The military has a dozen front doors for commercial technology, and a good dual-use company gets to learn all of them. It pitches the Defense Innovation Unit (DIU), then AFWERX, then NavalX, then the Army Applications Lab (AAL) — re-telling the same story each time, rewriting its paperwork to fit each door's intake, forking its codebase to survive each one's security review.

Then it runs out of money, because none of that was the work. We built the front doors to make commercial adoption faster. Collectively, we made it a full-time job — and the companies that can least afford a full-time job of paperwork are the ones with the technology we need most.

BLUF: The Software Weapons & Tactics (WepTac) Conference builds the range where code gets fight-tested. Its sibling, the Dual-Use and Innovation WepTac, does for the deal what that one does for the code — but it runs like a real Weapons and Tactics conference, not a live-fire exercise. You don't invent your best tactics on the range; you develop and ratify them at WepTac, and then the whole force flies to them. Do the same for acquisition: a quarterly, DIU-chaired conference where practitioners from the innovation organizations, contracting, the data office, and the program offices develop and ratify binding doctrine — one open intake, two rights menus, reciprocity by evidence, and the ordering guides that move the money. It keeps two doors open on purpose: small business at low maturity through Small Business Innovation Research (SBIR), and the commercial heavyweights at high maturity through the Commercial Solutions Opening (CSO) and Other Transactions (OTs) DIU pioneered.

The front doors don't talk to each other

Each innovation organization was built to be a fast lane, and each is — into its own cul-de-sac. Separate intakes, separate paperwork, a forked codebase as the cover charge, security settled with a binder instead of evidence. Clear one and you get no credit at the next. What you get instead is pitch theater: whoever tells the best story to the right senior leader wins a demo, and the warfighter waits.

The deepest version of the failure is data. When a program optimizes for a platform instead of a data contract — the Project Maven mistake — the government ends up renting access to its own information, locked to one vendor's tools. Do it enough times and "modernization" is a collection of moats you paid to dig. The companies notice. So does their capital.

A weapon stuck in the acquisition system is a weapon the enemy never has to face.

A WepTac, not a Red Flag

Here's the distinction the Pentagon keeps blurring. A Red Flag is a live-fire exercise — you fly the mission, graded, against a real adversary. A Weapons and Tactics conference is something else entirely: the deliberate forum where weapons officers and subject-matter experts take hard lessons, develop the tactics, argue them, stress them, and ratify the ones that survive into the manuals the whole force then flies. The Software WepTac is the live range for code. This one is the tactics conference for the deal — and the deal is doctrine, not a sortie you fly under a shot clock.

So run it like one. Organize the quarterly conference into mission-area working groups — interface and data, rights and risk, contracting and money, reciprocity and security — each staffed by the people who actually do the work: the Software Design and Development Supervisors (SDDSs) who own the mission threads; agreements officers (AO); the data office responsible for the Joint All-Domain Command & Control (JADC2) data schemas, zero trust architecture (ZTA) and service level agreements (SLAs) — specifically the Super Intelligence Office (SIO); and the innovation organizations (DIU, AFWERX, NavalX, AAL, et. al.).

Each group brings proposals — the acquisition equivalent of a tactics-improvement proposal — and each proposal gets argued hard and red-teamed on the ugly cases: a data-ownership clause that collides with a coalition export rule; a critical vulnerability in a shared library that an inherited authorization has to absorb with no bespoke paperwork. What survives gets ratified into binding doctrine; what doesn't goes back. And what comes out is never a slide — it's an intake rubric, two rights menus, a reciprocity standard, and a joint ordering guide a squadron can use the next week, with a named team to carry each one into a real contract vehicle within ten business days. No artifact or it didn't happen.

Two open doors

That bet is the whole reason to keep the doors open. The value of an open topic — and of the open CSO DIU built — is that industry gets to propose the thing nobody wrote a requirement for, and the government pivots to it. The moment you narrow intake to a pre-published interface or a platform wish list, you've closed the open door and turned the fast lane back into a shop that only buys what it already imagined. The SDDS who owns the mission thread is the one who recognizes the unexpected solution and decides to pivot to it — that judgment is exactly what the career path exists to supply.

