Ground Reality of AI Tools

The ground reality of AI capabilities is not what you see in your LinkedIn feed.


I’m getting overwhelmed, infuriated, outraged, and anxious all at the same time whenever I see posts claiming AI killed one tool (replaced one role) or the other.


Overwhelmed because every post in my feed is around some kind of AI automation.


Infuriated and outraged because I know most things are not as flawless or seamless as they claim.


Anxious because it makes me wonder if I am missing out on something.


I’m someone who always tries to find shortcuts to get things done faster with minimal effort. So when I heard about tools like GPTs, Lovable, Make, and Gumloop I was equally excited like others. I was happy with the prospect of getting more done with less, and avoiding the pain of convincing people to do what I want.


But if you ask me today, after publishing 150+ blogs in the last six months (which got half a million impressions so far), and spending 100+ hours trying to build 4 micro tools and 5 websites. I’m not as excited as when I first heard about these AI tools.


Let’s take GPTs for example.


Working with GPT is like working with an overconfident intern who thinks they figured it all out.


With a simple prompt, GPT can write a 700-1200 word response in seconds. At face value, it’s awesome, amazing, put whatever superlatives you want.


But when you take a deeper look, you realise all those words are fluff and surface level. The problem with AI content generation is, it is a mixed bag. You can’t reject or accept it completely.


It doesn’t make my job any easier. The quality of its output is as good as the depth of my knowledge and understanding of a particular topic.


The easiest part is I don’t need to convince it to do something for me. I don’t get pushbacks. All I want to do is just prompt. It can do as many iterations as possible.


For all those blog articles, I spent endless hours of prompting, writing, editing, and reviewing. I wouldn’t have done a decent job if I hadn’t had the hands-on experience.


GPT is at the top of the list when it comes to who apologised to me in my entire life.


It never cared to correct its mistakes based on feedback and makes the same mistakes every time. For example, in every other prompt I instruct it to avoid em dashes and it never follows.


Every time I point out its mistakes, it either compliments me or defends itself with a lame reason.


It doesn’t bother to double-check its work before sharing it with me. Whenever I ask to review its output, it will always come up with suggestions to improve it and shamelessly ask ‘would you want me to revise the response’.


It is far more difficult with other AI content tools because you won’t have the freedom like in GPT as they have a narrow vision. They are built to produce SEO content. The responses lack depth and are simply filled with words.


I can’t speak for coders/programmers but for marketing, these AI tools are just good interns. Yes, they are good at proofreading and coming up with initial research snippets. That’s it. They aren’t going to replace marketers, content writers, research analysts, or any other roles at least for now.


About Lovable, Make, Gumloop, Bolt, Cursor? In the next post.

Comments

Post a Comment

Popular posts from this blog

The 95/5 Rule

Some marketing gurus talk about focusing on the 95% of your target market who aren't ready to buy. A few days back, David Reed, ex-Head of Content at Gong, shared how focusing on the 95% helped Gong to grow . But it's never 95% vs. 5%. It's 5% + 95%. Why? You need to survive today to see tomorrow. The 5% in-market segment helps you see another day. So, focus on the 95% only after gaining enough traction from people who are in-market today. If brands want to invest in the 95%, they can emulate the Gong Lab series. And before you get started, here's why the Gong Lab series was a success to begin with. Gong's data-driven sales insights were first of their kind as it shared raw insights into what influenced deal closures. It helped sales leaders to work and coach their team not based on theories, gut feel, or past experience, but data. The content was written by SMEs (ex-sales leaders) like Chris Orlob and Devin Reed. It made a huge difference in the kind of data points...
Read more

What real preparation looks like

Image
  Not everyone can just wing it. When you see someone killing it with their presentation or interview or a sales call or pitching an idea, it's the culmination of hours of preparation. One of my mentees asked what exactly preparation looks like. I tried explaining it to him, but later I realized nothing can explain it better than this scene from Ford vs. Ferrari. Before leaving for Le Mans, Ken Miles walks his son through the entire race in his mind — from the critical first gear turn at Eau Rouge to pushing Top Gear at 210 mph down Mulsanne straight. He describes every corner, every shift, every braking point. From how the car behaves through the Dunlop Bridge, to letting it run free over the rise at White House. It’s the culmination of years spent tuning the perfect car and studying every inch of the circuit. That's what real preparation looks like. Most people think they can wing it, until they are in the moment. It's not just about having slides or data points. It's...
Read more