Something I noticed digging through Newton's current status: behind all the talk about agent marketplaces and "swarms" of composable bots, the thing that's actually live right now is a lot simpler. A Recurring Buy agent. That's it, so far.
Not knocking it — a working, policy-checked recurring-buy agent running on real infrastructure is worth more than a roadmap slide. But it's a useful reality check when you see the bigger pitch (an onchain marketplace where you discover and orchestrate whole swarms of agents). That's still ahead, not here.
I think this is actually the right order to build in. Ship one boring, narrow thing that works, prove the authorization layer holds up on something simple, then expand. The alternative — launching the whole marketplace vision before the base layer's proven — is how a lot of "verifiable automation" pitches end up being neither.
Just don't confuse the roadmap deck with what you can actually go use today.
@NewtonProtocol $NEWT #Newt
Poll: Which approach do you trust more from an infra project?
Ship something small and narrow first, expand later
Launch the full vision early, iterate in public
Not knocking it — a working, policy-checked recurring-buy agent running on real infrastructure is worth more than a roadmap slide. But it's a useful reality check when you see the bigger pitch (an onchain marketplace where you discover and orchestrate whole swarms of agents). That's still ahead, not here.
I think this is actually the right order to build in. Ship one boring, narrow thing that works, prove the authorization layer holds up on something simple, then expand. The alternative — launching the whole marketplace vision before the base layer's proven — is how a lot of "verifiable automation" pitches end up being neither.
Just don't confuse the roadmap deck with what you can actually go use today.
@NewtonProtocol $NEWT #Newt
Poll: Which approach do you trust more from an infra project?
Ship something small and narrow first, expand later
Launch the full vision early, iterate in public