I've been trying to write a post for several days on tokenonics and it's annoying me beyond belief because I just can't get to the 'so what?'.
It's not often I have to abandon a piece or completely overhaul, but when it happens, it's usually because I am bored (and you will be too) and that's not good enough.
As a rookie, it used to happen on FT features sometimes. You would get a brief for a piece, run your interviews and then write it according to the brief.
But you'd realise there was a substory that was more interesting - or someone said something that flipped the brief's assumptions on its head.
Plans are the same. For those people organised enough to make one plan and stick to it - it often becomes clear the plan isn't fit for purpose.
At these moments, it's so important to step back, make a cup of tea and think. Reorder your thoughts and go through cycles of 'so whats' or 'whys?'.
Your answers come by not treading the same path again and again - but looking at the path from all angles - and then walking away for a bit.
So the point I was trying to make about AI tokens is very simple:
Visible cost is not the same as real value. The moment work becomes measurable, the market starts pricing the measurable part and stops paying for the rest - until the rest turns out to be the thing that mattered.
Invisible value or waste is about to become more visible.
Elon's $1tn+ IPO suggests Anthropic could follow. Speculators suggest AI token prices will then rise significantly.
People are measured on cost vs outcome rate - but more and more people will be asked to commit to outcome-based delivery at a fixed price.
AI won't be judged this way until all of its tasks can be lined up in a supermarket and seen clearly with a label.
Until then, you - the buyer - will take the risk as to whether it will deliver the outcome this turn or next turn.
A bit like the one-armed bandit.
That is what I was trying to say. Sorry it took about three years to get there.
Have a great weekend.
Dan
