs of using AI products from leading developers like OpenAI and Anthropic are adding up quickly.
According to Vincent Gusdorf, head of AI analytics at Moody’s Ratings, ‘It’s very easy to create something you don’t need with AI.’ Gusdorf recommends a more disciplined approach to using AI, rather than simply maximizing token usage.
Some companies are now looking for alternative AI solutions, including open-source models from Chinese startups like Moonshot and Zhipu. These models offer similar capabilities to top US models at a fraction of the cost.
As the industry shifts away from tokenmaxxing, companies are focusing on finding more efficient and cost-effective ways to use AI. This includes using ‘model routing,’ which involves sending easier queries to cheaper AI systems and reserving more complex tasks for more powerful models.
Original reporting: KTBS 3 (Shreveport) — read the source article.
