The Daemon Counsellor
How I stopped worrying and learned to love the AI
I’m writing this post with a new laptop made for the 2020s: Intel Core Ultra 5, 16 GB RAM, and Intel Arc Graphics. 30 minutes before buying it, I didn’t know what Ultra was and thought “Oh no, you didn’t Jedi mind trick me again.”
I guess that’s the thing with computer sales (or more cruelly in my experience, service) people: They don’t tell you what you need to know, and they will repeat what you already knew. And this isn’t just in Asia, as I experienced the same thing with European Australian men. So, I asked the staff to lay down the catalogue on the table and asked Google Gemini on what did it think of the new model he offered.
“What a steal!”, Gemini exclaimed and explained to me that Ultra is the new generation of Intel Core processor, and then arguing on why the model I was offered would be a great purchase. About one hour later I asked it for lunch suggestions. Three hours later I consulted for bits and pieces in setting up my new laptop.
One year ago, I was still sceptical about AI, thanks to dodgy Google search results, hostility from commentators such as Justine Bateman, and personal experiences of mutuals dealing with horrible AI applications: Teachers receiving homework wholly done by AI, editors receiving pitches and drafts written by AI, and people hearing on how office workers were almost-mandated to use their office’s choice of AI monthly. Personally, I also dislike how corporations use AI-made graphics instead of paying a graphic artist.
What have changed over this year? Curiosity. Knowing that my questions to humans are unanswered on X and receiving shrugs irl. Stories on how people have been helped by AI. Thoughts by friends such as Hollis Robbins and Katherine Dee. Perhaps AI programs themselves have improved a lot over the past three years. And eventually, major improvements to my own life, for free.
1. That’s So Mario: Alternative to Wikipedia, alternate histories, and images
Many authors have written about how Wikipedia has declined over the last decade, especially on the Middle East and American politics. This is because like in X, many people have decided that history, politics, and everything else is about agenda, not facts (maybe it’s always that way).
The oldest AI question and answer that I still save is from 15 July 2025. I asked X’s Grok if 9/11 was preventable, and what it would have done to prevent it. It told me what I had known: CIA and FBI didn’t work together, FAA, airport security, and immigration department failed their parts, and it was the first airplane hijacking that ended with a suicide crash instead of hostage taking.
It suggests integrated info sharing, FAA’s and airlines’ focus on plausible terror scenarios, and more serious follow-up on leads. It identified what were the challenges: A very different America in the Millennium.
The ball rolled from there and Grok was the gateway drug. Its button explains inside jokes and subtweets my mutuals make. Before I knew it, I had Grok designed a series of sitcoms we could have had in the 2020s, better EA Sports FC and The Sims games, and modern paperbacks that never existed.
I moved on to Google Gemini as it attempted to replace the traditional method of Google search into AI summary. Since Indonesia bans access to Reddit, I found Gemini’s (and others’) research and harvesting into Reddit’s knowledge to be invaluable. Even if the access to Reddit had been possible, its chaotic format wasn’t for me.
It’s got to the point where I paid for ChatGPT and Claude. ChatGPT’s Dall-E designed those dream sitcoms, video games, and movies posters, and drew people from historical times. When Claude says it cannot draw images, I could ask it to explain why.
*
2. Better Life Through Artificial Intelligence
I’ve come to see Gemini as the evolution of the traditional ways we used Google: For directory, consultation, and suggestions. I’ve used Gemini for consumer advisory, meal suggestions, and repair ideas. Again, I don’t think it’s worse than consulting WikiHow or Quora, and maybe even better. I’m particularly inspired by Kat Rosenfield’s story on how ChatGPT helped her installing an antique mirror.
Their (referencing to my council of Gemini, Grok, ChatGPT, and Claude) greatest achievement so far is to solve my lifelong question which no one I personally know can answer: What should I have done to stay in Australia in the mid-2000s? All of them answered that achieving that with a Bachelor of Arts degree wasn’t the way – Australia didn’t need one.
ChatGPT suggested the extreme (for me, not for other Indonesians) solution of switching my major to Accounting or IT, the most common success stories at that time. Others considered my deficiency in computing stuff (on paper or on screen) and looked for other solutions. Graduate Diploma in Education and be a teacher in either regional Victoria or lesser popular states, such as South Australia and Tasmania.
Grok and Gemini, meanwhile, looked for other ways beyond the Skilled Occupations List. My two alternatives are a partner visa, as previously I asked about the viable ways to have got an Australian girlfriend in two years (“Get out of the international students bubble” was the consensus, something I kind of did but not enough); and regional government or business sponsorships.
They listed the subclass visas I could have applied and what were my career options. Selling lettuce and being a diversity hire at the city council were viable (even managing mercenaries!), and so was teaching Indonesian in Tasmania or Warrnambool. But being an international students’ officer in Melbourne was not, and probably not something I would have wanted.
No therapist in the world could have solved my rumination, with apologies to any therapist reading this. Too specific, too technical, too obscure and un-medical. But it all made sense and helped me understanding the paths that other Indonesians I knew had taken (quite sure at least one took the partner visa option, and not a woman).
AI tells me what no human told me: All the personal growth I could have had as a student, all the diligence and intelligence, would mean nothing for the Australian government: They just wanted an accountant, or a lettuce salesman.
*
3. Limitations and Potentials
“AI May Make Mistakes,”, and they do, a lot! I’ve corrected the disparity between day and date, basic information mistakes, and repeated, canned suggestions. And as a pretty un-social (not really antisocial) person I don’t see them to be inferior to the human touch. I understand the worries and the sentiments about communicating with machines instead of human, but well, I’ve been let down by humans, and have let down human too many times. As I explained above, no human could have answered properly how could I have stayed in Australia in 2006.
Now to the potentials and the works that come with them. Clearing my old life and making a new one. Settling well into my fifties. Navigating this bleak world that is considerably worse off than in 25 years ago, but also, for some ways, offering some possibilities that were impossible during the peak of human civilisation as we know it.