Keep both doors open, too, because dual-use shows up at two very different maturities. A small company with a low technology readiness level (TRL) comes through SBIR, where a competitively won award is the ticket to a Phase III vehicle that can scale it. A commercial heavyweight — the hyperscalers and frontier labs, the Googles and Amazons most of the economy already runs on — arrives with a high-TRL product that was never going to fit a small-business research pipeline, and it comes through the CSO and the OT rails DIU pioneered precisely to buy mature commercial capability fast. Different companies, different maturities, same conference — and the same rights, reciprocity, and ordering doctrine apply to both, so neither has to fork its code to fit the other's lane.

Interface conformance still matters — but as the transition discipline, not the bouncer at the door. Once a capability is picked, mapping it to a mission-thread interface, a service-level objective (SLO), and a contract vehicle is how it scales across the force and stays portable instead of becoming the next moat. That discipline belongs at the point of transition, after the government has pivoted to the idea — not as a filter that strangles the idea before anyone sees it.

Rights as a menu — and why the venture market is watching

Settle rights on a menu, up front, instead of as a fight the government loses after award. Two menus do it. The Data Rights Menu defaults to Government Purpose Rights (GPR), tool-agnostic and data-first, with export and portability baked in: the government can always get its own data out, in the published schema, and move it to another provider. The License Rights Menu settles the software question honestly — anything a government software factory builds is unlimited GPR; anything dual-use is bought as a commercial license where the vendor keeps its intellectual property (IP) and the government owns the data that software generates.

That second menu is where most reformers flinch, and it's the most important line in the post. Letting a dual-use vendor keep its IP looks like a giveaway. It is the opposite. The commercial license is the thing that leaves the company something to sell on the open market, and that commercial upside is exactly what a venture investor underwrites. Strip the IP and you strip the exit; strip the exit and the venture capital (VC) that funds the entire dual-use base — the private research budget that now dwarfs the Department's own — prices in the risk and routes around defense, taking the best technology with it. Buy the license and own the data and you do the reverse: you keep the venture market paying to invent your next advantage, and you still hold the one asset that compounds, which is the data. You never again pay a prime for a license to a problem only the government has, and you never again let a commercial vendor hold your data hostage to keep your business. Buy the license; own the data; stay free to leave.

Bring your code once

The third piece of doctrine is the one the post is named for. Reciprocity runs on evidence, not paper. Clear an accredited commercial-onboarding platform or cloud service once, present the pipeline evidence — a software bill of materials (SBOM), attestations, test results — and that continuous authorization to operate (cATO) carries across every innovation organization's task orders automatically. Bring your code once, inherit the accreditation, and compete on the mission instead of on who can survive the paperwork a fourth time. That single rule is what turns a dozen cul-de-sacs back into one road.

Buy effects, not artifacts

Aligned rights and inherited accreditation are worthless if the money can't move, so the conference rewires the buy too. Outcome contract line items (CLINs) pay when a mission metric moves, not when a milestone slips. A decentralized, sole-source Phase III vehicle lets any contracting officer order a proven capability without re-competing it. And budget-activity agility — the subject of a later post on how the money machinery actually works — lets a commander shift dollars onto a thread the week it goes green instead of waiting two years for the next budget cycle. The through-line from a hard-won award to a fielded effect stops being a cliff and becomes a ramp.

One output matters as much as any artifact: a market signal. The conference tells VC and the primes what the next four quarters will actually buy — by interface and effect, not by brand. The investors in the room don't pitch; they red-team the ordering guide and tell the government which contract terms will unlock capital. Dual-use only works if private capital stays in the game, and capital stays when it can see real, credible demand instead of a fog of one-off pitches and fiscal-year surprises.

