We simply wrapped up our Hacking for Protection class at Stanford.
This was the eleventh 12 months we’ve taught Hacking for Protection, and the influence of uneven warfare, (drones, off-the-shelf applied sciences, and many others.,) disruptive applied sciences (AI, business entry to house) and a startup pleasant DoW acquisition system – make it really feel like a a lot completely different class than the earlier courses.(I’ll summarize a number of the learnings about the usage of AI on the finish of this put up.)
Hacking for Protection is now in 70 universities, together with 20+ within the UK – and this 12 months in Poland and Germany – with groups of scholars working to know and assist clear up nationwide safety issues.
This 12 months’s issues got here from the Navy, Air Drive, Military Analysis Lab, Protection Innovation Unit, IQT, and NASA.
This quarter 9 groups of 42 college students at Stanford collectively interviewed 1132 beneficiaries, stakeholders, necessities writers, program managers, business companions, and many others. – whereas concurrently constructing a sequence of AI-driven minimal viable merchandise and growing a path to deployment.
We opened this 12 months’s last shows session with a fantastic discuss AI and protection – previous, current and future – from (Ret) LTG Jack Shanahan. Jack was the Director of the DoD Joint Synthetic Intelligence Heart (JAIC). Watching his speak is a worthwhile use of your time.
In case you can’t see Jack Shanahan’s video click on right here
Throughout the quarter visitor audio system within the class included Owen West – director of the Protection Innovation Unit, Mike Brown – associate at Protect Capital, (Ret) LTG Joseph McGee latest head of the Joint Workers J5 (technique, plans, and coverage,) and Hon Marise Payne Australia’s Minister for Overseas Affairs.
“Classes Realized” Shows
Every of the eight groups gave a last “Classes Realized” presentation together with a 2-minute video to offer context about their drawback. Not like conventional demo days the place groups exhibit, “Right here’s how sensible I’m, and isn’t this a fantastic product, please give me cash,” the Classes Realized shows inform the story of every staff’s 10-week journey and hard-won studying and discovery. It’s a curler coaster narrative describing what occurs once they uncover that the whole lot they thought they knew on day one was mistaken and the way they ultimately acquired it proper.
Whereas all of the groups used the Mission Mannequin Canvas, Buyer Growth and AI instruments to construct Minimal Viable Merchandise, every of their journeys was distinctive.
This 12 months we had the groups add two new slides on the finish of their presentation: 1) inform us which AI instruments they used, and a pair of) their estimate of progress on the Expertise Readiness Degree and Funding Readiness Degree.
Right here’s how they did it and what they delivered.
Crew Noctua – Began with an issue that mentioned, “Particular operators can’t detect drones passively, with out exposing their place.” They ended up understanding {that a} bigger drawback was, “Dismounted troops and base defenders lack a passive means to offer early warning detection of all varieties of drones, together with these which can be RF silent.
In case you can’t see the Noctura video click on right here
In case you can’t see the Noctura presentation click on right here
These are “Depraved” Issues
Depraved issues refer to actually advanced issues, ones with a number of transferring components, the place the answer isn’t apparent and lacks a definitive method. Most issues our Hacking For Protection college students work on fall into this class. They’re typically ambiguous. They begin with an issue from a sponsor, and never solely is the answer unclear however determining methods to purchase and deploy additionally it is advanced. Most frequently college students discover that in hindsight the issue was a symptom of a extra fascinating and complicated drawback – and that Acquisition within the Dept of Warfare is in contrast to something within the business world. 
As an alternative of admiring issues from inside a classroom our college students get of the constructing and study, discovery and iterate.
The determine reveals the varieties of issues Hacking for Protection college students encounter, with the commonest ones shaded.
Crew SwarmShield – The preliminary drawback was framed as, the price of utilizing costly interceptors to shoot down low-cost drones. By the top of the category the Crew realized the issue was constructing terminal steering that lets an inexpensive, throwaway drone discover and hit an attacker at night time.
In case you can’t see the SwarmShield abstract video click on right here.
In case you can’t see the SwarmShield presentation click on right here
Division of Warfare Listing – This 12 months the scholars had entry to a Division of Warfare Listing – primarily a phonebook of ~5,700 names of “Who buys within the Dept of Warfare?” 
The listing features a tutorial on how the DoW buys and the assorted acquisition and funding processes and applications that exist for startups. It gives particulars on methods to promote to the DoW and the place the Program Acquistion Officers (PAEs) match into that course of.
