01
What question is this site answering?
Whether you should spend thirty minutes with Bud about the next piece of work you’ll have to sign. Not whether to buy. That decision has its own conversation, after the call.
What we’d want you to see before the call
The state of the work: re:compound, as of September 2026.
We ask every producer to send their work across with the state behind it: what it rests on, who owns what, what’s still open. This page is ours. It’s written for the partner deciding whether to take thirty minutes, for the colleague they’ve sent to check, and for whatever AI they’ve asked to read this site for them.
02
What does the claim rest on?
Native
You can inspect it. Eleven organizations started, nine completed the ten weeks, nine of nine extended, two in flight, as of September 2026. Two quotes, verbatim: a managing partner in mid-market private equity on the problem, a managing director in mid-market advisory on the result. Ten published essays and a coined term with a public definition. The founders’ careers, verifiable on LinkedIn.
Relayed
What partners tell us after ten weeks; we haven’t measured it. The outputs make sense again. They know who owns what. They re-enter a matter or a deal at altitude. Their judgment carries past one decision and can be read. They see when an assumption has stopped being safe.
Provisional
Not yet tested. Every installation so far came through a prior relationship with Bud. That record can’t tell you how the offer lands with a partner who doesn’t know him. What it can tell you is that every firm that finished extended, first to the rest of the partner’s team and then to other partners, and that decision was made after seeing it run, not on his word.
Judgment
Ours. That two defaults in human–AI collaboration are close to fatal for serious knowledge work: the model developing the frame it’s handed instead of testing it, and the first fluent draft becoming the thing everyone edits. And that a checker at the end cannot be the whole answer.
03
Who stands behind what?
Bud Bhattacharyya: the method, the model-side operating assets, the public thinking, and every installation. Elissa Stewart: operations and practice; every installation is scoped through her desk and held to the week-ten bar. We are alternately reviewer and producer on each other’s work. re:compound LLC. Two people, no software.
04
What are the live alternatives?
- Keep doing what you’re doing. The tools stay fast; the partner keeps reconstructing or signing blind.
- Buy more speed. More tools, more agents, more output rolled into the same unreadable thing.
- Hire a consultancy for an AI program. Broader, longer, and it usually stays.
- Install this in one team and see whether the pen gets lighter in ten weeks.
We think the last is the right first move for a partner who signs. This site is the argument. The alternatives are real.
05
Where do we want your judgment in the call?
Whether the work you have in mind is consequential enough to show whether this took hold. Whether the team is the right first team. Whether your firm’s approved AI environment is one the assets can be configured in. Whether ten weeks from now is the right time. You’ll know the answers; we won’t.
06
What are we watching?
Whether partners who don’t know Bud recognize themselves in the first two sentences of this site. Whether the first cold conversations ask “what does this actually involve,” which would mean we haven’t answered the process objection, or ask “when.” We’ll update this page when we know.
A practical first step
Thirty minutes about the next piece of work you’ll have to sign.
Who’s bringing it, when, and whether that’s the team to start with. No deck, no demo, nothing to prepare.