SYDNEY J. FREEDBERG JR.
In 2023, the Pentagon’s push for a global battle network, aka CJADC2, was all about going fast, fast, fast. In 2024, it’s going wide.
“Last year, we essentially went with what we had for the most part … to get this minimum viable capability,” said Air Force Col. Matthew “Nomad” Strohmeyer. A former fighter pilot, “Nomad” now oversees the quarterly Global Information Dominance Experiments (GIDE) that thrash out how to make CJADC2 work, not in the lab, but in real military headquarters. “This year, we really want to dramatically open the aperture and provide every opportunity for any industry partner, or other government programs… to be able to contribute.”
Likewise, he said, “we had a lot of success with [Britain] this year and we’re expanding that out to the other ‘Five Eye‘ partners this year.”
That expansion begins this week with GIDE 9, which is being interwoven with the latest Army-led mega-exercise/experiment/field test of new tech, Project Convergence 2024. Earlier GIDEs had brought in British observers, while the previous Project Convergence, held in 2022, welcomed the British and the Australians.
Bringing in the Brits was a humbling but helpful learning experience, Strohmeyer said. “In GIDE 6, which was last June-July … we were looking at the UK screens and [realizing], ‘Oh, you can’t see that data?’” he recalled. But by GIDE 8, he went on, “we were able to … successfully share data.”
This year’s conjoined GIDE 9/PC24 aims to deepen allied involvement, as well as include more participants from the various US services and private contractors.
Afterwards, GIDE staff will not only brief senior leaders in the Pentagon and on Capitol Hill, Strohmeyer said, but also hold an industry day for interested companies. The best way for aspiring contractors to connect to GIDE, he told a conference room full of industry representatives, is to use the Pentagon’s Tradewinds website.
CDAO Martell: ‘CJADC2 Belongs to Us’
Strohmeyer spoke last week at a conference hosted by his boss, Chief Digital and Artificial Intelligence Officer Craig Martell. As the Pentagon’s central hub for artificial intelligence and big data, Martell often emphasizes, CDAO’s job is not build AI in-house but to empower the rest of the Defense Department to build AI and data analytics tailored to their specific missions — a key component for the “joint” part of CJADC2. (Much of that empowerment, he emphasized, is the unglamorous work of cleaning up and cataloging hitherto incompatible databases).
“Our job is to make sure they have the tools, the whole department has the tools, to be successful,” Martell told reporters at the same CDAO conference where Strohmeyer briefed industry. That said, he emphasized, “CJADC2 belongs to us.”
So what does “CJADC2” stand for? As Martell broke it down: “C” for “combined” means with foreign and partners; “J” for “joint” means all US services; “AD” for “all domain” means across land, sea, air, outer space, and cyberspace; “C2” means “command and control.”
The military’s been doing “C2” forever, Martell emphasized, but technology means the scale and the means have changed.
“In Napoleon’s day, you’re a general standing on a hill, watching your troops, and you send messages,” he said. Today, both the “watching” and the “sending” parts are digital — and tremendously complex.
As early as fall 2015, then-Chairman of the Joint Chiefs of Staff Joe Dunford publicly called for strengthening the Joint Staff to keep better track of global threats that operated across the boundaries between regional Combatant Commands. At the time the concern was that the Pentagon’s “four-plus-one” threats — Russia, China, North Korea, Iran, and terrorists — increasingly exploited America’s artificial lines between Europe, Mideast, and Pacific, between war and peace.
The problem with Dunford’s “just add staff” proposal was that current bureaucratic processes can consume essentially unlimited amounts of human labor.
When the 2023 GIDEs tried pulling together global plans from regional commanders’ inputs, Strohmeyer said, “we discovered that… the way it currently happens is each of the combatant commands will aggregate up all sorts of different options … in briefing slides, in emails, in conversations, in video teleconferences. [Then] the Joint Staff has to aggregate all that…It takes hundreds of man-hours. It takes hundreds of emails.”
So between GIDE 6 and GIDE 7 last year, the team spent a lot of time developing new streamlined staff processes for collaboration between commands.
