Grokking the Anthropic Forward Deployed Engineer Interview
0% completed
Extended Thinking and Effort
Should a model think longer before every answer? A common first answer is yes, because a careful answer seems better than a fast one. But reasoning is paid for in output tokens and in waiting time.
At the accounting firm, the answer was yes for one sub-agent and no for another. Anomaly review found many more real problems when it reasoned more. Invoice matching gave the same results at low and high effort and only became slower and more expensive.
.....
.....
.....
Like the course? Get enrolled and start learning!
Reading Progress
0%