“How It’s Done – Rock Version” by Kenshi, Kim X, Gogy
I try exceptionally hard to not write on this blog about two things: Stock. and. AI.
But how can one look at the world and avoid talking about either when we are leaders, or generally people, in games? If you avoid the subject you look incompetent, yet if you have a strong opinion on either of these things you set the tone for your team, your friends, and others’ belief in you all while hoping your communications department still loves you, and you have not become the vector for a PR disaster. This is the nightmare pendulum of being a manager person right now.
No pressure.
Opinions are fun regardless of employer.
And because I turned off analytics on this blog for my own architectural reasons, I have literally no idea if any person reading is that super nice fan who was screenshot to me with their name blurred as “loving my blog” (hi, thank you, wow. I have fans. That’s wild.) or Kotaku. F*ck! Please don’t quote me! I am just a manager trying to survive in an incredibly demonic timeline where no one feels loved at the moment and has never felt more like a number in a spreadsheet.
How did our ability to generate so much cr*p so quickly with tooling get out of hand so fast.
I see the promise of AI, I do. I don’t need to be convinced. Not 10 years ago. Not today when we are reinventing the wheel of step functions and lambda as AI agents even if not every version of AI is meant to be natural language and there are problems with trying to squeeze every process ever of mankind into a natural language human computer interaction.
Remember the development of the mouse? At one point we thought the Rollkugel-Steuerung was going to be where mice made it and…now people use trackpads and tablets with wireless pens. We have a lonnng way to go, friends. Having an opinion now or saving it won’t save you from potentially being misaligned with anyone in your future. Might as well lead in the open.
But I also heard from one of my friends that every time he gets on LinkedIn now he takes psychic damage, and this made me ugly laugh so I didn’t cry.
As games stock continue to plummet with the rest of the world’s stocks, from decisions made by people who never actually made a game and maybe also never played them either, and we collectively in a misery-hates-company sort of way decide whether AI will be able to replace all our jobs, I asked a Slack group of industry friends if I should “Buy the dip or pee my pants” and one astute colleague’s response was the “Por qué no los dos?” meme.
You cannot make this up. AI cannot make this up. Only we can write the above because we are living it in the fastest training speed which is real-time.
But my favorite part is AI hasn’t actually helped me with my hardest problems and I am still sitting here wishing so much that it had. I wish AI could make decisions for a team on which tools they had so passionately built, and love, and see value in, to deprecate against other tools that other engineers had also so passionately built without it hurting someone who spent a long time on a project. Can it compare the features? Sure. Can it help me understand which ones have less liveops problem? Not without a ton of input from teams and customers first which is the hard legwork of collecting data.
“But Molly! AI can collect data for you!”
YES BUT AI CANNOT MAKE PEOPLE GIVE A SHIT :).
Ask anyone who has ever sent a survey or gone door to door and you would understand that AI absolutely cannot force people to want to listen to a Jehovah’s witness.
And somehow, in our efforts to make people give even more sh*ts about our stuff we now use it to make slides prettier and more slides, lots more slides, more verbosity and text, and this has made people give even less sh*ts and exclude even more people because there isn’t enough space for all our collective us + now AI talking against all the actual time we have in the world.
By the way, I don’t have a good quantification of this except the absence of data that people want to collect because everyone in the world wants to care about productivity over inclusivity, which absolutely are at odds with each other right now.
People are using, and are going to keep using, AI to build new games. Will they build better games? Some sure. But they will also use it to build shittier games. Woohoo!!! This is my favorite irony about the stock market!
We think AI tooling is going to BUILD ONLY BETTER games.
No! It is going to also build SHITTIER GAMES!
It’s people who build things.
Making games fun was already exceptionally difficult. Understanding what people find fun is hard as sh*t and always has been.
The most commonly requested feature from game devs? “If you could help us find fun faster – you would have the most important tool.” All the tools creators build are in service to this ideal – whether they are AI tools or not – not make games faster. Find fun faster. Because fun is what ensures we still have a job tomorrow.
Curation requires human input.
I’m reminded of this physical process we used for 21 Potions at Ker-Chunk Games with Drowning Monkeys (it’s been years). We prototyped everything via cards by hand. I used to carry around 5 decks in my purse, not because we couldn’t do it digitally. Sure we could. But because iterating on a new idea was easy then. It made us care more. It made me feel more in touch with the game’s design – no guardrails, no tools problems. Just us and the mechanics. I think AI can make iterating and trying a new idea just as easy but that’s just the thing – it may make trying some new ideas faster. It doesn’t make people care about those ideas faster.
And this is the feeling I keep coming back to in my soul: Curation is caring. It’s about identifying and being brutally honest about what you care about and what your colleagues and what your customers, players, and consumers care about and why you care about them – you care about them for the customer but, hopefully, also each other. AI can help you maybe identify this – it can help you figure out things to try. It can help you help people try them faster, sometimes at the expense of quality, but it absolutely cannot make someone care.
AI can help you automate. It can help you design systems and tools to put stuff into spreadsheets. But it cannot force a leader to stop making another team manually copy and paste alerts into cells because that leader feels that the process should be that way as a matter of practice and hygiene and you may see it as controlling and bullying. These are perspectives for which nuance and debate can only solve.
AI can help you find issues in your stack with advanced anomaly detection in monitoring. It can help you investigate. But it cannot help your partner team apologize after they grilled a team at 5 PM on a weekday, wouldn’t listen when they said they needed to train them in order to help them investigate the problem, and then escalated it up to make their weekend even shittier.
Frankly, AI doesn’t solve the hard problems managers deal with as part of conflict resolution. Because when I say I am sorry my sincerity matters but when AI does it it just needs to make sure that next time 2 + 2 actually does equal 4 this time before managers across the board start using it all over the place in their review cycles to cut out the time in a crushed org chart, the very place where humanity is needed.
And finally, making someone care requires them to have space. And the one thing AI is absolutely crap about – is helping you to have it.
So, as a leader, which you are too, I’m now sitting here deciding where I want it in my life. Which space it gets to have and which space it doesn’t. I wake up just hoping that I won’t have to tell Alexa for a 3rd time how to turn off my lights hoping someday, someone may train her just a little bit more on women’s voices. I hope that I won’t have to upgrade to Alexa Plus just to find out if the version that cannot answer how it uses and stores my data (I asked when trying Alexa Plus), can still actually turn off my lights without also asking me if I was interested in if dogs could eat peaches or if I want the crypto update of the day.
I’ll still talk to Claude a bit more and build some mega cool spreadsheets and feel the glee of knocking out a frontend. But we ain’t ever going on a date and I’m not feeding it pictures of my kid.
If you are feeling lost – it is okay to admit that. That’s the sign of someone truly growing.
Some of you are following leaders who act like they have all the answers and wondering “How they can possibly have AI all figured out when I still have to sometimes switch to Google provided DNS because AT&T in my own home goes down and that requires expertise to know what that means” and “It’s so easy to create monotony now, prettified and call it strategy and leadership.” All of these things can be true.
I’m going to rip of the bandaid and here are my thoughts – is AI replacing jobs? Yes. Yes it is – I’ve seen it as my friends who are writers are failing to get employed. To what scale? No idea. Do we have to use it? Yes. We are living with this everywhere – you need to try.
But will it make games better? Only the players can tell you that. The customer has all the power and always will – and people are customers, not AI.
Header Image by Aron Visuals from Unsplash.
