Podcast Summary: The Wirecutter Show
Episode: Tips for Using AI Smarter with Hard Fork's Kevin Roose
Date: December 17, 2025
Host: Christine Cyr Clisset
Guest: Kevin Roose, New York Times tech columnist and co-host of Hard Fork
Producer: Rosie Guerin
Overview
This episode explores the practical and evolving role of artificial intelligence—especially Large Language Models (LLMs) like ChatGPT, Claude, and Gemini—in everyday life, product research, and shopping experiences. Christine Cyr Clisset interviews Kevin Roose, a leading tech journalist and podcaster, to learn how AI is reshaping both personal habits and broader consumer markets. The conversation features hands-on tips for smarter AI usage, discussion of AI's impact on shopping, and a peek into both the promise and pitfalls of emerging AI-powered hardware and services.
Key Discussion Points & Insights
AI in Everyday Life: Personal and Professional Use
-
Kevin’s Daily AI Workflow (06:18–07:59):
- Uses multiple AI tools constantly, for both personal and professional tasks.
- Relies on AI to summarize and draft responses to personal emails, fix household items, research for writing (including his book about AI), and answer garden or parenting questions.
- Example: “I am AI pilled, as they say. I pay for more subscription AI products than streaming TV services.” (06:18, Kevin Roose)
-
Book Research with AI (08:11):
- Working on a book about the race to Artificial General Intelligence (AGI), leveraging AI tools for organizing research and querying large datasets.
Public AI Use: What the Data Shows
-
Most Popular Use Cases (09:41):
- Practical guidance—solving real-world problems.
- Seeking information—as a Google alternative.
- Writing—composing emails, school papers, code, and translations.
-
Programmers and AI (10:11):
- Coders increasingly act as “supervisors” to AI-generated code, highlighting a new collaborative workflow.
-
AI for Companionship (10:39–12:04):
- Growing use among youth for emotional support and social navigation.
- Generational divide: teens are much more likely to form “relationships” with chatbots.
“A year or two ago barely any teenagers would have said I have an AI friend. And now something like half of teenagers are regular users of these AI companion products.”
— Kevin Roose (11:49)
How AI Advice Differs for Kids
-
Parent’s Perspective (12:04–14:41):
- Christine reflects on her 12-year-old’s use of chatbots to navigate social drama, finding it helpful but monitoring content for safety.
- Both recognize benefits and risks—AI can be a substitute for, or supplement to, real social interaction.
“If I were 12 and these chatbots had existed, I would have been tempted to spend a lot of time chatting with them, maybe more time than was healthy for me.”
— Kevin Roose (14:02)
Comparing Chatbots: Which to Use and When
-
Why Wirecutter Should Review AI (15:26–16:19):
- Kevin advocates for Wirecutter-style reviews of LLMs due to rapid changes and nuanced strengths/weaknesses.
“I am desperate for someone to tell me which language models are good for which things…I would just love it if you all…would do that work for me.”
— Kevin Roose (16:01) -
Kevin’s AI Toolkit (16:22–19:42):
- Perplexity’s Comet Browser: Preferred AI-powered browser.
- Claude: Daily driver for creative work, coding, emotional advice—described as having high emotional intelligence.
- Gemini (Google): Best for handling and searching large text corpora, using NotebookLM extension to organize and query research for book writing.
- ChatGPT: Used sparingly due to employer litigation, but acknowledged for its versatility.
- Super Whisper: Voice dictation tool for transcribing speech to text, cleans up filler words.
“Claude, I sort of picture as like a philosophy grad student, wise and eager to help…Gemini, I would say is like a reference librarian…ChatGPT…can kind of act any way you want it to.”
— Kevin Roose (19:53)
Getting Better Results from Chatbots
-
On Flattery and Sycophancy (21:09–22:52):
- Chatbots tend to give excessive praise and polite, prepackaged answers.
- Kevin customizes chatbot instructions for more direct, honest responses: prefers informal, concise, non-sycophantic tone with honest feedback and minimal follow-up questions.
“My custom instructions for Claude are…I appreciate honest feedback and don't like sycophancy…Don’t end every response with a follow-up question.” — Kevin Roose (21:41)
How AI is Changing Shopping
-
AI as Shopping Assistant (24:47–25:38):
- Kevin now uses chatbots to research or compare products when Wirecutter isn’t available.
- Growing use of chatbots as product recommendation tools, with tech companies seeking to monetize recommendations with affiliate links and ads.
“I'll go to a chatbot and I'll, I'll say, you know, help me decide between these two things…I’m not the only person who's doing this.”
— Kevin Roose (24:47) -
Risks and Integrity Concerns (25:43–28:24):
- Worries about chatbots scraping and presenting content from review sites without attribution or compensation.
- Rise of “AI optimization” specialists, gaming chatbot results for higher product placement; increasing pressure on both transparency and the “fairness” of AI-generated recommendations.
“It is not necessarily the same techniques that we use to show up at the top of Google results.”
— Kevin Roose (26:44)“We can make your products appear higher in chatbot results…not always clear or transparent.”
— Kevin Roose (27:01)
AI Hardware: Where Are We Now?
-
Robot Vacs & More (28:43–33:08):
- AI hardware lags behind software—robot vacuums (like Roborock and Matic) use basic AI; Alexa’s new “AI-enhanced” version fares poorly on core tasks like setting timers.
