ChatGPT and the A.I. Education Arms Race

(Plus Thoughts on Becoming an A.I. TikTok Creator!)

👋🏽 Hi friends!

I want to try something new. Last time I updated you, it was about my podcast with actor Greg Grunberg! This time, I wanted to share something a little different — an analysis of the rapidly-moving A.I. revolution and how it will impact education and society.

As part of this change, I’ve renamed my newsletter to “The Social Analyst” — the name of my columns at both Mashable and CNET. It encapsulates what my newsletter is about and what I love to do — analyzing and understanding how technology affects society.

There will also be personal updates in my newsletters and ways we can connect. (Scroll to the bottom to hear about how I somehow became an A.I. educator on TikTok!)

One more thing: I’ve moved this newsletter to a new home — Substack.

Now, back to the topic of this month’s The Social Analyst.

---

Is the high school essay dead?

A few weeks ago, I asked ChatGPT to write me a 5-paragraph essay about the themes of Catcher in the Rye. Then I asked it to rewrite the essay for the college level.

It took me 15 seconds. 15 seconds to do something my 8th-grade self would have needed a few hours to properly write.

As my co-founder Matt likes to say, this is the worst A.I. will ever be in human history. And it's already pretty mind-blowing.

Why would any kid, or any college student, use anything else but ChatGPT to write their essays — or at least write the core of their essays and then edit it from there? My 8th-grade self would have absolutely used this technology.

Very few people realized that artificial intelligence had advanced so far to the point that it could write legal contracts, create award-winning art on command, or even write newsletters. (I used an A.I. tool called Lex to help me write this one.) With the right prompts, you can churn out a hundred unique blog posts, code an entire video game, or write lyrics for your next album.

I recently predicted that someone is going to make a 7-figure business out of being an expert at writing prompts for A.I. In fact, I’m already starting to see a cottage industry of prompt-writing experts dishing out advice on TikTok.

I don’t think there is any going back now. Not just for the college essay — but for society as a whole. The last few weeks of fervor around ChatGPT and the capabilities of A.I. have convinced me that we’ve entered a crucial point of no return.

A fundamental transformation is about to overtake education, business, and almost every knowledge worker-heavy industry. Society is not ready.

The A.I. Education Arms Race

In the next two years, schools are going to attempt to win an arms race against A.I. Students will “magically” start writing much better essays. Teachers will only allow students to write essays in class (perhaps ignoring more research-intensive, take-home essays). They will ban ChatGPT from being accessed, and I suspect they will use new tools to tell if an essay was written by A.I.

But they can never win this arms race — GPT-4 is around the corner and will wipe out the effectiveness existing A.I. detection software. Then there will be GPT-5, not to mention other language models. Hell, you can just ask Copy AI to write a blog post, then ask ChatGPT or QuillBot to rewrite it in the style of a cowboy, a professor, or Shakespeare and it will beat the A.I. detection software 99% of the time.

The reality is that the knowledge workers of the future will use A.I. in every part of their daily work. In the same Millennials must use calculators and Excel at work, Gen Z and Gen Alpha will have to use tools like ChatGPT to be competitive in the job market.

Why would any company hire a copywriter or marketer that can’t keep up with someone who uses A.I. to be 2x, 3, or 5x more efficient? Why wouldn’t a student use A.I. to get over writer’s block and get a first draft done in seconds? It won’t be long until there’s a sophisticated A.I. tool for every industry, creating legal contracts, replacing stock imagery, and providing second medical opinions.

So What Will Society Look Like in a Post-GPT world?

Since it’s the end of the year (aka prediction season!), I have some predictions for how the A.I. revolution will play out in our society:

1) The pervasiveness of A.I. in the workplace will be swift and total. Companies using A.I. to automate more of their business will outperform those who don’t in significant margins. As the economy recovers, tech companies especially — now deeply scarred by multiple rounds of layoffs — will increase their use of A.I. and automation in order to minimize their costs so they never experience something like the layoffs of 2022 ever again.

2) There will be significant resistance and backlash to the rapid rise of A.I. in daily life. Ethical, social and legal questions will be raised around the data sets used to train different A.I.s and whether A.I. will make us worse at critical thinking. (This backlash is already happening in the art world.) It’s the same kind of blowback we saw with smartphones.

3) The A.I. genie is out of the bottle. Tools like ChatGPT are here to stay. They will be ubiquitous and easily accessible (cheap or free). New A.I. tools are being built and adopted in rapid succession. (I’m tracking them all.) Regulators are not great at keeping up with new, rapidly-changing technologies, and for most of society, the usefulness of A.I. greatly outweighs any negatives. You will see A.I. become a standard part of writing, programming, marketing, and sales.

4) Education and society will move from resistance to adaptation. For a year or two, schools may ban ChatGPT, but the best educators will incorporate it into their lesson plans. Several teachers on TikTok have told me they will have their students review essays written by A.I. for errors and language improvements. Other teachers have told me they will have students work on things A.I. cannot do currently, such as cite sources and write out the critical thinking they use to come to conclusions about new works and current events.

