When the Algo Tries to Be a Journalist

When the Algo Makes Insane News Feel Normal

This week on Kill Your Algo, the guys try to figure out what the internet wanted them to believe: Trump assassination-attempt coverage, political violence narratives, gerrymandering, Iran war messaging, UFO apathy, and the increasingly weird problem of deciding whether anything online is real anymore. Jeff starts by fighting the machines at work, Bill brings fact-mode energy to viral political claims, and Bob investigates the missing/dead nuclear scientist narrative only to find that the algorithm may be better at connecting dots than proving patterns. Plus: Iran Lego propaganda, AI videos, cyanide bombs in the American West, billionaire discourse, puppy yoga, and a coyote mistaken for a lost dog. Show Notes In this episode: * Jeff opens with life inside the AI/code machine: vibe coding, AI slop, and whether more machine output actually means better work. * The hosts compare how their feeds framed the Trump assassination-attempt story: serious threat, staged narrative, Secret Service failure, media apathy, or conspiracy object. * Tweet of the Week includes Spencer Pratt as an unlikely LA political character, Iran Lego propaganda, and a viral charity-statistics fight. * Bill reviews his fact-mode homework and how political stories get distorted by timeline jumps, implication, and partisan framing. * The guys dig into corruption, insider trading, Congress, and how the algorithm points outrage at the other team instead of the system. * Gerrymandering becomes a case study in how both parties frame power-grabbing as democracy protection. * Bob breaks down the missing nuclear scientist narrative and asks whether it is really a pattern — or just modern amplification doing what it does. * Conspiracy Corner covers AI realism, UFO apathy, whisper campaigns, and cyanide bombs in the American West. * Doggie Day Care brings the needed reset: puppy yoga, husky howling, and one almost-rescued “dog” that turned out to be a coyote. * This week’s homework: figure out how to trade feeds or compare feeds with someone else for a week.

When the Algo Lets You Spike the Football

S1 Ep8: When the Algo Lets You Spike the Football What happens when your feed doesn’t just inform you—but hands you something to celebrate? This week on Kill Your Algo, we run a live experiment: what happens when you deliberately retrain your feed with 20+ new follows? The result isn’t just new content—it’s a shift in tone, confidence, and what feels true. We break down how algorithms don’t just show you information… they reinforce it. From “spiking the football” moments to strange, unverified stories that feel important but lack confirmation, we test where signal ends and bias begins. * A real-time experiment manipulating the algorithm * How quickly your feed adapts—and what it prioritizes * Why “spiking the football” content dominates engagement * The feedback loop: liking → more of the same → stronger belief * Stories that show up in your feed but nowhere else * How different feeds create completely different realities If your feed keeps making you feel right… how would you know if you’re wrong? We decide the next experiment live on the show. Want to participate? Try this: * Follow 15–25 accounts outside your normal viewpoint * Engage (like, comment, watch) for 2–3 days * Track what changes in your feed * Notice what starts to feel “true” Your feed isn’t neutral. It’s trained—by you. If this episode made you think differently about your feed, share it with someone who lives in a completely different one. Kill your algo.

When You Train the Algo

What happens when you start nudging the system—and it nudges back? In this episode of Kill Your Algo, we talk about a live experiment: can you push an AI or a feed to move in a direction you choose? Not by hacking it—but by subtly changing your inputs. Same prompts. Slightly different framing. Watch how the responses shift. Then we take it a step further. Because this isn’t just about AI—it’s about your feed. Every like, follow, pause, and comment is a signal. Over time, those signals shape what you see, what gets amplified, and eventually, what feels true. The question isn’t just what is the algorithm showing you? It’s what are you training it to show you? This week’s experiment: Step outside your normal feed. Follow voices from a different perspective. Engage—lightly, intentionally—and observe what changes. Your recommendations. Your “For You.” Your sense of what’s trending. We’ll do the same—and compare results next episode. Kill Your Algo.

When the Algo Shows You Both Sides… But No One Tells You What’s True

Can AI actually cut through bias, misinformation, and media narratives—or does it just reorganize them? This week, we tested it. We built prompts designed to turn AI into an evidence-first journalist—one that separates fact from inference, flags weak sourcing, and avoids partisan framing. What we found was useful—but incomplete. AI can surface facts. It can show you both sides. It can even highlight what’s missing. But it can’t tell you what’s true. That part still belongs to you. * AI is a powerful signal filter, not a truth engine * “Both sides” ≠ clarity — it can actually delay conclusions * Verifiable facts are easy; truth requires interpretation * Most media optimizes for engagement, not accuracy * Even with better tools, human effort is still required 0:00 — Cold open: “AI won’t take a side” 0:20 — The experiment: can AI act like a journalist? 6:30 — The problem: regurgitation vs insight 12:00 — Fact vs truth (Indiana Jones reference) 20:00 — Can AI actually resolve ambiguity? 28:00 — The real issue: people don’t want to do the work 33:00 — What your algorithm is actually feeding you 36:00+ — Tweets of the week + narrative breakdown Copy this into your AI tool and paste in any article: You are an evidence-first news analyst. Analyze the article or text below using the following framework: 1. Identify all verifiable facts (objective, checkable claims). 2. Separate facts from: 3. Evaluate the strength of each claim: 4. Flag: 5. Do NOT assume any claim is true without evidence. 6. Do NOT provide opinions. 7. Avoid partisan framing. Output format: * Verifiable Facts * Inferences * Speculation * Source Credibility Assessment * Missing Context / Red Flags Copy a full news article Paste it into your AI Run the prompt You’ll get a structured breakdown of: * What’s actually known * What’s assumed * What’s weak or missing This does NOT give you truth. It gives you cleaner inputs. You still have to decide what those inputs mean. AI can cut through the noise and surface the signal. But meaning doesn’t come from the machine. That still belongs to us. 🧠 Key Takeaways⏱️ Timestamps (approx)🧪 Bob’s “TB Mode” Prompt (Try This Yourself)How to Use It⚠️ ImportantFinal Thought

When the Algo Tries to Be a Journalist

In this episode of Kill Your Algo, Jeff, Bill, and Robert explore a big question: could an AI ever act like a journalist? As algorithms increasingly shape what we see in our feeds, the idea of an AI that can separate facts from opinions — and flag claims that might not be entirely true — becomes more than just a thought experiment. The hosts dive into narratives showing up in their own algorithms this week, share their Tweets of the Week, and unpack how different information ecosystems can shape the way we see the same story. Then they explore the central experiment of the episode: what would it take to build an AI that behaves like a journalist with integrity? No conclusions yet — just a curious attempt to understand the systems influencing us all. Homework for next week: Each host will attempt to create a prompt that turns an AI into a journalist-style fact checker — something that separates facts, opinions, and questionable claims. If you try the experiment too, send your prompt to: killyouralgo@proton.me The best ones may be featured in next week’s show. AI journalism algorithm bias media literacy AI and truth news algorithms information ecosystems AI fact checking AI ethics Kill Your Algo podcast