Crew Weapons With out Wait – The preliminary drawback for this staff was “Retool and scale protection manufacturing capability to replenish vital munitions on the tempo required by sustained, high-intensity conflicts.” That is what I name a “boil the ocean” drawback” – large and huge – and obscure. By class finish the staff realized what was quickly achievable (and wanted) was inexpensive, licensed munitions for small drones produced on the point-of-need.
In case you can’t see the Weapons With out Wait video click on right here
In case you can’t see the Weapons With out Wait presentation click on right here
It Began With An Thought
Hacking for Protection is constructed on the identical methodology as Lean LaunchPad class I created at Stanford in 2011. It was adopted by the Nationwide Science Basis (NSF) because the NSF I-Corps (Innovation Corps) to coach Principal Investigators who needed an SBIR grant. Now in its second decade and in 100+ universities, I-Corps has change into a typical for science commercialization on the NSF, Nationwide Institutes of Well being and the Division of Vitality – coaching 3,251 groups and launching 1,400+ startups thus far.
Crew IonX – IonX additionally began with a “boil the ocean” drawback – The US wants a safe uncommon earth provide chain. They ended up with an issue extra tangible and deliverable – Mineral processors throughout markets can’t determine and take a look at higher chemical reagent schemes.
In case you can’t see the IonX video click on right here
In case you can’t see the IonX presentation click on right here
Origins Of Hacking For Protection
In 2016, brainstorming with Pete Newell of BMNT and Joe Felter at Stanford, we noticed that college students in our analysis universities had little connection to the issues their authorities was making an attempt to resolve. We realized the identical Lean LaunchPad/I-Corps class would offer a framework to take action. That 12 months we launched each Hacking for Protection and Hacking for Diplomacy (with Professor Jeremy Weinstein and the State Division) at Stanford.
Crew Cheese on the Moon – Began with a mandate to seek for mineral deposits on the moon. By class finish they realized that to try this lunar missions must know what’s on and below the moon not solely to mine, however to land.
In case you can’t see the Cheese on the Moon video click on right here
In case you can’t see the Cheese on the Moon presentation click on right here
Objectives for Hacking for Protection
A decade in the past, our objective for the category was to show college students Lean Innovation strategies whereas they engaged in nationwide public service. We needed to familiarize college students with the army as a occupation and assist them higher perceive its experience, and its function in society. We additionally hoped the category would present our sponsors a technique that builds drawback understanding earlier than writing necessities.
The category nonetheless does all this, however now that the DoW is shopping for from startups and protection enterprise capital is considerable, the category has changed into a nationwide safety incubator. Most of our groups kind protection firms.
Crew Gasoline Forge began with the issue that fight items must generate energy and gasoline regionally. They ended with a extra fascinating statement that they might build networked, on-site hydrogen nodes to gasoline drones in ahead, contested environments the place resupply is in danger,
In case you can’t see the Gasoline Forge video click on right here
In case you can’t see the Gasoline Forge presentation click on right here
Go-to-Market/Deployment Methods
The preliminary objective of the groups is to make sure they perceive the issue. The following step is to see if they’ll discover mission/answer match (the DoW equal of business product/market match.) However most significantly, the category teaches the groups concerning the troublesome and complicated path of getting an answer within the palms of a warfighter/beneficiary. Whereas the DoW has made large strides in reforming how and who they purchase from, college students nonetheless must know: Who writes the requirement? What’s an OTA? What’s colour of cash? What’s a Program Supervisor? Who owns the present contract?
Crew Luminarch – Began with Tactical items lack the aptitude to visualise, handle, and adapt to the electromagnetic spectrum in actual time. They ended with Tactical items lack low-cost, attritable RF sensors that may be deployed at scale, limiting their capability to detect threats, handle signatures, and talk.
In case you can’t see the Luminarch video click on right here
In case you can’t see the Luminarch presentation click on right here
Crew Tessellate– Began with the statement that drone missions don’t scale. And ended by realizing what’s lacking is US multi-drone doctrine doesn’t exist and present drone warfare adjustments are occurring sooner than the software program lifecycle.
In case you can’t see the Tessellate video click on right here
In case you can’t see the Tessellate presentation click on right here
AI Within the Class Room
AI has had some apparent and never so apparent impacts on our class.
First, right here’s a abstract of how our college students used AI in each courses I taught this quarter.
In case you can’t see the AI Use In Class slide click on right here
In case you can’t see the AI Rap Video click on right here
AI Instruments Used
Claude + Granola – had been the AI instruments utilized by everybody.
