You might consider adding some new things to do with your time so that the coding is not such a total aspect of your life.
I suffered chronic depression for decades, and the thing that no one ever told me is that there is no One Thing.
The exercise didn't fix it. The therapy didn't fix it. The meditation didn't fix it. The hobbies didn't fix it. It was Everything. When I'm not happy with my art, I do more exercise. When I'm not happy with my exercise, I do more meditation, etc. More social interactions, stronger networks, deeper support group. Eating better. Less social media. Less infotainment. More water. Everything. All the things.
A multitude of small changes has accumulated to a state in which I am now bothered by very little and can easily shrug off a wide variety of things that would previously have shut me down for hours or days.
After getting to know the thing beyond first impressions, you'll learn whether or not it actually turns your gears. If it doesn't, drop it and pick something else. Eventually, you will luck into something that excites you -- and it very well may be something that you never would have thought would excite you.
It's got the same kind of "how do I make what is currently nothing much in to something useful and or beautiful" problem solving, so it tickles the same kind of neurons, I suppose.
If the desire is the tinkering of old, find yourself something with bad documentation and figure it out, or dive in to something outside your typical realm: infosec, encryption, video encoding/decoding, etc. If the joy is the tinkering, do the tinkering yourself.
If the desire is "be employed in tech", then explore the realm of "agentic engineering". Spec-driven development, prompt engineering, context management, etc. There are still open questions and questions that are currently "solved" with less than optimal solutions. Alternatively, there are many projects that are refusing AI contribution, so any of those companies are places where you'd be on a level field with your coworkers, though I suspect many of them will still use LLMs for research, prototyping, and rubber-ducking.
There's a large gulf between "build me a web browser" and the amount of work that would go in to prompting to get something that competes in compatibility and performance with the current leaders.
But, there's two ways to use the tools at hand. Using it to do your work for you is leading to disillusionment. You might find that using it only for research (as opposed to manually digging through 8 year old reddit posts and poorly written man pages) is a better fit for you.
I was just thinking about the historical precedent for this pattern. Imagine you're a farmer when internal combustion engines are first becoming popular. Your neighbor has one and he planted twice the acreage you have in the same amount of time. If your love is the farming itself, you've an opportunity to acquire and learn the use of a tractor. If your love is the horses, you'll need to weigh that love against the financial incentive of the tractor. You might figure out that you can keep a horse or two just in case the tractor's down, or just for your love of horses. However, refusing the tractor outright will likely lead to your farmer acquiring your land in some future.
Art is another example. I know my way around a digital editor. I've seen some impressive generated results. I don't make any money from my art, so the decision is perhaps easier. I derive joy and calm from the manual process of dragging graphite across paper, and massive joy from the smiles my gifts produce. I'd argue that any such gift created by a prompt would never generate such a smile, and I'd get none of the meditative calm from the creation process.
This is the same. There's a new tool that's all the craze. You can either go all in, stick to your guns, or keep your favorites around for fun. You just have to decide what's right for you, and nevermind about the neighbors, because if it's not for you no amount of their enjoyment will transfer to you.