Mythos and Marshmallows
Ben Thompson, discussing Anthropic's Safety Superpower:
To that end, it has long been clear to me that the frontier labs have the economic imperative to move closer to the user. If you own the user touchpoint, then you have meaningful lock-in, and the best way to own the user touchpoint is to be the canvas for everything they need to do. This, by extension, means that the frontier labs are on a collision course with software companies: it’s software that owns the user touchpoint, and it’s in the frontier labs’ long-term interest to not simply be a commodity input into software but to simply replace software outright.[...]
Note also the virtuous cycle with moving up into user touchpoints: the more workflows that are done directly with Claude or Codex, the more data each company gets to feed back into their training, which makes their products that much more capable and useful, expanding the number of workflows they can serve, expanding their access to data.
[...]
What if, however, the companies that give in to Anthropic’s data policies get better results right now? Or what if existing companies resist, leaving the door open for new companies — or the model makers themselves — to outcompete them in the market?
This feels like a marshmallow test for companies.
Delay gratification and risk getting left behind. Or eat the marshmallow now and surrender your data leverage.
Except the marshmallow test was always messier than the myth.
It was not simply a test of discipline. It was a test of trust, context and circumstance.
Do you believe the second marshmallow is really coming? Do you believe the adult will keep their promise? Do you believe the room is safe enough to wait? Can you afford to wait?
That feels like the real enterprise AI question.
Not: should we adopt this tool? But: do we trust the company trying to become our canvas?
There is no clean answer. Eating the marshmallow might be the rational move. Waiting might be the strategic one. Either choice can make you brilliant or foolish in hindsight.
It depends on where you're headed and what you are protecting.
Market demand? Distribution opportunity? Speed? Capability? Customer intimacy? Institutional memory? Data leverage? The right to remain your own canvas?
Because once the frontier lab owns the workflow, your usage becomes part of its compounding advantage.
That is the sliding door. Not adoption versus abstinence but leverage now versus leverage later.