I am not a popular person on LinkedIn with the “LLM are bad” crowd. Most of my friends are in this unpopular group, since we have all just decided that we would rather explore the potential of these tools, rather than pretend to spit in the direction of their headquarters. You can find these conversations if you really want to. I am not going to promote them. I have been tagged in posts by enough irrational haters for one weekend.
Make no mistake. I have yet to have many healthy conversations with people who have the “LLM are bad” merch proudly sewn into their shirt or stapled to their forehead.
If you believe “LLM are NOT bad”, be prepared for some angery words. I am an enthusiast for the artistic output of next-generation creators using LLM tools, and heartily cheer their work.
I am not ashamed to admit this. I will say that it has gotten me blocked by some people on LinkedIn, tagged in threads by people with very mean words next to it, and generally shocked at the sheer level of toxicity in people’s opinions.
I have even had to start deleting things from threads on my page and putting in a warning that I consider the boundary of Godwin’s law to be the line I draw comments at. You can Google that if you don’t know what that means.
It is absurd and crazy to me that no one wants to accept an “agree to disagree” position here; for some, it has to be all or nothing.
Generally, after it is clear that I am not moved by their tantrums, petty name-calling, and emotional foofaraw, they drop the following bomb.
“You will be sorry when the bubble bursts.”
I want to unpack that a little as I rant about the future of AI and the future of games.
For starters, I think it might come as a shock to some of the folks out there that this is not society’s first tech bubble.
There was a web3 bubble, a mobile 1.0 bubble, a casual games bubble, a social platform bubble, the dotcom bubble, and, for that matter, the “Atari ET buried in the landfill bubble.”
I have lived and worked through most of these. And ultimately, while some percentage of investors with actual dollars took a bath in some of these bubbles, we all managed to survive, and in some cases, we continued doing what we were doing.
Yes, there will probably be an AI bubble. We are going to see a lot of people lose investment money in the frothy run-up to this bubble, and it might be happening RIGHT NOW.
That does not mean AI is going to go away. Even if Anthropic, Claude, and all these big frontier model companies stopped making new models today, I would still be running a local model that does some of the same things their current models do, maybe 25% slower or 25% worse. The gains from using LLM tools are there. I wish I remembered the study link where someone provided proof that 80% of all business activity that people do is better done by AI.
If you find that study, please drop me a note. I would love to have it handy when angry, flustered meme-able people stick their face in mine and scream “SOURCE?!”
It happens quite a lot.
It does say to me that we are going to be seeing a massive change in the future of work. There are whole industries of middle-tier people whose job right now is to take some level of information, manipulate it or synthesize it, and then regurgitate a report to some high muckety-muck. In the process of doing so, they can make recommendations, suggestions, and improvements that will get them promoted.
The last part of that sentence, the recommendations, suggestions, and improvements? Those items will include some level of easy, low-hanging fruit that can be generated by an LLM. Some of it will be stuff that requires what I have called “Platinum Collar Thinking,” which is a fancy way to describe the remaining 20% of work that is best done by humans.
That will still exist, but people with Platinum Collar Thinking can’t just sit around and do the 20% of the work that used to exist and expect to be rewarded 100% for it. They are going to have to do four to five times as much work as they did before, in ways that matter to their employers.
We are all going to have to get much better at doing our everyday things when we drop some information into the Claude-inator 2000 and have it emit some CSV or a report that might need a little finessing or “interpretation.”
So let’s bring that to games.
We are seeing new inventions, new tools, and new technologies leveraging LLMs showing up every day.
We are not yet in the territory of having our progress bounded by stuff like Moore’s law. I have twice run into technology-based companies that are doing two or three orders of magnitude improvements over existing LLM capacity, quality, or speed. Some of these improvements are even stackable.
While these advances are largely bleeding-edge and are only now entering the wide, wide world of people who are “level 1” on the four levels of AI adoption, we are going to see huge transformative changes in the way we think and the way we work. I attended an event called “FrAIday” here in my town, where we had a deep hour-long conversation about this.
But the biggest thing that is going to change?
How we play.
This is really why I get called all kinds of names and people block me for my “AI-First” position.
We have been through multiple waves of technology change in the past thirty years. I have had a front-seat view of much of it.
The internet. The first smartphones. The social internet. The touchscreen smartphone (mobile 2.0).
I also meandered through a few near misses. Virtual reality. Web3. The Metaverse. All three of these will manifest themselves in coming waves of technology and how we interact with machines, and each other, and how our machines interact with other people’s machines.
The wedge in previous iterations of technology, going all the way back to before my time and including things like the printing press and television, has always been entertainment.
And the same thing will be true here with LLM tools.
And this is what is generating all the fear and confusion, and people throwing tantrums.
We are going to see new genres and new forms of entertainment using LLM technology coming fast. We are going through most of the last waves of “just use LLMs to supplement parts of the pipeline to make existing things.”
Consumers are not excited by that, and the industry does not love it because it is already in a massive down cycle for other reasons. If you spend less than five minutes thinking about the art that is possible with LLM technology, it is really easy to get a low-level threat on Maslow’s Hierarchy of Needs and react with “Oh Oh, It Is Coming For My Jerb.”
This is especially true if you have already lost your job. But right now, that job isn’t coming back anytime soon. And this is a convenient scapegoat for it.
The pattern I see in some people with a lot of grit and a lot of curiosity and a lot of the people I have been fortunate to work with is to lean in, learn something new, and continue to make things. This is true for people I have worked with for one year, and this is true of some of the people I have known and appreciated their professional partnership on things for thirty years.
I spooked the living shit out of a senior creative director, showing off my vision for the future of entertainment. “In the world you are describing,” he said slowly, “The job I have right now goes away.”
Yes, it does. But a dozen new jobs will arise in its place.
Much like when the printing press arrived, and the ownership of books and the ability to learn things and get educated transformed with the lowering of the barrier of cost of creating books, we all benefited, and society moved forward.
You see lots of big celebrities and senior creative people talking about their concerns. You are damned right. They have been gatekeepers for professional success in the world of leisure. And those gates are being blasted open. The ability to create something will no longer require up to nine thousand highly paid people to generate something that is worth consuming.
Solo creators and smaller teams will throw their creations together using well-articulated designs, LLM tools in partnership with niche craftsmen and tastemakers, to redefine the very things we do for fun.
The problem with this is probably one of timing. The hits to the creative industry caused by interest rates, changes in consumption patterns caused by COVID, and eventually the “end” of COVID, and other factors for games like the growth of Roblox, and in Hollywood, things like the writer’s strike and the inability for studios to take serious risks on fresh approaches to content have created a window of opportunity for new things to emerge.
In the next five years, we are going to see a further decrease in the number of AAA titles that are shipped. We are witnessing the slow meltdown of the console space, where everyone has raised their hardware prices due to the component demands, which will complicate next-generation designs and launches.
We have never seen such a radical transformation from top to bottom of just about everything.
I have spent most of my thirty-year career doing zero-to-one startups and dealing with crazy levels of uncertainty. I was built for this kind of low-oxygen experimentation, and I love the challenge of trying to understand how to thrive when this level of uncertainty exists.
I also think this is probably the biggest technological change we are going to see for a while, and there are only going to be two or three waves as big, or bigger, possibly ever.
So why am I writing this?
Because if you have your head in the sand right now, you are going to miss the opportunity of a lifetime.
I really don’t care about all of the name-calling, blocking, and gaslighting I have received simply for having this opinion. I don’t care for it either, honestly.
But I want to explain why I am excited for the opportunities that are unfolding and why I always lean into what is possible.
I am curious to see how this all shakes out by 2030.
How about you?