5) It will take a few decades for physical work and labor to be supplemented by A.I. and automation (e.g. construction, trucking, etc). It’s just easier to automate things that only require software. Hardware breaks down and is clumsy. Humans had millions of years to evolve opposable thumbs and precise muscle control. But eventually physical tasks will become much more automated with A.I. These jobs will never be completely replaced, but we will need far fewer people to build a building or run a warehouse.

6) Within the next generation or two, the vast majority of the human race will not need to work in order for society to function. This opinion is a bit more controversial. I don’t believe we will need as many people to generate the same amount of economic output once A.I. becomes ubiquitous. A.I. just doesn’t need a lot of managers. In the long-term, this will be amazing for the human race, as we open up more of our time to exploring the things we love and spending less of our time working on things we hate just to survive. This is something I’ve been predicting since 2016.

This transition will take time. It’ll happen over decades. And our society isn’t yet prepared to make this transition. But when we do make the transition, society will be better off. This is a topic I will talk about much more in-depth in this newsletter.

Octane AI Spent 6 Years Building Towards the A.I. Revolution

As you can imagine, the recent spike in interest in A.I. has been validating for the Octane AI team — especially for my co-founder Matt Schlicht and me, who have been bullish on the potential of A.I. since we started the company in 2016.

Matt in particular has been working with GPT-3 since the private beta in September 2020. He used GPT-3 to build Octie.ai, an A.I. assistant for the ecommerce industry. He’s been hard at work with the team, building some new and amazing A.I. products for the thousands of ecommerce brands using Octane A.I. We will be talking more about what we’ve been working on soon.

Matt Schlicht was able to build Octie in a cave! With a box of scraps!

We literally started Octane AI as a chatbot company and now — 6 years later — chatbots are back in vogue — and actually realizing their full potential. We firmly believe that chat is the most natural interface for any product. In fact, it’s a key prediction my co-founder made all the way back in 2016.

Matt and I are also working with several startups we advise on how they can use A.I. to supercharge their businesses. Ping me if you want to talk more about what we’ve learned.

The TikTok Era

TikTok is insane.

A few weeks ago, I had 500 followers on TikTok. Now I have over 14,000 followers.

It all started, funny enough, with A.I. and schools. A few weeks ago, I made a 3-minute TikTok video about how A.I. was going to radically change education.

The result: my TikTok went viral with over 1.4 million views, 200,000 likes and 5000+ comments. My video’s received more views than there are people in San Diego! Hundreds of teachers are still debating the merits and drawbacks of A.I. in my comments section.

Comments section of my TikTok on education goes on forever!

It was eye-opening for someone who has been in the business of social media content creation his whole career. It rekindled my love for creating content.

Since then, I’ve been experimenting with TikTok, and the results have been fascinating. It’s hard to predict as a creator what will and will not go viral on TikTok, but I’ve found a niche exploring the background of A.I. and the layoffs at tech’s biggest companies.

I’ve even started a TikTok series on amazing A.I. tools that will make your life easier. I think it’s worth checking out if you’re interested in what’s new with A.I. It’s all part of my in-depth research into the wider A.I. landscape.

My next goal is to reach 100,000 followers on TikTok. I have no idea how long it will take. If you regularly use TikTok, I’d appreciate a follow. The more people who watch, like and comment on my videos, the more quickly my educational content will spread.

Gives and Asks

I recently attended a Brex Supper Club for L.A.-based founders, and my favorite part was “Give and Ask” — when each founder offered something and asked for something from the other founders in the room.

It’s the holiday season, so I thought something similar would be appropriate!

Give:

1. Founder-to-founder advice. I’ve been trying to help my founder friends through very difficult times. Founders are facing impossible decisions in this economy. Maybe my experience or just an open ear can help.

2. A casual chat with me about what’s happening in A.I. I’m oceans deep and happy to share what I’ve learned.

3. Recommendations for amazing restaurants, bars, coffee shops, and things to do here in Culver City, in the heart of L.A. If you are coming to Culver City, let me know.

Ask:

1. An introduction to amazing and accomplished experts in A.I.! Matt and I are constantly trying to level up our knowledge in the field.

2. Suggestions for things to do in Japan. Deborah and I are going there for the first time this year, and we want to see and try as much as we can.

3. Suggestions for A.I. tools that have made your life easier. I am looking to feature at least 100 A.I. products on my TikTok, and there are plenty I know nothing about (yet).

4. If you enjoyed this newsletter, please share it with your friends!

One Last Thing…

Happy Holidays! To celebrate, I asked A.I. to paint Santa riding a reindeer while holding a menorah. Here’s the result:

A.I. clearly doesn't know how many candles are in a menorah.

A.I. is amazing, but it’s still far from perfect. I’ll try this again a year from now and we’ll see how much it improves.

Happy Holidays,

~ Ben