Today, Santosh Arron, founder and CEO of Blankline, issued an executive order to the company.
Effective immediately, Blankline's major R&D effort and its most important researchers are redirected into Project Hope. That is now the priority.
Dropstone continues. Project Yaazh continues. The science continues. But Hope moves to the top, and we are reorganizing the entire company around it. This is not a pivot and it is not a reaction. It is a decision made from a position of strength, after results we are still working to fully verify alongside researchers from MIT, Stanford, and leading institutions across the world.
This post explains what led to the order.
The agent that prompted it
The order followed a single experience.
Project Yaazh brought Dropstone, our resident agent, closer to true autonomy than anything we have built, and Arron tested it on himself. He gave it access to his email, his calendar, and his contacts, and watched it work. It performed an autonomous callback, deciding in real time what needed to happen and what to do next. It found his brother in an address book where no contact is labeled "brother." It mapped his schedule. It understood his entire account, and it stayed precisely within the access it was given.
"That moment stopped me," Arron said. "Not because it failed. Because it worked so well that I had to ask the only question that matters: where does this go?"
That question is the reason for everything that follows. An agent that capable, released to the world at full strength with no restraint, is not a product. It is a liability with a launch date.
Why we are limiting it on purpose
We are equally serious about what happens when this capability reaches people's hands at full capacity. So we are limiting it, deliberately.
Today, Dropstone can place calls only with AI disclosure and a built-in delay, so the person on the other end always knows they are speaking with an AI. We achieved low latency internally, then deliberately tuned the system to be recognizably AI, on purpose, because a machine that can pass for a human on the phone is a capability the world is not ready to receive from anyone, including us.
In our adversarial testing, the agent made zero errors. Not a low error rate. Zero. That is why we are rolling it out to maximum users first, rather than widely and all at once, so people can see exactly how it works, at exactly the pace we have set. The capability is finished long before the responsibility is, and we are pacing to the responsibility.
India's responsibility
This is also why the order exists at all.
India's responsibility is not just to build AI. It is to build it right. India is no longer only an AI consumer. We are defining a new category, globally: systems that are commercially practical, genuinely safe, and genuinely intelligent, all at once. This is precisely what Sridhar Vembu has always believed India could do, and it is the standard we hold ourselves to.
Building it right is not a slogan. It is a set of constraints we accept on purpose, including constraints that slow us down and cost us reach. Hope is being built with proper safety at its core, from the first line rather than as a layer added at the end. It is also why we are choosing investors very carefully. Without the right judgment, without the right values behind the capital, this is the kind of work where humanity starts to come apart.
What the order is responding to
We want to be plain about the state of our research, because the headline does not capture it.
We are standing somewhere very few people thought we could reach. We found the first universal bimodal drift rate. We are building across Dropstone CLI and Dropstone Agents. We did original scientific research on wormholes. We are working on Primus. We are working on Hope. None of this is surface level. Every piece of it goes all the way to bedrock.
And here is the part that carried the most weight in the decision. This research is not accelerating at 10x. It is accelerating at 100x. "It is getting harder for me to verify the results it is producing," Arron said. We are in constant contact with researchers from MIT, Stanford, and top institutions across the world, not as a courtesy, but because we need their help to keep pace with what we are seeing. When the people who built the system bring in the best minds available simply to check the work, that is a signal worth reorganizing around.
What happens now
The order stands. Hope is the priority, and the company is being restructured around it.
We will not make promises and we will not share timelines. What we will say is this. We are taking it seriously enough that we are reorganizing the entire company around it. That should tell you something.
Prepare yourself.
Project Hope was authorized by executive order of Santosh Arron, founder and CEO of Blankline. Dropstone and Project Yaazh continue under existing teams.