Cheeky Pint: Garrett Langley of Flock Safety on Building Technology to Solve Crime

Podcast: Cheeky Pint
Host: John Collison (Stripe cofounder, referred to as “A”)
Guest: Garrett Langley (Flock Safety founder, referred to as “B”)
Date: March 5, 2026

Episode Overview

John Collison sits down with Garrett Langley, founder of Flock Safety, for an in-depth conversation about building a technology company to solve crime. Over pints, they cover Flock’s founding story, technical innovations, their explosive growth (now covering over half the U.S. population), and deep insights into modern law enforcement, privacy debates, drones, crime fighting technology, and challenges in hardware entrepreneurship. The tone is candid, humorous, and highly informative, providing unique access to the day-to-day realities and moral questions underlying modern safety infrastructure.

Key Topics & Insights

1. Flock’s Origin Story & Founding Motivation

• Initial Spark: Langley started Flock Safety in 2017 after a gun was stolen from his neighborhood, and police were unable to help due to a lack of actionable intelligence—specifically, a missing license plate.
• First Solution: Built a “camera scout” with a college friend to track neighborhood vehicles and identify non-residents.
• Breakthrough Moment ([02:15]): After a similar crime, data from their cameras led directly to an arrest and gun recovery:

  "That car gets put on what would be called Ebola. ...hours later, they find that vehicle, the gun's in the car, the person goes to jail." (B, 02:15)

• Media as Growth Channel: Early growth came organically from local news coverage of crimes solved with their system, prior to any real business strategy.

2. Evolution of the Product

• From Neighborhoods to Nationwide Networks: Flock started by serving legally incorporated neighborhood associations, then evolved to city-wide deployments and integrations with broader law enforcement databases.
• Community-Centric Approach:

  "You had to build a safety system for a community... every single security system was focused on the individual." (B, 04:44)

• Expanded Capabilities:
  • Real-time 911 call analysis using AI
  • City-wide camera integration (including private cameras)
  • Advanced search (e.g., identify people by attributes: "white Converse sneakers")
  • Drones for real-time aerial support

3. By the Numbers: Scale and Impact

• Current Reach:
  • Over 6,000 U.S. cities (about 50%+ of American population, excluding Manhattan)
  • ~$500 million in Annual Recurring Revenue (ARR)
  • More than one million crimes cleared or arrests aided last year
• Breadth of Product: Cameras, drones, "Flock OS" for software integration
• Customer Mix: Law enforcement, municipalities, Fortune 100+ corporates (retailers, healthcare, logistics), with $100M+ ARR now from private sector

4. Law Enforcement Structure & Coordination

• U.S. Uniqueness:
  • Over 17,000 localized police agencies (vs. single national force in most countries)
  • Historically poor inter-agency data sharing; cloud adoption only legalized in many states in recent years (e.g., Florida in 2022).
• Flock’s Unlock: Platform enables effective multi-jurisdictional operations (e.g., human trafficking busts across 4 states).
• Coordination Tension: Skepticism and legal boundaries exist around federal vs. local cooperation:

  "People have this big distrust... places like California, there is state legislation that law enforcement is not allowed to collaborate with federal authorities..." (B, 13:20)

5. Technical Challenges & Solutions

a. Hardware & Networks

• Self-powered, self-contained cameras for deployment where there is no power/fiber
• Edge computation: Must operate on low power (no GPU), solar-powered, with 5G backhaul
• Precision: Able to photograph and analyze cars at 100+ mph

b. Real-Time Crime Response

• 911 Call AI: Parses live calls, surfaces actionable leads instantly (“amplified intelligence”—leveraging LLMs to deliver data to officers)

• Multi-modal search: Search recent footage for attributes (e.g., “white Converse sneakers”) ([05:32])

• Drones: Rapid response (e.g., reducing 911 response time from 7.5 minutes to under 70 seconds); used in pursuits, surveillance, search & rescue

• Notable Moment:

  "In this case, this person was from a 911 call to an arrest in about 17 minutes." (B, 07:10)

c. Crime and Data Ecosystem

• Stolen Vehicle DB Integration: Real-time sync with both national (FBI-maintained NCIC database) and local “hot lists”; local updates happen instantly, national still faces a ~24-hour lag due to archaic CSV-over-FTP workflows.

  "It will take 24 hours to make it to the FBI hot list. And that's a CSV file that gets sent around on FTP servers across the US." (B, 17:13)

• Anomaly detection: Identifies “cold plating” (stolen plates), simultaneous presence of vehicles in different locations, misaligned make/model vs. plate—flagged as potentially criminal activity

6. Privacy, Accountability, and Controversy

• Perceived vs. Actual Privacy

  "Because we operate in the physical world, we are held to a higher standard..." (B, 21:00)

• Data Auditing and Retention:

  • Strict logs of access, full auditability, short-term data retention by default
  • Limit data lifetime (7-30 days) unless extended via authorized procedure (e.g., warrants)

• Societal Debate: Understandable privacy concerns—Langley advocates for more cameras with more severe use restrictions, not less surveillance ([82:03]).

