Claude in Excel is really impressive. So its weird that using Microsoft's own Excel agent using Claude 4.5 usually yields weaker answers, It seems to happen because the Microsoft Excel agent relies o…


LinkedIn Content Strategy & Writing Style
Associate Professor at The Wharton School. Author of Co-Intelligence
5 people tracking this creator on Viral Brain
Ethan Mollick positions himself as the premier academic practitioner of the generative AI era, bridging the gap between rigorous research and hands-on experimentation. His content strategy centers on "stress-testing" the frontier of LLM capabilities, moving beyond hype to provide a nuanced value proposition of practical utility for knowledge workers. What makes him notable is his refusal to remain a detached observer; he operates at the intersection of pedagogical theory and technical prototyping, often building his own tools to demonstrate how AI reshapes human expertise and organizational behavior. By blending dry academic skepticism with a playful curiosity about "little computer people," he offers a sophisticated roadmap for co-intelligence that is both deeply technical and accessible.
381.3K
3.1K
585
—
14.6
19
14
Claude in Excel is really impressive. So its weird that using Microsoft's own Excel agent using Claude 4.5 usually yields weaker answers, It seems to happen because the Microsoft Excel agent relies o…

As jobs begin to shift with AI, I think there will be increasing numbers of people feeling like Aditya (former CTO Dropbox, early Facebook engineer, etc). I think figuring out what comes next is chal…

Every few months, I write an updated, idiosyncratic guide on which AIs to use right now. My new version has the most changes ever, since AI is no longer just about chatbots. To use AI you need to und…
Moltbook seems to have "broken through" to a wider non-AI audience. That makes it a good chance to explain the present: this time, Moltbook was mostly roleplaying by people and agents, very little of…
As a business school professor, its striking that a lot of programmers, as they increasingly delegate authority to coding agents, are re-encountering the basic problems that underlie management theory…
AI agents have rather suddenly gotten good enough at long horizon tasks that I think it represents inflection point in the impact of AI at work. Agreement on this very rapid trend from METR, GDPval &…

14.6 posts/week
Posts / Week
0.5 days
Days Between Posts
14
Total Posts Analyzed
HIGH
Posting Frequency
584.7%
Avg Engagement Rate
STABLE
Performance Trend
95
Avg Length (Words)
HIGH
Depth Level
ADVANCED
Expertise Level
0.82/10
Uniqueness Score
NO
Question Usage
0.3%
Response Rate
Writing style breakdown
<start of post>
The transition from tools that follow instructions to agents that exercise judgment is the defining shift of this year. We are seeing this play out in how models handle "taste" & subjective quality, rather than just raw logic.
In a recent paper, researchers found that even smaller models can be trained on citation data to predict which scientific breakthroughs will have the most impact. This suggests that quality isn't just a human intuition we can't replicate, but a pattern of signals—citations, shares, & peer sentiment—that can be ingested & understood. When I use GPT-5.4 Pro for my own research, I am not just looking for a summary; I am looking for the model to tell me what matters.
What is important is that we don't just let the models decide for us. We need to be deliberate about the signals we use to train them, or we risk creating a feedback loop of mediocrity. (And we all know how much the internet loves to reward mediocrity.)
Paper: it https://lnkd.in/eEX2BwUJ
<end of post>
Sign in to unlock the full writing analysis
Nail your LinkedIn strategy with ViralBrain.
Analyze and write in Ethan Mollick's style. Grow your LinkedIn to the next level.