Even with streamlined processes, however, it takes more than human minds to manage the modern flood of data from spy satellites, scout drones, and other forces deployed across the planet. Just weeks ago, the four-star tapped to lead Indo-Pacific Command, Adm. Sam Paparo, said publicly that current intelligence analysis methods can’t keep up with an increasingly aggressive China. Instead, Paparo said, the best way to get “indications and warning” of an impending surprise attack would be to employ AI.
“Paparo’s comment is in line with what I was saying,” Martell told reporters. “Our job is to make the data flow to the right people at the right time, so that particular commanders can build the tools they they need to build … [e.g.] if he needs a particular indications and warning AI tool … [and] the way we’ve been figuring out what that looks like is through these GIDE experiments.”
Strohmeyer added: “One of the things that we have discovered is when we begin… aggregating this data together… we’re starting to really get insights” into what adversaries are doing.”
Pearl Harbor, The Blitz, The Need for Speed
The “minimum viable capability” version of CJADC2 is “real and ready now,” Deputy Defense Secretary Katherine Hicks declared proudly at the CDAO conference. But what does that mean?
It’s a spin on the software development term “minimum viable product,” Martell explained, but renamed because CJDAD2 is not a specific product but an aggregation of software code and staff procedures. The idea is what’s called “agile” development: Instead of seeking perfection, developers give a bare-bones prototype to real-world users to get their feedback fast, then rapidly fix the problems they identify, add new features they request, then throw the upgrade right back to the users for another round of experimentation, feedback, and improvements.
“It’s not going to be perfect,” Martell emphasized. “It’s not going to be the end of a long, five years spec/requirements process.”
But why can’t the Pentagon wait five years to get CJADC2 right, especially since it’s so complex?
“I get so much feedback of, ‘This is messy, Nomad, this is hard, can’t we slow down? Can’t we only do one GIDE a year or two guides a year?’” Strohmeyer told the industry briefing. “So I apologize in advance, if you’re one of the partners that we have [for] the pain that you’ve had that you’ve had to go through and the pain you’ll continue to go through this year.”
There are risks to going this fast, with numerous imperfections and frequent outright failures to fix. In fact, “I actually want you to fail because that’s where we identify the issues with what we’re doing and how we can better,” Strohmeyer told the contractors. “The Deputy Secretary of Defense has specifically told me… ‘Nomad, I’m going to ask you where you failed. And if you do not identify for me where you actually failed, you have failed.’”
The model here is not the typical painstaking Pentagon process but startups like SpaceX, he said, where the control room burst into cheers when a rocket exploded during its ascent because of all the data they had gathered. Likewise, he said, “the warfighting user at the edge … they’re actually quite risk tolerant, because they feel the day-to-day warfighting requirement every day and they know that they you know, are maybe challenge to do the job.”
The urgency stems less from the stunning speed of technological advances, though that’s definitely a factor, and more from the alarming degeneration of global stability.
“Think about great power competition,” he said, using the Pentagon’s understated term of art for the escalating Second Cold War with Russia and China.
“Probably none of us have really lived through the things that happened in the 1940s,” he told the room of rapt civilians. But on an earlier tour at Pacific Air Forces HQ in Hawaii, he recounted, “I spent about three years in PACAF headquarters, every day walking in and seeing the bullet holes on the wall from when it got strafed by Japanese Zeroes.”
“We found ourselves reeling suddenly from a massive naval loss at Pearl Harbor,” he continued. “We are not only ill-prepared … we didn’t have the concepts. We hadn’t thought through the new ways of warfighting.”
Strohmeyer also spent three years at Lakenheath airbase in England, imbibing Royal Air Force lessons from World War II. In 1940, the RAF was able to withstand the Nazi blitz — “barely,” he emphasized — “because a few of them had both the imagination to see that it was coming.” Those pre-war few took a new invention known as radar, married it to still-maturing technology of radio, and developed a system of command-and-control — the Dowding System — that enabled an outnumbered RAF to respond rapidly to Luftwaffe air raids.
“We live in a digitized age, an informationized age,” Strohmeyer said. “If we wait until we’re actually attacked [and] we find ourselves that next great war — If we wait until then it will likely be too late. And that is the urgency that drives what we are doing.”
No comments:
Post a Comment