• Legislative Best Practices: Praises states like Virginia/New Mexico for clear, enforceable laws mandating audits, retention limits, and defining authorized use ([83:36])

7. Tech Trends, Offense-Defense, and Future Challenges

• Emerging Crime and Technology:

  • Cell phone records (CDRs) critical in crime resolution; criminals rarely leave them behind.
  • New “asymmetric warfare”: Gangs/cartels using drones for casing homes, smuggling into prisons.
  • Drone countermeasures for law enforcement are hamstrung by FAA restrictions ([26:47]).
  • Criminal sophistication shifting toward organized cargo/distribution theft—e.g., using legitimate business fronts for fraud ([41:38]).
  • Sharp increase, then decline, in violent crime attributed to COVID-19's psychological and social disruptions ([32:37]).

• Policing Technology:

  • Movement away from subjective, bias-prone policing toward objective, data-driven patrols (e.g., focusing on stolen vehicles rather than suspicious individuals).

8. Business Challenges and Hardware Manufacturing

• Scaling Pains, Product Line Diversification:

  • Built too many hardware products/segmented too broadly, leading to difficult capex and forecasting issues.
  • “Every decision you make in hardware is millions of dollars… in software, there are always two-way doors. In hardware, it’s always a one-way door.” (B, 73:12)

• Supply Chain Complexity:

  • Competes for components (e.g., capacitors, memory) even with giants like Apple, compounding design and procurement challenges.

• Maintenance and Field Ops:

  • Onsite field service teams; predictive, preemptive replacement schedules (main failure: IR cut filter—one moving part in the camera).

• Humorous Note:

  "A third of our employees dig holes for a living. AI is a long ways away from replacing that." (B, 53:26)

9. Philosophy, Social Impact, and the Future

• Prevention Over Prosecution:

  • Flock’s long-term vision is fewer crimes and fewer people going to prison, investing in “thriving cities” (capital for new jobs initiatives in safe municipalities).
  • Support for alternative, non-carceral responses for non-violent, opportunistic first-time offenders.

  “Failure for Flock is prison population goes up. ...The whole concept of prison will never work. ...What could Flock be doing to ...prevent kids from becoming criminals?” (B, 100:51 and after)

• Expanding Internationally:

  • Significant nuance in law enforcement and infrastructure globally. Economic competition (especially with state-subsidized Chinese companies) creates barriers outside of the U.S.

Notable Quotes & Memorable Moments

• "No one really steals a car for fun... You steal a car to go do something bad." (B, 17:52)
• "The majority of criminals, let’s call it 99%, are not evil people. They're not like evil is a random act of violence and that is exceptionally rare. It's all opportunism." (B, 39:49)
• "Prison can't be the answer. It just doesn't work." (B, 104:01)
• "One of the things the government did figure out—maybe you need to have an RFP for a $200 million contract. What about a $10,000 contract? ...But no, it's tough. I would not wish upon anyone selling to local government. It's more negatives than positives in a lot of ways." (B, 87:26)
• On unintended consequences:

  "Historical precedent gets thrown away when things are much, much cheaper and much, much easier." (B, 99:46)

Important Timestamps

• [00:24] – Atlanta founding story, crime catalyst
• [02:15] – First neighborhood deployment → first solved crime
• [05:21] – Description of Flock's current products and national footprint
• [09:52] – By the numbers: cities, ARR, market share
• [15:16] – Deep dive: technical challenges of camera deployment
• [17:13] – Data pipeline & 24-hour lag to FBI stolen vehicle database
• [21:00] – Privacy, data retention, accountability discussion
• [23:54] – Cell phone data and its role in solving crime
• [26:44] – Drones, criminal use vs. law enforcement limitations
• [32:07] – Shift from bias-prone policing to data-driven objectivity
• [37:11] – Social media’s role in criminal culture & recruitment
• [41:38] – Shift from retail theft to distribution theft
• [44:07] – Private sector segment: scale, use cases, employee safety
• [53:26] – Hardware as competitive moat ("third of employees dig holes")
• [72:01] – Hardware lessons: forecasting, supply chain
• [80:13] – Balancing surveillance with civil liberties
• [83:36] – Legislative frameworks for responsible tech use
• [100:51] – Mission focus: prevention, not just prosecution

Closing Thoughts

The episode gives an unusually transparent and technically authoritative look behind the curtain at one of America’s most consequential public safety technology companies. Garrett Langley and John Collison cover tactical, technical, and moral aspects, sparking a nuanced discussion on crime, technology, privacy, and societal tradeoffs. Langley's optimism about prevention and nuanced realism about the complexities of surveillance make this a must-listen for those interested in crime, cities, and civic innovation.

For more, listen to the full Cheeky Pint episode, or visit Flock Safety and Stripe.