Giant Language Fashions Used
– Claude, Claude Code, Claude Chrome extension, Claude Cowork, Claude Design
– ChatGPT
– Gemini
Observe taking
– Granola
– Twinmind
Shows
– Perplexity
Constructing prototypes
– Replit
– Lovable
Creating Artificial Customers
– Hear Labs
– Viewpoints AI
Summarizing Analysis
– Google NotebookLM
– Notion + G Suite (not strictly AI, however used as a part of AI workflows)
Different
– Ultralytics YOLOv8 (utilized by the SwarmShield H4D staff for drone detection/monitoring MVP)
The apparent and optimistic adjustments of AI had been that groups had been capable of do buyer discovery extra effectively. The not so apparent change was that creating merchandise quickly allowed groups to make dangerous concepts go sooner.
Prior to now, MVPs had been an indication of a groups technical competence, however now spinning up one thing in hours that beforehand took weeks, implies that an MVP is now not proof of vital considering and speculation testing.
This meant scholar studying was unbalanced. A finished-looking product felt like success. College students confused a refined deliverable with the necessity to deeply perceive the wants of all of the stakeholders, in addition to the necessity for Buyer Validation. For protection startups meaning understanding a path to a CRADA, or to a analysis or manufacturing OTA. We wanted to sluggish the groups down. Going ahead we’ll have college students come into class with a prototype however subsequent time accompanied by the express hypotheses and experiments they’ll use to validate whether or not the prototype solved an precise drawback.
Extra about this in a separate weblog put up.
It Takes A Village
Whereas I authored this weblog put up, this class is a staff undertaking. The key sauce of the success of Hacking for Protection at Stanford is the extraordinary group of devoted volunteers supporting our college students in so many vital methods.
The educating staff consisted of myself and:
- Pete Newell, retired Military Colonel and ex Director of the Military’s Fast Equipping Drive, now CEO of BMNT.
- Joe Felter, retired Military Particular Forces Colonel; and former deputy assistant secretary of protection for South and Southeast Asia, and Oceania; at the moment Director of the Gordian Knot Heart for Nationwide Safety Innovation at Stanford which we co-founded in 2021.
- Steve Weinstein, associate at America’s Frontier Fund, 30-year veteran of Silicon Valley know-how firms and Hollywood media firms. Steve was CEO of MovieLabs, the joint R&D lab of all the key movement image studios.
- Chris Moran, Govt Director and Common Supervisor of Lockheed Martin Ventures; the enterprise capital funding arm of Lockheed Martin.
- Jeff Decker, a Stanford researcher specializing in dual-use analysis. Jeff served within the U.S. Military as a particular operations gentle infantry squad chief in Iraq and Afghanistan.
- Jillian Manus, a enterprise associate at Protect Capital and Senior U.S Enterprise Advisor for the European Innovation Council
Our educating assistants this 12 months had been: Evan John Twarog, Varsha Saravanan, Breno Casciello, and Luke Andrews.
34 Sponsors, Enterprise and Nationwide Safety Mentors
The groups had been assisted by sponsors and mentors.
Sponsors had been originators of the staff issues. They gave us their hardest nationwide safety issues: Owen West, Will Ryan, Phillip “Donna” Smith, Joel Uzarski, Alexandra Bissey, Mark Breier, Jonathan Inventory, Trent Emeneker, Matthew Anderson, Ana Alvarez, Jonathan Boltersdorf.
Nationwide Safety Mentors helped college students who got here into the category with no data of the Division of Warfare, perceive the complexity, intricacies and nuances of these organizations: Katie Tobin, Kelly McGannon, Rachel Costello, Henning Heine, Josh Edwards, Marco Romani, Tom Schmitz, David Vernal, Wealthy Lawson, Dan Ruttenber, Ashley Perry, Sophia Vahanvaty, Rick Lu, Chris O’Connor
Enterprise Mentors helped the groups perceive if their options may very well be a commercially profitable enterprise: Doug Seiche, Jeremy Schoos, Adam Waters,, Matt Croce, Isobel Porteous, Eric Byler, Diane Schrader, Donnie Hasseltine, Mark McVay.
Sponsoring Organizations: Gordian Knot Heart for Nationwide Safety Innovation, Widespread Mission Mission, Lockheed Martin, Boeing, BMNT, Protection Innovation Unit.
Due to all!
Filed below: Hacking For Protection, Instructing |
