Therapy

Everything is AI now. Almost.

In my continuing experiment to make everything AI because I have apparently given up on thinking for myself (apart from this blog, which is entirely written by me, warts and all) I have started to lean on AI tools to help me deal with my mental narratives.

I’ve mentioned before that I go to see a therapist. She is a wonderful human being who takes precisely zero shit from me, calls me out when I’m prevaricating or just making shit up, and nothing beats the experience of talking to a professional about yourself for an hour.

Like I said, I go to a professional. In the same way I don’t try and do plumbing or electrics myself. The consequences of not using a professional are potentially catastrophic.

But my therapist’s schedule is occasionally interrupted by - you know - holidays and other trivial stuff that means she’s unable to sit with me and hear me talk about myself for an hour. Devastating. For her, that is - I hear myself talking all the time.

So I turned to AI this week to help me tease out some thinking on a particularly tricky work problem I have at the moment. After 8,000 words on the topic - back and forth - I had a brainwave.

“Help me condense my usual long ‘checklist mentality’ that I use to approach my sessions.” I said. “I end up coming away feeling like the session isn’t a success, simply because I didn’t cover everything on my list. Use all the data you’ve got from our conversation.”

Good old GPT 5.6 Terra on High got right down to business, and came up with some damned good ideas, and helped me frame some of my recent discussions in context:

• not treating the session as a checklist, in fact not going with a list at all
• how sitting in discomfort is not a task, but an exercise that is meant to be repeated
• and a host of other things, besides.

This is not to say that the AI tool has replaced my therapy. Like any good AI, it becomes a useful tool that - when steered in the right direction - can help and guide and reframe, rather than solving a problem.

There are still minor issues, but the better the frontier models get, the better they are at catching my patterns and adapting to suit.

That - alongside the exercise of posting every weekday on the accursed hellsite packed with AI and useless tips - has boosted my confidence and given me an insight into the workings of my own brain that I feel more equipped to explore with the guidance of a bloody good professional brain examiner.

So, as with most things, when used properly, with respect and adapation to the shortcomings, AI becomes a useful tool that can be applied to a host of challenges; giving me an expert for a hugely reduced fee.

When OpenAI and Anthropic stop subsidising the tokens we all use, we are all going to be screwed.

And suddenly way dumber.