Soon, we won’t need to write much any longer. Artificial Intelligence will do it for us. What should we do with all the free time? Paint? Play chess? Write music? Well, AI could do that for us, too. Maybe we should give our best at what AI simply cannot do in our place: we should think.
For 13 billion USD, Microsoft got hold of 49% of OpenAI.2 Microsoft now plans to integrate ChatGPT into Word, PowerPoint, Excel, and Outlook. Today, the majority of business matters are still written in MS Office. As things stand, soon, ChatGPT will make our economy’s heart beat. As an employee, your job will suddenly become very easy. You can automate most of it. To make sure that everything goes as planned, you monitor it from the sidelines, and when necessary, you’ll type in a couple of fuzzy commands to guide it in the right direction.
ChatGPT can read your emails, process your documents and answer your requests for you. It’s only a matter of time, until ChatGPT graduates from Office to Windows itself, runs the whole system, and makes your apps feel like the programs you used to run from cassette tapes.1 Did Microsoft really think this through? Because…
Sooner than later, Artificial Intelligence will get rid of Windows, too, yeah, forget laptops, desktops, and monitors… Artificial Intelligence will run everything from a phone, or a watch, or, a precious little finger ring. AI will do everything for you in a galaxy of clouds far, far away, through an ultra-light client in slightly over 12 parsecs.
It looks like the road to Heaven…2
You’ve heard it before: AI might have the power to radically change society, streamlining business processes, market communication, and challenging traditional hierarchies.
- The end of apps: If ChatGPT is the centerpiece of all apps, you don’t need them, you just need a couple of plugins that render your processed barks.
- The end of business communication: If no one writes or reads emails, Slacks, or Trellos anymore, we might as well stop sending them altogether.
- The end of bosses: If everyone uses the same brain, we don’t need anyone telling us what to do.
Instead of typing boring requests in Outlook, instead of cropping stock imagery and power pointing the newest Lego-block business-clichés, instead of braining up Word memos that no one will read, you can give your computer a couple of approximate orders and the machine processes the text for you. Working on a screen has never been this easy. So… take it easy. It was all bullshit anyways. AI will catapult the existing business circus into a Spaceballs dimension.
The briefings, meetings, memos, proposals, timesheets and invoices, the deliverables, code and code reviews, the marketing campaigns, and analytics… Work on the screen has all been a simulation of itself before AI started simulating them. Time to leave the simulation of simulations to the machines. Let them print money while you focus on things that matter.
…but it could be the road to hell
In a darker corner of the same future, machines simulating work, individuality, and the recognition of language’s limitations challenge the necessity of traditional income, personal identity, and verbal communication as a whole.
- The end of income: What do we need work for when a machine simulates work better than yourself?
- The end of you: What do we need you for when a machine simulates you better than yourself?
- The end of language: If we all know that language is not real, we don’t need to speak anymore.
If no one understands what no one thought, and no one reads what no one has written, we may as well abolish the whole circus of language and enjoy the silence. Wouldn’t it be nice? Jokes aside, it looks like AI devalues all productivity apps, not just the Office Suite, but work on the computer is getting superfluous.3 Screen work has been hit by a Weimar scale inflation.
Time to Think and Rethink
If artificial intelligence frees us from business as usual, what will we do with all our free time? Play chess? Paint in oil? Compose symphonies? Read a printed book? Write on paper? Maybe we do nothing. Maybe we just take it easy. Or maybe it’s time to get real, think, and rethink what we know, what we do, what we aim at and who we are.4
What can we know?
AI has made clear that, to a large degree, we all simulate knowledge and meaning. AI is so good at simulating school and business language because a lot of our own understanding in both spheres is largely simulated.
But to the same degree that AI has made clear how much of our own knowledge is simulated, it can also teach us by counterexample to be more real. AI makes perfectly clear how text sounds and feels and it is just simulated knowledge.
If you have worked with ChatGPT for some time you will recognize its patterns. Working with simulated intelligence can—if you’re not too lazy—make you better at being yourself, stop pretending to know what you don’t know, make you take care not to express what you don’t feel, and only give shape to words that you have felt.
What should we do?
It has become a lot easier to program, market, learn and write content. It has become a lot easier to simulate understanding.
Now, you can let Artificial Intelligence do boring tasks for you. You can let it speak for you, when you don’t care that much. Independent, of what you allow it to do, it can’t think and feel for you. So as a guideline, you shouldn’t let AI take decisions for you. On the contrary, using AI you need to think harder, deeper, clearer. You have more time now, so use it to do what the machine cannot. Use them reasonably. Know what you say. Question your understanding. Use AI as a workout to make sure that what you say is what you know and mean to say.
We need to rethink ourselves—and we need to observe AI. And we need to observe it now, as it still makes obvious mistakes and reveals through its bugs how it works. Once it simulates us perfectly, we won’t understand it at all anymore. We won’t be able to discern it from us and understand how it works.
Right now, and that not something one should say lightly, AI acts a lot like cancer. It grows uncontrolled out of our organic knowledge, and it grows where that organic knowledge already has developed some carcinogenic tissue. Visually, AI generated pictures often look like a form of cancer. The current Dall-E 2 and Stable Diffusion pictures do not convey a healthy, meaningful, and happy world. It’s one dark horror show of stuff growing out of what used to be organic. Soon, this will change and AI-engineers will learn to simulate happy images.
