Navigating Noise

Navigating Noise

By Filter LabsNewsPolitics
Download on the App Store

Navigating Noise episodes

  • Why Influence Operations Still Can't Prove They Work, with Carrick Longley @ ZenithFlow

    For two decades, the information operations community has sold activity. Messages disseminated, impressions served, reach and click rates. Carrick Longley spent his career inside that business and came to a blunt conclusion: almost nobody can point to a single decision that changed.


    Longley is the founder and CEO of ZenithFlow and a former Marine Corps major who spent years in signals intelligence, electronic warfare, and information operations planning before moving into industry. He holds a PhD in Information Sciences from the Naval Postgraduate School, where his dissertation examined how he and a small team pushed a field data collection tool through a change-resistant Department of Defense. He is also a vendor in this space, and he asked to be held to the same standard he holds everyone else to.


    Erol takes him up on it. The conversation runs from baselines and attribution to why counter-narratives can amplify the thing they're meant to counter, why the risk of inaction almost never makes it into a decision brief, and what it would take for defense contracts to pay for outcomes instead of hours.


    In this episode:

    - Why the adage that it's easier to drop a bomb than to send a message still holds

    - How stacked selection bias makes most social data useless for measurement

    - The case for judging every operation, kinetic or not, by its cognitive effect

    - Why information professionals have never had a real training range, and what Range is meant to fix

    - What outcome-based contracting would require from both sides of the table

    - How a self-taught engineer with a ham radio license built a 21-person company that ships the same day a customer asks

    Subscribe to Navigating Noise for new episodes every week.

    Available on YouTube, Spotify, and Apple Podcasts.

    Timestamps:

    00:00 Intro

    01:23 A true or false to start

    02:32 The 244-page dissertation almost nobody has read

    06:01 Two decades of selling activity instead of outcomes

    07:38 Impressions, reach, click rates, and what they miss

    10:32 Define the objective first

    11:39 Stacked selection bias in social data

    14:03 The risk of inaction nobody briefs

    14:52 When a counter-narrative amplifies what it counters

    17:37 Every action has a cognitive effect

    19:34 Information forces never had a training range

    22:26 Building Range

    24:44 Inside the synthetic environment

    26:29 Where the name ZenithFlow came from

    29:49 Teaching himself to code in order to teach it

    31:53 From one engineer to twenty-one people

    35:15 Operational context as the real moat

    38:36 Paying for outcomes instead of hours

    42:21 The billboard question

    45 min
  • How to Turn a Noisy Internet Into Actionable Signals for Irregular Warfare (with Jonathan and Erol)

    One of these founders spent his twenties reading Augustine of Hippo in the original Latin, watching how a theologian made sense of an empire falling apart. The other spent his career in Ramadi, South Sudan and Anbar, then wrote about how Augustus Caesar built an empire using the oldest propaganda machine in history. They run the same company. This episode is the story of why that's not a coincidence.


    Erol Yayboke and Jonathan Teubner co-host Navigating Noise and co-build Filter Labs, where the bet is simple and strange: the way information moves through a community is just as real a form of power as ships, bombs or currency, and almost nobody is reading it well. To open Season 2, they answer listener questions about the company's origin, their sharpest disagreement, and what they each learned approaching the same problem from opposite ends of history.


    Along the way: a rooftop promise that became a two-week deadline, a defense of "boring policy," a critique of airport-book intellectualism in tech, and the data frustration that started in a house in Iraq.


    In this episode:


    - Why a book on Augustine of Hippo and a piece on Augustus Caesar point at the same idea about power

    - How Filter Labs actually started in 2022, the night Putin cut off Western social media

    - The "founder hubris" moment: promising a demo before telling the co-founder

    - Why Jonathan thinks much of tech's "big ideas" culture is faux intellectualism

    - The data mistake Erol watched the US make in Ramadi, and what he'd fix

    - Why religious and "arcane" language is real power that policymakers keep ignoring

    - What Season 2 is really about, and the serious voices you won't see on cable news


    Subscribe to Navigating Noise so you don't miss Season 2.

    Listen on YouTube, Spotify and Apple Podcasts.


