The Rise of Chain-of-Thought AI: A Tale of Two Trains
Consider a train leaving Chicago traveling west at seventy miles an hour, and another train leaving San Francisco traveling east at eighty miles per hour. Can you figure out when and where they’ll meet?
Introducing Chain-of-Thought AI
The idea behind chain-of-thought processing is that the AI model can detail the sequence of calculations it performs in pursuit of the final answer, ultimately achieving "explainable" AI. Such explainable AI could conceivably give humans greater confidence in AI’s predictions by disclosing the basis for an answer.
A Tale of Two Chatbots
To explore the matter, I put OpenAI’s o1 against R1-Lite, the newest model from China-based startup DeepSeek. R1-Lite goes further than o1 to give verbose statements of the chain of thought, which contrasts o1’s rather terse style.
The Results
I submitted the famous trains math question to both R1-Lite and o1 preview. Both chatbots start out with fairly simple routes to a solution to the famous trains problem of grade-school math. Both models came up with similar answers, though the o1 model was noticeably faster, taking five seconds to spit out an answer, while DeepSeek’s R1-Lite took 21 seconds.
The Chain of Thought
When I asked both models to compute roughly where the two trains would meet, meaning what U.S. town or city, the o1 model quickly produced Cheyenne, Wyoming. In the process, o1 telegraphed its chain of thought by briefly flashing short messages such as "Analyzing the trains’ journey," or "Mapping the journey," or "Determining meeting point."
The Verbose Approach
In contrast, the DeepSeek R1-Lite spent nearly a minute in its chain of thought, and, as in other cases, it was highly verbose, leaving a trail of "thought" descriptions totaling 2,200 words. These became increasingly convoluted as the model proceeded through the chain.
Conclusion
The test goes to show that in these early days of chain-of-thought reasoning, humans who work with chatbots are likely to end up confused even if they ultimately get an acceptable answer from the AI model. While OpenAI’s o1 model wraps up its work fairly quickly, DeepSeek’s R1-Lite, on the other hand, goes through a long and winding "thought" process that becomes increasingly complicated and distracting.
Frequently Asked Questions
Q: What is chain-of-thought AI?
A: Chain-of-thought AI is an approach in generative AI that details the sequence of calculations it performs in pursuit of the final answer, ultimately achieving "explainable" AI.
Q: What is the difference between OpenAI’s o1 and DeepSeek’s R1-Lite?
A: OpenAI’s o1 model is more concise and provides a brief chain of thought, while DeepSeek’s R1-Lite is more verbose and provides a longer chain of thought.
Q: Why is the chain of thought important?
A: The chain of thought is important because it allows humans to understand how the AI model arrived at its answer, giving them greater confidence in the prediction.
Q: Is chain-of-thought AI a new concept?
A: No, chain-of-thought AI is an extension of existing AI approaches, but it is a new way of presenting the information to humans.
