10 Comments

All I know is that the next time I neglect to pause and paraphrase, I now have a go-to second response. “I’m sorry, excuse me while I back off this verbal rake I’ve step on.”

Expand full comment

Good point: I find that simply pointing out how awkward a conversation is getting can actually save that same conversation (some of the time).

Expand full comment

This needed more Dink.

Expand full comment

Alley, you have logged into Rudy's device again.

Expand full comment

Perhaps an interesting parallel here with AI. DeepSeek freaked a lot of tech companies out for cost and capability reasons but the really big breakthrough from a UX or human interaction perspective was adding stepwise reasoning. Now you can see what the AI was ‘thinking,’ how it came to its brilliant (or stupid/incorrect) conclusion. Just like human interactions, the more we understand thought process and intent, the more we trust.

Expand full comment

Oh yes, I was absolutely thinking about R1 and "reasoning" models like this. When I talked about "show your work", that wasn't all I was thinking about, but I kept coming back to this.

Expand full comment

Reasoning was a good advance, but DeepSeek nailed the show-your-work element. Seems so simple right? No one likes a black box

Expand full comment

Perplexity has had a similar idea (pre-reasoning models, they let you see more behind-the-scenes stuff w/links, etc). Very good early adopters.

Expand full comment

Alternatives:

"Stop and Switch It Up"

"Cease and Convert"

"Halt and Hypothesize"

On second thought, maybe yours works?

Expand full comment

"Acquit and Make Sure the Glove Fits" is my final offer.

Expand full comment