What should we do? It’s always easier to say what we shouldn’t do. We certainly shouldn’t let a machine generate random patterns and then let other people use their time and energy to understand them. What, again by counterexample, we should probably do is use AI for ourselves, to be more, not less, human, to take care that what we say makes sense. We should use our time with AI to waste less of other people’s time.
We can enter into dialogue with AI to test our thoughts, correct involuntary errors, deepen our knowledge by investigating what we don’t know, improve our words, results, calculations, code, and designs. The same way we use spellcheckers, calculators, and design grids to improve our text, thought, and work. In all we do, we need to continue to question the machines, ourselves and how we use them.
What can we hope?
We need to rethink how we spend our free time. AI will have eaten all our hobbies long before it fired us from our job. It beat Kasparov at chess 25 years ago. It doesn’t beat Bach at composing, Picasso at painting, or Shakespeare at writing—but it already makes it easy for a hobbyist to simulate them. Why resist?
The broader point is: Our life goal cannot be to finish our meaningless work quickly and then do something imperfect but meaningful in our free time. This was always a bad take on life. Work has impact. What you do at the office affects others. If you let the machines do meaningless work, the effect will be that meaninglessness will continue to perpetuate.
If you give up on thinking for yourself, on asking yourself, if you mean what you say, feel what you thought, if you give up on understanding, you become a machine. As the machines start out simulating us in almost every aspect, it is not unlikely that, as it happens so often when we produce technology, that at some point we start adapting to them.
Still, there is hope. The hope is that through the right use of machines, we learn to be different from them. If you use AI mindfully, not to fool others but to question yourself, to think deeper and to better control your self-expression instead of automating it, there is a chance that you will understand more and not less.
There is a chance that AI can become a gym for your mind. If you have used AI yourself, you have learned that the output of AI improves the more you understand what you are doing. This is perfectly clear with image generation, but it’s the same when it comes to text and coding. Only if you know what you want and understand what you are doing, can you use your tools appropriately to achieve the results you desire.
AI doesn’t possess human understanding. It simulates it. How so? It doesn’t have a body. It doesn’t sense what it calculates. It doesn’t feel what it processes. It cannot. We also simulate more than we realize, but we can understand. We can feel that we think. We feel and know that we are.
AI feeds off our understanding. It feeds off our understanding in the material it uses to simulate us, and it feeds off our understanding the very moment we use it. We can hope that this will not lead to us being completely drained of any understanding, but rather, to strengthen it.
What does it mean to be human?
To not get sucked into that black hole too soon, let’s shine a light on the idea of human language. Language connects us. Language connects one human being to another. Through space and time. Language transports meaning between minds, sense between bodies, it can make us understand each other and ourselves. It can make us feel what others feel. Language is a bridge.
Language is a bridge. If we disconnect one side of the bridge, the bridge falls. If one listener or speaker, writer, or reader stops feeling what is said, the bridge crumbles. Language without a body is senseless, meaningless, void. As the machines start outdoing us technically, let’s not neglect that understanding can be real and that, at its best, it can be mutual. The end of writing is to understand each other through time and space, to feel what others have felt and to make others feel through time and space how we feel now. AI can mess this up, badly, by blurring what was written by humans and what was generated without feeling or understanding—but it cannot end it.
Realizing what it means to be human may not always turn out to be a pleasing experience. We do not always mean well. But there is an almost unspeakable beauty in the moment where two humans that do mean well, realize that they understand each other.
- “It’s so good that I have investors asking me how they pulled it off, or why OpenAI would even do this,” Microsoft’s $13 billion bet on OpenAI carries huge potential along with plenty of uncertainty, CNBC
- iA will not let AI write your text without what or ifs. We want you to feel what you want to say, think what you want to write, and write so clearly that others will think what you think and feel what you felt. We will continue to use algorithms if they encourage you to think harder. We have been doing that for years (with Syntax Highlight and Style Check). Using AI to correct spelling errors, point out holes or check for factual or logical errors—anything that makes you think more is great. But, whatever you we, we want you to think more, not less. If you’d rather have ChatGPT write for you than feel, write and think for yourself, you can do that elsewhere. To use ChatGPT you don’t need us. It’ll be everywhere else. ↩
- Quoting Freedom, by George Michael. The song’s lyrics are about his own struggle with fame and his complicated relationship with the music industry. It was playing in the background while I wrote the article, and it shows: “Well it looks like the road to Heaven … But it feels like the road to Hell … And everybody’s got to sell … And some mistakes were build to last … That’s what you get, for changing your mind … And after all this time … I just hope you understand … Sometimes the clothes … Do not make the man … All we have to do now … Is take these lies and … Make them true somehow…”
- The popular argument is that those who will be using AI will win and those who do not will be left behind. As long as AI remains a tool and we stay in control, that argument is fully plausible. But we are already so much following technology in how we communicate with each other. The world is full of noise because we are not in control of our information technology but the other way around. When it comes to writing, things are not that easy. If you do not understand the foundation of what you write, it’s not likely that you will end up with a text that has been felt and understood in every aspect. Writing is rewriting and rewriting until the thought becomes clear. AI may help here and there pointing you to unclear elements, but if AI writes for you, you will stop thinking. ↩
In the Critique of the Pure Reason (1781) Kant formulated these four questions to sketch out a framework for understanding the human experience and the nature of knowledge and morality:
1. What can I know?
2. What do I have to do?
3. What can I hope for?
4. What is the human being?”
–Kant, Critique of the Pure Reason, KrV, A 805/B 833
The four fundamental questions appeared already in 1765 in his Lectures on Logic (Log, AA 9, 25).