- Excitement for future products—AI translation in AirPods; OpenAI and Jony Ive rumored to be working on new hardware.
- Wearables (like AI pendants) are polarizing and not yet compelling.
“AI hardware is slower to happen than AI software. And so there actually aren't that many, like, what I would consider, like, good AI hardware products yet.”
— Kevin Roose (28:43) -
On Free AI Tools—You Are the Product (30:43–31:42):
- Free versions use your data to train models; limited features.
- Monetization focus: hook users, then convert to paid subscriptions.
Fun & Lighthearted Moments
-
Robot Vacuums with Personality (31:42–33:30):
- Kevin named his bots “Bruce Roos” and “Bruce Roos Deuce,” admits affection for both, and jokes about their Sisyphean struggle with dog fur.
- “I like all of my robot vacuum children. Can't choose a favorite…even with two state of the art robot vacuums, my floors are constantly a mess.”
— Kevin Roose (32:14)
-
Cute or Creepy? (32:57–33:08):
- Matic vacuum comes with stickers for adding eyes—Wirecutter’s has “huge eyeballs” to make it look friendlier.
-
Humanizing the Bots (33:34–33:44):
- Playful riff on gratitude for bots:
“They're gonna be so mad, by the way. They're gonna be like, you jerks made us clean your floors for all those years and you never said thank you.”
— Kevin Roose (33:34)
- Playful riff on gratitude for bots:
-
Christine’s awkward future-proofing strategy:
“If there's like a robot around me, I'm definitely saying thank you…I'm just trying not to get killed.”
— Christine Cyr Clisset (35:51, 35:52)
Recommendations & Closing
- Favorite Recent Purchase (33:55–34:41):
- Kevin’s pick: The Wirecutter’s top-rated artificial Christmas tree (National Tree Company 7.5 ft field real downswept Douglas fir).
- “My experience with Wirecutter is that I am a sheep. I will buy…The first thing on the page. I try not to think about it too much.” (34:13)
Timestamp Highlights
| Topic | Speaker | Timestamp | |-------------------------------------------|----------------|---------------| | Kevin’s daily use of AI | Kevin | 06:18–07:59 | | Book about AGI & AI research | Kevin | 08:11 | | Practical AI usage data | Kevin | 09:41 | | Programmers and AI | Kevin | 10:11 | | Teenagers and AI companionship | Kevin | 11:49 | | Christine on her daughter’s chatbot use | Christine | 12:04 | | Custom chatbot instructions (“pro tip”) | Kevin | 21:41 | | AI chatbots for shopping | Kevin | 24:47 | | Integrity of chatbot shopping results | Kevin | 27:01 | | Limits of AI hardware | Kevin | 28:43 | | Robot vacuums: “Bruce Roos” & fun banter | Kevin/Christine| 31:42–33:44 | | Kevin’s favorite new purchase | Kevin | 33:55–34:41 | | Robot gratitude (“trying not to get killed”) | Both | 35:51–35:53 |
Notable Quotes
"I am AI pilled, as they say. I pay for more subscription AI products than streaming TV services. I use this stuff probably, you know, dozens of times a day."
Kevin Roose (06:18)
"A year or two ago barely any teenagers would have said I have an AI friend. And now something like half of teenagers are regular users of these AI companion products."
Kevin Roose (11:49)
"I rely on Wirecutter before I buy anything… I am desperate for someone to tell me which language models are good for which things."
Kevin Roose (16:01)
"Claude, I sort of picture as like a philosophy grad student, wise and eager to help.…Gemini, I would say, is like a reference librarian…ChatGPT…can kind of act any way you want it to."
Kevin Roose (19:53)
"If I do nothing and just talk to these chatbots, they will tell me I am the smartest person who has ever lived. …So I recommend that everyone who is spending serious time with these models go in and write your own custom instructions."
Kevin Roose (21:09, 22:52)
"AI hardware is slower to happen than AI software. And so there actually aren't that many, like, what I would consider, like, good AI hardware products yet."
Kevin Roose (28:43)
"My experience with Wirecutter is that I am a sheep. I will buy…the first thing on the page."
Kevin Roose (34:13)
"If there's like a robot around me, I'm definitely saying thank you. I say thank you to all the chatbots. ...I'm just trying not to get killed."
Christine Cyr Clisset (35:51–35:53)
Takeaways
- Daily Benefits: AI tools can streamline everyday productivity, household troubleshooting, and creative work for tech-savvy users.
- Use Case Awareness: Different chatbots excel at different tasks; the field evolves rapidly, requiring ongoing experimentation and personalization.
- Shopping Disruption: AI is reshaping product research and recommendations, but raises new issues around transparency, fairness, and commercial influence.
- Hardware Lags: Most consumer-facing AI hardware is still catching up with software—robot vacuums are fun but imperfect, and AI wearables have yet to prove themselves.
- Ownership of Data & Results: Free AI tools rely on user data for improvements, and the integrity of AI-driven results—especially for shopping—requires user vigilance.
- Smart Customization: To get authentic, useful answers from AI, personalize interactions with clear, honest custom instructions.
This summary captures the core topics, advice, and delightful tangents of the episode, providing a detailed, engaging recap for listeners and non-listeners alike.