    Timestamps:


    00:00 Intro

    01:59 Season 2, and why this one is a solo episode

    02:36 Listener question: two founders, two men named Augustus

    03:12 Augustine of Hippo and how he thought about conflict

    06:24 Augustus Caesar, early pioneer of propaganda

    07:17 Coins as a distribution network

    07:46 The faux intellectualism problem in tech

    10:03 Why you need irrational people chasing big ideas

    11:54 From reading Latin to a machine learning model on religious violence

    12:40 Five million words, Oklahoma, and a contested authority

    15:07 Lahore, and why religious language is real power

    16:13 The border ceremony outside Lahore

    17:52 Ramadi, post-surge and pre-ISIS

    18:59 Two data sets, and who really decided where the money went

    20:06 Where his frustration with data began

    21:30 The voices he would have prioritized

    23:01 The night Putin cut off Western social media

    25:45 A rooftop promise and a two-week demo deadline

    27:12 What Season 2 is, and what it isn't

    31:45 The one thing they actually disagree on

    36 min
  • How to Turn a Noisy Internet Into Actionable Signals for Irregular Warfare (with Jonathan and Erol)

    One of these founders spent his twenties reading Augustine of Hippo in the original Latin, watching how a theologian made sense of an empire falling apart. The other spent his career in Ramadi, South Sudan and Anbar, then wrote about how Augustus Caesar built an empire using the oldest propaganda machine in history. They run the same company. This episode is the story of why that's not a coincidence.


    Erol Yayboke and Jonathan Teubner co-host Navigating Noise and co-build Filter Labs, where the bet is simple and strange: the way information moves through a community is just as real a form of power as ships, bombs or currency, and almost nobody is reading it well. To open Season 2, they answer listener questions about the company's origin, their sharpest disagreement, and what they each learned approaching the same problem from opposite ends of history.


    Along the way: a rooftop promise that became a two-week deadline, a defense of "boring policy," a critique of airport-book intellectualism in tech, and the data frustration that started in a house in Iraq.


    In this episode:

    • Why a book on Augustine of Hippo and a piece on Augustus Caesar point at the same idea about power

    • How Filter Labs actually started in 2022, the night Putin cut off Western social media

    • The "founder hubris" moment: promising a demo before telling the co-founder

    • Why Jonathan thinks much of tech's "big ideas" culture is faux intellectualism

    • The data mistake Erol watched the US make in Ramadi, and what he'd fix

    • Why religious and "arcane" language is real power that policymakers keep ignoring

    • What Season 2 is really about, and the serious voices you won't see on cable news


    Subscribe to Navigating Noise so you don't miss Season 2.

    Listen on YouTube, Spotify and Apple Podcasts.

    36 min
  • The Influence Operation That Never Sent a Single Message | Stephen Hunnewell

    Navigating Noise is a weekly interview show about how the world's most critical decisions get made, and what happens when the data behind them breaks down. Hosts Jonathan Teubner and Erol Yayboke talk to the people working at the edge of intelligence, influence, and strategic competition.

    Stephen Hunnewell is a Senior Principal within MITRE's Strategic Non-Kinetic Effects Mission Center, where he focuses on strategic influence and economic statecraft. Before MITRE, he served two tours in Afghanistan, first as an infantry officer, then as a civil affairs officer in the Pech River Valley, one of the deadliest stretches of the war.

    In this episode, Stephen reconstructs an operation the Army called Crop Diversification, one that got left out of the definitive book on the valley: a program that paid Afghan farmers to stop growing the dense, unmanaged cornfields the Taliban used for cover and ambush positions. He points out partway through that he never once said the word "psyop." Jonathan presses him on what "influence" actually means once you strip away messaging, and the two get into why the discipline keeps sliding back into silos inside the U.S. system that's supposed to run it.


    In this episode:

      Operation Crop Diversification: paying Afghan farmers to stop growing the cornfields the Taliban ambushed from
      How he got through the entire story without once saying the word psyop
      Money as a weapon system, and what economic statecraft looked like before anyone called it that
      Night letters, and why Taliban retaliation against villagers became the measure of effectiveness
      Why influence became a dirty word inside the system that runs on it
      What the OSS had that the US has never rebuilt, and why nobody wants to be the propaganda czar
      How defense and influence work went from unfundable to fashionable


    Subscribe to Navigating Noise for new episodes every week. Watch on YouTube, or listen on Spotify and Apple Podcasts.


    44 min
  • How to Fill an Information Vacuum Before Your Enemy Does with Kristian Ljungberg

    On February 4, 2025, ten people were murdered at a school in Örebro, the deadliest mass shooting in Sweden's history. Then came a silence that the internet filled for days. Kristian Ljungberg holds national responsibility for how the Swedish Police communicate on days like that, and his read on what went wrong is not what you'd expect from a serving official: the first error was a decision on an org chart. The agency treated a national trauma as a regional incident, which meant no national chain of command and nobody authorized to speak for the country while rumor took the field.

