I think CoT may be very valuable in helping capabilities researchers understand their models’ shortcomings and design training curricula that address those deficiencies. This sort of falls under “business value” itself. I currently think this is one of the stronger arguments for why we won’t switch to neuralese.
(The post only vaguely gestures at this argument, so your interpretation is also fair, but I think business value and capabilities gains are pretty closely related rather than being opposing factors.)
I think CoT may be very valuable in helping capabilities researchers understand their models’ shortcomings and design training curricula that address those deficiencies. This sort of falls under “business value” itself. I currently think this is one of the stronger arguments for why we won’t switch to neuralese.
(The post only vaguely gestures at this argument, so your interpretation is also fair, but I think business value and capabilities gains are pretty closely related rather than being opposing factors.)