Hi, there!
Welcome to the 2nd edition of Work in Beta.
Before we start, quick shoutout to Leena, Deep, Darshan, Praneeth, Varun and Chirag for their inputs & feedback on the previous edition.
I started ‘Work in Beta’ to help you find the most fundamental unlocks to master AI - something that I discovered over the last 3 years as a CPO / VP - Product and it has absolutely changed the way I work now. I want the same for you.
In this edition, I am excited to present:
A simple framework to understand and fix AI slop in your writing
De-AI - access to my free CustomGPT to help your writing look less AI generated (even if it is)
Two course from Anthropic Academy that will help enhance foundational understanding and usage of LLMs
THE ‘HOW-TO’ PLAYBOOK
Why your Writing looks like AI Slop (and, how to fix it)
Image Credits: Nano Banana Pro / Work in Beta
AI Slop is not about typically over-used AI words or phrases or even em-dashes.
It is about your writing pretending that you experienced something when you didn't.
You can write text like ‘let's delve into…’ or ‘Not only…, but also…’ and still sound perfectly human if there is an actual lived experience behind what you're saying. If there is none, no level of ‘authentic’ writing will make it real.
So, the real tell of AI Slop is ‘hollowness’. The signs of the missing reality or lived experience. You can always describe Rome in a nuanced way if you have been there versus have just read about it in books. AI Slop sounds like the latter.
Let me give you an example:
Statement 1: "I use hyper-customized projects and CustomGPTs to run my AI-enabled business flows"
Statement 2: "After probably 20-odd iterations, I finally had a functional CustomGPT where when I input a client's website URL, it gives me 3-4 pages of bulleted recent data about them - including LinkedIn profiles of key leaders and what are they writing or talking when it comes to their AI initiatives"
You can clearly distinguish which statement is just words and which is a story than can only come from someone who has built the thing.
So, the fix to slop is simple: connect your writing to your reality.
Add one specific detail. Cut the over-explanation. Replace jargon with what actually happened.
But, doing this every time might get tedious for you. It did get for me.
So I built De-AI.
It's a free CustomGPT that scans your content for hollow claims, forces you to add specific details from lived experience, and strips out over-explanation and jargon.
More details (including link to access) on it below.
THE TOOL STACK SECTION
De-AI (remove AI Slop from your content)
Image Credits: Nano Banana Pro / Work in Beta
De-AI is my go-to tool for grounding my writing in reality and I am excited to invite you to use it as well.
How to use it:
Paste your draft writing into De-AI (and, give some context if you want)
It rates your writing (out of 10) and points to specific problems in it
It asks you some clarifying questions on the above points
And finally, it rewrites while also clarifying what it cleaned up
Use it on your next LinkedIn post, client proposal, or newsletter. I would love to know about your experience using it.
THE ‘WHAT-TO’ PLAYBOOK
2 Foundational Courses from the Makers of Claude
Image Credits: Nano Banana Pro / Work in Beta
In this edition, I am recommending two foundation AI courses from Anthropic (the makers of Claude). I have myself completed both these and found a number of very simple but easily missed lessons.
The first course is Claude 101. The reason why I liked this one is that it fundamentally helps you move past surface-level AI usage and starting adopting mechanics of seeing an LLM as a productivity enhancement system.
The second course is AI Fluency: Framework and Foundations. The reason why I highly recommend this course is that in a very simple manner, it helps you look at something what Anthropic calls the 4D Framework. This framework helps you understand not just how to use AI tools but when to trust them, when to push back, and how to maintain quality control.
THE OUTRO
That is all for today.
Thanks for taking 7 minutes to read through this newsletter. Please do let me know what did you think of it especially if it was useful. And, where can I improve.
What did you think of today’s newsletter?
If it is easier for you, you can also reply to this email with 1 - 2 lines of feedback or even a voice note. You can also DM me on Linkedin or if you are interested we can also connect on a call.
In case you are new to Work in Beta, you can access all previous newsletters and articles here.
Have a great week ahead and see you again next week!
-PD