    Before the police, Kristian ran crisis communication at the Forsmark and Ringhals nuclear plants and in healthcare. Three industries, one problem: what you say in the first minutes when people are terrified and the facts aren't in. He co-wrote Sweden's 2026 national guideline on joint communication during hybrid incidents, which makes him one of the few people who has actually written down how a democracy should talk to its citizens while under attack.

    Jonathan Teubner sits down with him on the sidelines of the Riga StratCom Dialogue for a conversation about the trade-off at the center of the job: getting it right versus getting it out, and what happens to a country when the silence becomes the battlefield.


    In this episode:

      Why the "golden hour" of crisis communication is now, in his words, five minutes
      How the Swedish Police published their first message on the 2017 Stockholm terror attack in three minutes
      The one time Kristian deliberately made a story harder to understand, and sent reporters an 800-page Interpol protocol
      What actually happened inside the Örebro information vacuum, and the decision he calls a strategic mistake
      Why the police eventually released footage from inside the school, blood on the floor, and why the criticism stopped after they did
      How he tells scared people getting it wrong apart from a deliberate influence operation, and why the response is nearly identical
      What Sweden means by "vaccinating" a population against disinformation, and why he thinks a skeptical public is a healthy one


    Subscribe to Navigating Noise for new episodes on how the world's most critical decisions get made, and what happens when the information behind them breaks down.

    Listen on YouTube, Spotify, and Apple Podcasts.

    More on Filter Labs: filterlabs.ai

    39 min
  • Why People Believe AI Videos Even When They Know They're Fake

    In May, Ukrainians woke up to an avalanche of posts about hundreds of thousands of migrants flooding into their country, exempt from the draft, replacing Ukrainians, even renaming their cities. Influencers made TikToks about it. Public broadcasters ran explainers debunking it. Within four days it was the number one topic in the country. None of it was true.

    Roman Osadchuk is Head of Threat Intelligence at LetsData, where he investigates information operations, and he has co-authored research on Russian influence campaigns with the Atlantic Council's DFRLab. He joins Jonathan in Riga to walk through what he actually sees in the data: leaked Russian internal documents with content quotas and performance dashboards, cloned news sites that check your IP address before letting you in, and a strategy that no longer bothers trying to make you believe anything in particular.


    In this episode:

      How a fabricated migration wave became Ukraine's top story in four days
      Why "does it feel true" has replaced "is it true," and what that does to a society
      Inside Russia's leaked playbook: 300,000 comments a month, quotas, dashboards
      How doppelganger sites steal the credibility of the Washington Post and Fox News
      Why Russia's own doctrine has no word for "offensive," only "counter campaigns"
      What AI genuinely changed in influence operations, and what it still can't do
      The one question Roman says protects you better than any media literacy course


    Subscribe to Navigating Noise for new episodes every week. Watch on YouTube, or listen on Spotify and Apple Podcasts.

    More on Roman's work at LetsData: letsdata.ai

    48 min
  • How to Tell What's Real Online in the Age of AI Bots | Tetiana Haiduchyk

    For 6 years, Tetiana Haiduchyk has run a controlled experiment most people never hear about. Her team goes onto the open market, pays a few hundred euros for fake social media engagement, and measures whether the world's biggest platforms can detect and remove it. Mostly, they can't.

    Tetiana is Head of Analytics at Trementum, the small Kyiv data-science firm that serves as the analytical engine behind NATO StratCom's flagship "Social Media Manipulation for Sale" experiments. This year's run generated more than 100,000 interactions from over 30,000 inauthentic accounts across seven platforms, for about €252. She does this work from inside Ukraine, where the information operations she studies target the country she lives in. That makes her the rare guest who is the show's protagonist made real: the analyst under pressure, testing a contaminated information environment, in a country at war.

    This conversation goes inside that job. How the market for inauthentic engagement actually works, why the bots stopped shouting and started blending into real conversations, and why a trained human's judgment is the filter that still holds up.


    Here are the biggest highlights:

      How the experiment buys engagement from commercial vendors, and why the payments run through crypto
      Why "spotting the bot" is over now that they comment like real people
      Which platform removed 82% of the fake accounts, and which removed almost none
      What following the crypto revealed about who runs the market for inauthentic engagement
      Why speed favors adversaries who ignore their own rules, and what open societies can do about it
      How she taught her own mother to check a source before trusting it
      What it takes to "feel alive" under constant threat


    Subscribe to Navigating Noise for the real story behind information warfare. Watch on YouTube. Listen on Spotify and Apple Podcasts.