The objections, answered

A reform that touches intake, rights, and money at once draws objections from every direction. Good — here they are, with the answers, because the mechanics live in the answers.

"Open topic is just chaos — you can't buy what you didn't ask for." You can, and the disciplined version of it is the entire point of a WepTac. Open topic isn't the absence of a standard; it's a bet that the best idea isn't on your requirements list yet, run through a forum built to recognize it, stress it, and ratify it. The SDDS owns the pivot; the working group ratifies the doctrine. That's structure — it's just structure aimed at recognizing good ideas instead of pre-approving them.

"Letting vendors keep their IP is a giveaway." Demanding the IP is the giveaway to the enemy. Insist on owning every line of a dual-use company's code and you kill its commercial exit, its investors walk, and you end up paying full freight to build in-house what the market would have funded for you, while the best content gets acquired and deployed faster by our near peers. Own the data instead — the asset that actually compounds — and let the vendor keep the code it can sell elsewhere. The alternative gets you neither the leverage, the living supplier base, nor an advantage over the enemy.

"Reciprocity lowers the security bar." It raises it. A binder of control narratives tells you what a vendor promised a year ago; an SBOM, real-time pipeline attestations, and an inherited cATO tell you what the code is doing now. Evidence is a higher bar than paperwork — and the only one that scales across a dozen doors.

"This favors the big commercial players over small business." Two doors, not one. The SBIR lane exists precisely to protect the low-maturity small-business entrant; the CSO and OT lane just stops forcing a mature commercial company through a small-business research pipeline it was never built for. Nobody gets crowded out; each comes in on the terms that fit what it actually is.

"It's just another conference." A conference produces slides. A WepTac produces ratified doctrine and a named team that installs it in ten days. If an intake rubric, a rights menu, or an ordering guide can't survive the working group's hardest case, it doesn't get signed — and if nothing gets signed, nothing happened. That is the opposite of a conference.

The fix

Run the Dual-Use and Innovation WepTac quarterly, DIU-chaired, paired with the Software WepTac that feeds it the interfaces. Govern it on the "two dads" split this series treats in full later: DIU owns the standards, the reciprocity, and the oversight; the services own the dollars and the billets. Keep both doors open — SBIR for the low-TRL small business, the CSO and OT rails for the high-TRL commercial heavyweight — and use both open-topic, so the force can still be surprised by a good idea and a pre-planned group of requirements aligned methods to evaluate. Publish the intake rubric, the two rights menus, the reciprocity standard, and a joint ordering guide, and name a team to wire each one into a real vehicle within ten business days. It kills pitch theater, forked codebases, paper security, and accidental data lock-in in a single exercise — and it does it without a new law.

What you can do Monday

  • Unit commanders: Name your priority mission thread and put an SDDS on it who can recognize a novel fix when industry brings one. The clearer your thread and the sharper your integrator, the faster an open-topic solution reaches you.
  • PEOs / PAEs: Adopt the conference's ratified artifacts and the joint ordering guide instead of standing up one more bespoke intake, and treat an existing Phase III vehicle as a landing zone, not a threat.
  • Contracting officers: Default to the Data Rights Menu — GPR, government owns the data, exportable in the published schema — write outcome CLINs on decentralized Phase III vehicles, and buy the commercial license, never the lock-in. Honor reciprocity: if a company cleared an accredited pipeline once, don't make it re-prove security by binder.
  • DIU and the service innovation organizations: Chair it, and run one open, interface-aware intake across all of you instead of a dozen platform wish lists. Keep the open-topic lane and the open CSO genuinely open — the whole value is being surprised by what industry brings.
  • OSD: Sign the "two dads" agreement so standards and money stop fighting, and make budget-activity agility real enough that a commander can fund a green thread the same week — not the same biennium.
  • Congress: No new authority needed. Ask two questions of every innovation dollar: do the front doors share one open intake, and when the contract ends, does the government own its data and can it switch vendors? If the answer is no, you're paying to rent your own information — and to scare off the capital that would have funded the rest.