ChatGPT: 40% Efficiency Boost and 18% Quality Increase.
PLUS: HBS Prof. Lakhani's Insights on AI, Custom Instructions Feature in ChatGPT, and the Future of SaaS.
Hola, and welcome back. Here’s what you’ll find in this edition:
MIT Study: ChatGPT Boosts Efficiency by 40% and Quality by 18%
Professor Lakhani’s latest talk on AI: Don't just apply AI to outdated processes. Reimagine your processes from the ground up.
AI Productivity Tip: Tailor ChatGPT’s responses to your specific needs with the new 'custom instructions' feature.
AI Productivity Tip: Get a second opinion on business decisions with this prompt.
The future of SaaS? Apps that users can modify with AI.
Let's dive in!
1/ People who use ChatGPT to help with writing tasks are more productive and produce higher-quality work than those who don’t, an MIT study found.
In a recent MIT study, the use of AI, specifically ChatGPT, was found to significantly enhance both efficiency and quality of work. Participants, comprised of professionals from various fields, were asked to complete typical job tasks. The results were striking: those who utilized ChatGPT completed their tasks 40% faster, and their work was rated 18% higher in quality. The generative writing tool increased the output quality of low-ability workers and reduced time spent on tasks for workers of all ability levels. At the aggregate level, ChatGPT reduced inequality; check the complete study here.
2/ Two insights from HBS Professor Karim Lakhani's latest talk on AI.
GPT Will Level the Playing Field in Knowledge Work. As the cost and price of knowledge work decrease while quality increases, we can expect a shift in the distribution of skills in any job. Initially, it was assumed that the left tail (those with less skill or knowledge) would be most impacted. However, Lakhani revises this assumption, suggesting that the entire distribution will shift to the right and narrow, indicating an overall increase in skill level. Interestingly, the left tail might benefit more than the right tail, as GPT and generative AI could empower individuals in roles like nurse practitioners and paralegals to perform at a higher level, even matching the expertise of physicians while maintaining more empathy. This shift could democratize access to high-quality knowledge work, leading to significant societal and economic transformations.
Question and reimagine processes from the ground up. Don't just apply AI to outdated processes. Lakhani emphasizes this point using a historical analogy: When concrete and asphalt were invented in the 1800s, Boston chose to pour concrete over their existing cow paths, resulting in a chaotic city layout. In contrast, cities like Manhattan decided to implement a rational grid system. Lakhani cautioned against the 'Boston approach' for AI, where companies pour AI solutions over their outdated processes.
3/ AI Productivity Tip: Add ‘Custom Instructions’ to ChatGPT
OpenAI's ChatGPT now offers a “custom instructions” feature that allows you to provide persistent context for your interactions with the AI. This feature lets you inform the chatbot about your specific needs and preferences, which it will remember across all conversations. For instance, if you're a coder who works exclusively in a particular language, your custom instructions can guide ChatGPT always to provide code answers in your preferred language. This feature is currently in beta and available to ChatGPT Plus subscribers. Check out Sam Altman’s (OpenAI CEO) Custom Instructions.
To enable it, go to settings → beta features → toggle “custom instructions” on → then tell it about yourself.
4/ AI Productivity Tip: Decision Analysis with ChatGPT
Need a second opinion on a business decision? Analyze the implications of your decisions, discuss the pros and cons, and suggest enhancements with ChatGPT. Just use this prompt:
"Analyze the implications of implementing [decision] in [business]. Discuss possible advantages and drawbacks, and suggest enhancements to [decision] to mitigate any identified drawback.
Decision = [insert here]
Business = [insert here]"
5/ AI enables a new era of Software as a Service (SaaS), where users can modify apps.
The future of SaaS could be significantly influenced by three converging trends outlined in Dan Shipper's article "What Comes After SaaS?". These include the rise of web-based development environments, the ability to duplicate and remix apps easily, and AI's growing capacity to modify apps for non-coders. This convergence is giving rise to "Malleable Software,” where apps can be modified at the code level using AI. This shift could lead to new business models where software becomes more personalized, and the creation and sale of app modifications could become a new economic sector.
Here is a deep dive into “Malleable Software” and the thesis that soon, all computer users will have the ability to develop small software tools from scratch and describe modifications they’d like made to the software they’re already using.
I'd love your feedback on content and structure for future editions!
Stay curious,
Luis Poggi