    44 min
  • Why the US Keeps Losing the Wars It Spends the Most to Win | Dr. Joshua “Bugsy” Segal

    For ten years, Dr. Joshua "Bugsy" Segal's job was certainty. As a US arms-control negotiator and on-site weapons inspector (the Chemical Weapons Convention, UNSCOM in Iraq), he built the machinery that proves whether a country is lying about its weapons. Then, one month after 9/11, he walked out of the Pentagon, put on a Navy uniform, and spent two decades in the one domain with no inspectors, no treaty, and no way to verify anything: the information environment.

    In this episode of Navigating Noise, Segal tells hosts Jonathan Teubner and Erol why America is losing the information war, and it isn't for lack of money or tools. It's that the US never built a narrative, never defined its objectives, and therefore has no way to measure whether any of it works. He backs it up with a four-star's private grade of the US effort against China ("abysmal"), a working theory of when Americans lost their herd immunity to Russian disinformation, and a blunt read on why Ukraine, fighting for its life, out-adapts the most expensive military on earth.

    This is a rare conversation with someone senior enough to say what analysts and info-ops practitioners think every day but rarely hear out loud.


    In this episode:

      Why a weapons inspector says you can verify a centrifuge but you can't inspect a narrative
      The four-star official who privately graded America "abysmal" against China
      How Russia "poured gasoline on both sides" of America's political divide
      Why strategic communication is "90% deeds, 10% words" and the smoke plume that proved it
      What Ukraine's cheap, fast innovation exposes about US spending
      The single word Segal would put on a billboard for billions to see


    Follow Navigating Noise so you never miss an episode. Available on YouTube, Spotify, and Apple Podcasts.

    55 min
  • NATO's StratCom Director Breaks Down the AI Information War | Jay Janzen

    In 2024, Jay Janzen told this show weaponized AI content was "minutes away." Two years later he's back with a blunt update: "We're definitely there." And the adversaries deploying it have no limits.

    Janzen is NATO's Strategic Communications Director and a former Canadian Brigadier-General, and he sits in the two-star meetings where military strategy and messaging now fuse. Recorded in Riga with Erol, this conversation gets into how NATO runs an information war without abandoning its values, why the alliance now treats deterrence as something that happens inside an adversary's head, and why the West's obsession with escalation is exactly the trap Russia wants us in.

    What you'll hear:

      Inside NATO's "shark tank" in Rennes, where 100 companies pitched agentic AI for cognitive operations
      Why Janzen rejects fighting disinformation with disinformation, even with one arm tied behind our back
      The three spaces where NATO denies Russia advantage so "the math never adds up"
      How the comms team was asked to shape Steadfast Deterrence 2027 three years in advance
      The Nazi propaganda myth Janzen believed for 30 years, and what it taught him about bias in AI
      What "good data" means when garbage in means garbage out at machine speed
      The one message he'd put on a billboard for billions to see


    Follow Navigating Noise so you never miss an episode. Available on YouTube, Spotify, and Apple Podcasts.

    Learn more about Filter Labs at filterlabs.ai

    40 min
  • Introducing Filter Labs’ New CEO: Why We Made the Transition and What’s Next

    On July 1, 2026, Jonathan Teubner walked out of the CEO seat at Filter Labs and Erol Yayboke sat down in it. This conversation was recorded that same day.

    Jonathan decided years ago he wouldn't be the lifetime CEO. The founder instincts that keep an early company alive — high risk tolerance, selling a product that doesn't exist yet, chasing every opportunity — become a liability at scale, and he wanted out of the seat before that turned existential. A recruiter told him he couldn't hire his way out of it. So he watched his successor grow into the job instead, without telling him it was a test.

    Erol lays out what he's keeping (the nerdiness, the flatness, the allergy to meetings) and what changes as Filter Labs becomes a Washington company selling into defense, intelligence, finance, and marketing. Then the conversation turns to the reason the company exists: Russia's move against Telegram just made a generation of single-source monitoring tools useless, and a decade of answering data quality problems with data volume is catching up with everyone.


    Highlights:

      Why Jonathan started planning his exit two years before it happened
      The exact moment a founder's risk appetite becomes a liability
      How Erol was tested for the CEO role without knowing it
      Life in the seat: five years without a single day fully off
      What the Telegram shutdown reveals about single-source intelligence tools
      Why "more data" stopped being the answer, and what good data looks like instead
      The Harvard Square DNA Filter Labs keeps as it becomes a DC company


    Subscribe to Navigating Noise for conversations on information, influence, and the data underneath both.

    Listen on YouTube, Spotify, and Apple Podcasts.

    Learn more about Filter Labs: filterlabs.ai

    39 min

About Navigating Noise

From the publisher's feed

Navigating Noise is a weekly interview show about how critical decisions get made when the data behind them breaks down. Hosts Jonathan Teubner, founder of AI intelligence company FilterLabs, and Erol…