Oct 25, 2026 · 32 min
Listen from 11:08
Listen at 11:08
Claude Duet puts personal AI workflows under scrutiny
How I'm Actually Using AI Right Now (Including Claude Duet)
As AI agents move from chat interfaces into everyday work, the episode tests whether they are genuinely useful across research, coding, strategy, content, and analysis.
- 1Claude Duet provides a lens for examining how AI currently supports strategy, research, coding, content creation, and data analysis.
- 2The episode compares Claude Duet with ChatGPT Dots while questioning whether proactive personal agents are ready for everyday workflows.
- 3Nathaniel Whittemore treats a new AI product as an opportunity to audit practical habits rather than simply celebrate its capabilities.
Don't miss
Whittemore turns Claude Duet’s introduction into an audit of how AI already supports his work across several domains.
The brief
Nathaniel Whittemore introduces Claude Duet, then uses the personal agent as a prompt to examine how AI fits into his existing work.
The discussion spans strategy, research, coding, content creation, and data analysis, keeping the focus on practical workflows rather than abstract capability.
Claude Duet is considered alongside ChatGPT Dots as Whittemore asks whether proactive personal agents are becoming useful additions to everyday work.
The episode’s central tension is simple: new AI agents may promise initiative, but their value depends on how well they fit real habits and tasks.
What was said on this episode
30 statements · 17 positive · 9 negative · 1 mixed · 3 neutral
Personal-agent form factors remain experimental despite early adoption.
“I believe pretty firmly that we have to treat the form factor of these personal agents as still experimental.”
Listen at 1:31
Personal agents are seeing early traction, especially Muse.
“we are absolutely seeing early traction and uptake, especially with Muse.”
Listen at 1:40
Opus 5.5 is Nathaniel’s preferred model for strategic externalization.
“Opus 5.5 is my favorite model that I've used in a very long time for this.”
Listen at 4:39
Astra and Opus are used about equally for research.
“for this use case, I actually use Astra and Opus About equally.”
Listen at 5:29
Cross-review between Astra and Opus adds confidence to research.
“Astra is reviewing Opus and Opus is reviewing Astra to add that additional layer of confidence”
Listen at 5:48
Nathaniel retains his initial title about half the time.
“about 50% of the time I stick with exactly what I had”
Listen at 7:19
GPT image models now dominate Nathaniel’s image-generation workflow.
“At this point, image generation has pretty much just become the GPT image models.”
Listen at 8:12
Midjourney can outperform alternatives for certain artistic image-generation uses.
“Midjourney is just its own weird beast with a lot of strange taste and artistry that, depending on the particular use case, can sometimes just be way better.”
Listen at 8:34
Astra inside Codex provides Nathaniel’s best data-analysis results.
“For me, it provides the best results and has that sort of concision of thought that I think is really valuable for this type of use case.”
Listen at 9:24
Claude Code with Opus 5.5 is Nathaniel’s primary building environment.
“for me, where things have landed pretty consistently is ClaudeCode. with Opus 5.5 being the default model”
Listen at 9:54
Fable substantially reduced the effort required to complete projects.
“Fable just brought that completion energy way down and matched the activation energy.”
Listen at 11:08
Claude models have become highly capable at video editing.
“the models in Claude have gotten enormously good at video editing.”
Listen at 13:34
Proactive AI workflows have not justified their setup and maintenance costs.
“up until now, for me personally, the juice wasn't worth the squeeze.”
Listen at 19:47
The chief-of-staff agent did not provide enough value to justify maintenance.
“it candidly just wasn't adding enough value to deal with the maintenance and upkeep.”
Listen at 20:28
Nathaniel is unlikely to use personal agents for personal tasks.
“I was very unlikely to use these sort of agents for personal uses.”
Listen at 21:39
Explaining flight-booking constraints to an agent is no faster than using the airline interface.
“it would take absolutely as much time to explain to an agent the variables and contingencies as it does to just look at their specific web interface.”
Listen at 22:11
The monothread paradigm centralizes context and retrieves it as needed.
“you instead have a single thread that has all of your context, and pulls up what it needs to do when it needs to do it”
Listen at 23:15
AI labs have improved context compression and continuity across conversations.
“the labs have all gotten much better on the back end about compressing the context and carrying it forward”
Listen at 24:05
Nathaniel currently advises against investing heavily in Google’s AI ecosystem.
“Google I still, at the moment, I'm sorry to say, I don't think it makes sense to invest a lot in.”
Listen at 25:34
Google’s position in personal agents may improve in the future.
“I am optimistic that it will change as they get back up on the horse.”
Listen at 25:41
Solo users should favor nimble, fast-moving, non-enterprise-focused AI tools.
“Being totally solo, I'm always going to bias towards the much more nimble, fast moving versions that aren't heavily enterprise focused.”
Listen at 25:57
The dot is currently weak at strategic engagement because its personality is flat.
“the dot's personality at this point is so flat that it almost stonewalls that sort of strategic engagement.”
Listen at 27:12
The dot is not particularly proactive.
“My dot I have found not to be particularly proactive”
Listen at 27:46
The dot could be valuable as a centralized context holder.
“it does seem like it could be valuable as a master holder of context.”
Listen at 28:17
Duet may overestimate the value of proactive tasks.
“Duet is so biased towards proactivity that I found that it maybe overestimates how valuable that set of work is”
Listen at 28:59
Successful email drafting increases Nathaniel’s optimism about Duet’s proactive value.
“This makes me more optimistic for the value of Duet for this proactivity.”
Listen at 30:07
Nathaniel is optimistic about Duet’s task-oriented approach.
“I'm already much more optimistic about this approach.”
Listen at 30:47
Duet is substantially worse than Claude’s native interface for visual design.
“doing it inside Duet was much worse than doing it in Claude's native design experience”
Listen at 30:56
Successful personal-agent features will spread across competing ecosystems quickly.
“within weeks if not days, it's going to become part of all of them.”
Listen at 31:26
Personal agents currently function more as delegated helpers than primary AI interfaces.
“It is much less a big shift towards it being the main way I interact with AI, and much more an easy-to-set-up little baby version of myself that does a bunch of stuff that I don't want to do.”
Listen at 31:51
Statements are attributed to the speaker as said on the episode and reflect their view at the time, not PodLume's. They are not advice.
Books & mentions
Some links are affiliate links — PodLume may earn a commission if you buy.
Weekly email
State of the Pod, every Monday
The week's most-discussed names and stories from podcasts, with the moments behind them. Free, five minutes, unsubscribe in one click.