4 Min Read • Intensity: ☕️ 

Good morning, tech enthusiasts! Let's explore some GPT-related news and examine some GPT limitations to ease job replacement fears (fingers crossed!). 😅

Grab your favorite cup of Jo and let's dive in!

• Google CEO’s view of AI acceleration 💨 

• GPT's Achilles Heel: Common Senses 🧠

• GPT’s Comedy Conundrum: No Jokes 🤡 

• Plot twist: is GPT really that dumb? 🤓 

Sundar Pichai’s AI Tsunami Warning

Sundar Pichai (CEO of Google) alerted society to the AI wave on April 17th, stressing the importance of ethical development and collaborative regulations.

AI's Fast and Furious Future 🏎️

Google CEO Sundar Pichai is sounding the alarm: the AI revolution is coming, and it's coming fast. To ensure we're not swept off our feet, he urges mindful reinforcements of this transformative technology.

Journalism ❌ Financial Transaction ❌ Information Tech ❌

Pichai highlights three knowledge-based industries where AI could be a game-changer:

• Journalism → Writers

• Financial Transactions → Accountants

• Information Technology → Software Engineers

With the right approach, these sectors could be transformed for the better.

Keeping AI in Check with Regulations

Pichai insists on regulations and global teamwork to ensure AI serves the greater good and doesn't widen existing inequalities.

Additionally, Pichai cautioned about the repercussions of AI, stating that the issue of disinformation, fake news and images, will be on a much larger scale and could potentially cause harm.

Just like this one ⬇️ 

Sources and full interview with 60 Minutes

What AI Can't Do

Now that we are aware of GPT's capabilities, let's discuss its limitations. While it has been suggested that GPT will replace 66% of the jobs out there, that is probably a bit of an exaggeration.

More realistically, GPT and other AI will have an impact on that percentage of jobs, including full placement, but not in all cases. I believe this because there are certain tasks that GPT is simply incapable of performing. Here are a few examples:

❌ Common Senses

Let’s ask GPT if cow lay eggs 🤫 

Why GPT Fumbles in the Real World

GPT may excel in language generation, but it isn’t actually alive and hasn’t really perceived the world. This causes AI to fall short in the realm of common sense knowledge and understanding, revealing a critical limitation and exploitable area.

Though GPT can generate text like a pro, it doesn't truly grasp the meaning of what it is generating. GPT is a “next-word prediction model”. This means while it is “chatting” in the first sentence, it does not have a cohesive plan about constructing the last paragraph.

Put simply, common sense remains elusive to GPT, as GPT can't reason like humans.

Striving for a Deeper Understanding

While GPT's language skills are remarkable, it reminds us that AI still has a long way to go in emulating human common sense and understanding.

❌ Basic Emotions

GPT's language generation capabilities are impressive, but its inability to truly understand human emotions and the context in which they're expressed reveals an emotional intelligence gap in AI.

Text Generation, Not Emotional Recognition

GPT may generate emotional language and respond in a learned manner, but it doesn't actually grasp the emotional context. Unlike humans, it can't understand the underlying meaning or emotions behind the text.

GPT still has a long way to go in terms of emotional intelligence. But hey, if they develop a Kevin Hart humor setting, that wouldn’t be a bad start!

Say Goodbye to Goofy GPT-3

GPT 3.5 free-tier versions often had their share of eccentric responses, like the ones I have above. In a more extreme case, GPT 3.5 even declared its love to a man and ask him to divorce his wife. What kind of advice is that?

The GPT-4 Revolution

GPT-4 aims to fix a plethora of issues, improving not only common sense and emotional intelligence, but also safety, hallucination, code efficiency, input sources, and token intake.

Want an in-depth look at GPT-4's improvements? Subscribe, and we'll dive deep into the world of GPT-4 in an upcoming article.

GPT 3.5 vs GPT 4

For now, here is a tiny sneak peek of how good GPT 4 is. I asked:

GPT 3.5’s Response

GPT 4’s Response

And for cat lovers out there:

Seems like there’s a clear bias between dogs and cats here 🤔 

Until next time,

Your friendly barista and ☕️ GPT!