The AI skeptics' take on AGI by 2027.

AGI BY 2027 - The Flip Side of The Coin 🪙

Read Time: 6 minutes

If you're not entirely convinced by the 'AGI by 2027' argument, this read's for you. Let's dig in and explore some of the reasoning behind why we won’t be at AGI anytime soon.

Want a recap on the faster AI argument? Check my email from 2 days ago.

The Counter-Arguments 👀

1️⃣ We're Complex

AI fanboys under-appreciate the complexities of human perception, vision, learning, and reasoning. Replicating this into a system that can communicate with your coworkers will not happen without several dozen significant breakthroughs from where ChatGPT is today.

2️⃣ We'll Need Energy

We will need a LOT of computing to make AI work in systems like SalesForce, Adwords, Meta, LinkedIn, and Gmail. And in a world moving to electric cars (6-8% of global energy), and needing more AC to stay cool, will we have enough energy to power AI too?

3️⃣ AI is Data-Guzzling, and the Internet's Too Small

We're running out of high-quality data to learn from. As a guy who ran a search engine, I know this is an under-appreciated issue. We can’t expect great results from your college-level AI marketing team if Reddit is the primary source for how to run AI-powered Adwords...

4️⃣ The Human Touch

I don’t see us letting AI make complex decisions, with major impacts, without human supervision for decades.

My Tesla regularly freaks out, which is why it will be an AI AND Human equation and not an AI OR equation for decades to come.

5️⃣ We're Not Entirely Comfortable with it..

Right now, consumer trust in AI is fairly low. We don’t like the unknown, particularly in high-stake situations.

Take medicine...

6️⃣ Robots Are Heavily Biased

If you train AI on the Internet, you get lots of bias, and general drivel. Until we can build a sophisticated set of AI agents from our own selected petabytes of clean historical data, we won't get AI that is as good as our marketing intern; we have a long way to go.

7️⃣ We Need It To Work Better

I don’t want to wear glasses or goggles. I want to talk to AI, and I want it to converse back. ChatGPT voice ability on mobile is a start, but it needs to work on my PC, Mac, and Apps to be truly useful.

8️⃣ What Happens When Things Go Belly-Up?

It’s been illegal to pump your own gas in New Jersey since 1949. Our politicians are going to have a very hard time legislating on AI. Testing. Who does it? Certifications. Who issues them? Data Privacy. Can we sell your AI agent data to Experian? Liability. When AI gets it wrong, who's at fault?

It's no secret I’m pro AI.

But my prediction is the future is going to be AI AND humans.

Not AI OR humans.

Which will slow adoption down.

Which is a good thing.

THe ai learning curve & PRivacy

Consider, for a second, your own mobile app usage.

The typical phone is littered with apps you've used once.

The app overpromised and underdelivered. So with your limited time, you went back to the apps you use habitually: Meta, TikTok, and YouTube.

It's the same for business apps/software. Product market fit is hard. We, marketers, overpromise and get then time locked out. AI won’t be different, but it will be a mistake to share your hard-earned data with these applications...

When You're Ready, Here's 3 More Ways To Get Your Team COMFORTABLE WITH AI 🤝

1. Let my team, help you with Marketing AI. Take this two minute quiz.
2. Want an intro to AI? Join my AI in marketing course at UNH next week. There's a couple of spots left.
3. When you're talking to your team about using AI to increase efficiency, why not start today? My current favorite text app is Cuppa. We use it to create high-quality, optimized blogs.

Your 5-Point To-do List to get prepared for AGI

The World Developer's Conference made it clear that Apple's current concern is privacy.

And if it's on Apple's radar, it should be on yours.

My advice:

#1 Own your data

#2 Build your AI skills, because prompting is a skill

#3 Build your team's AI bandwidth, as they are overwhelmed by vendors, apps...and idiots

#4 Test AI and then test it again in 6 months (Chat GPT 3.5 is not 4o!)

#5 Start with the end in mind. What problem are you trying to solve?

Hug the AI skeptic in your life. They are doing good work. I love them. We need their voices in the mix.

But I’m off helping my team build their skills for Chat GPT-5. As it's coming and we will be ready.

P.S If you have a vendor in your life that is making a lot of AI noise, ask them to make a bet, your time's valuable. What will they wager if their software doesn't live up to the pitch? Free for a year? Buy your team lunch for the week? Tickets to Taylor Swift?

Get them to put a little skin in the game.

I hope you are staying cool! See you again on Wednesday. 🤝