Ian Thomas – Podcast Stage | Craft 2026 (Conversation Outline)

Listed as “Ian Thomas – Podcast Stage,” but the scheduled guest had a conflict, so this became an ABK / Leadership Anonymous “Leadership News” episode recorded live at Craft 2026 — a casual, news-driven roundtable (the hosts liken it to “having a beer and listening”). Hosts: Dávid Schneitoffer (“Kispócok” — startup alumnus, beginner surfer, engineering director), Gábor Náday (“Máté”/“Mefi” — community builder, lifelong blogger, engineering director), and Fanni / Karolina Tóth (cognitive scientist → high-performance IT coach → sports dive leader; “expert in psychedelics and reinventing myself”). The podcast (abkpodcast.hu, ~4 years) usually runs monthly in Hungarian; “Leadership News” is a newer, shorter weekly format (their longest regular episode — on hiring — ran ~3h40m). Topics, mostly AI, with a companion newsletter that lists every article referenced. Fanni’s role: “they bring the news, I bring the champagne.”

1. Anatomy of an AI-Native Org

1.1 The 30-year pyramid

• For 30 years software orgs were shaped like a pyramid, with most head count in the middle doing translation work — turning what the business wanted into a language “us nerds could understand so we can tell the computers what to do.”

1.2 Translation, not code, is what’s getting cheap

• Everyone talks about code/features getting cheap; the article’s real point is that the translation work is getting cheap — so the pyramid narrows in the middle.
• The layered model: why (top — why the company exists) → what (middle — what to achieve) → how (bottom); managers sat atop the “how,” translating between business and engineering.

1.3 The provocative line, and the pushback

• Quoted line: “If a manager isn’t contributing to the why, the what, or the trust system that holds the how, it’s hard to say why they’re doing it anymore.” — i.e., pure translation is “solved” by agents, so fewer such managers are needed.
• Dávid: it’s likely a “false prophecy” that managers vanish — their role and size change, but the persistent problems remain; a good manager already solved the translation and managed people.
• Fanni’s take: soon whoever wrote the article will realize you need managers for meaningful communication between people, teams, and projects.
• The manager’s real work isn’t tactical (engineers are good at tactical issues) — it’s collaboration, how teams work together, unclear business logic (“I’m biased, I’m a manager comforting myself with these sentences”).

1.4 People problem, not tools problem

• The article calls out managers denying the change.
• Recurring lesson: it’s usually a people problem, not a tools problem — don’t “fix” it with a double-blind, VPN-anonymized feedback system; teach people to have tough conversations and not save feedback “until it’s an apocalypse.”
• Fanni’s company’s twist: feedback is anonymous, but everyone adds ”feel free to share my name next to this feedback” — because knowing the source lets you actually talk about it.

2. iPhone Anti-Snatching & Hardware Lock-in

2.1 The feature

• Apple is developing an anti-snatching feature that locks stolen phones immediately — curiosity about how the instant lock triggers (handover? pocket?).

2.2 Why iPhones are already hard to resell

• Each component is cryptographically bound to the specific iPhone: swap a display (or even a cable) into another device and it won’t work unless a certified repair shop authorizes it via an admin panel.
• “That’s why they’re so expensive.” Skeptic’s note: “a night market here will solve that issue in a minute.”

2.3 Fanni’s stolen-iPhone recovery story

• Her iPhone 14 was stolen; from a friend’s car she tracked it via Find My iPhone to a sketchy-neighborhood mobile-repair shop.
• She saw a man leave with a phone matching hers (no case), confronted the shop owner, then chased the man down in the parking lot, stopped his car, and he admitted he “found it.”
• She recovered it; on restart it seemed fine, then started beeping and announcing it was stolen and belonged to “Fanni Karolina,” demanding return to the owner — confirming iPhones are genuinely hard to get rid of.

3. Meta Glasses vs. the Instagram Chatbot Exploit

3.1 The delight (Meta glasses)

• At a coffee shop (Dorado, 7th District, Budapest) she tried a stranger’s Meta glasses: asked the weather (a 10-sentence humidity answer), tried Hungarian (it could, until the owner’s phone dropped Wi-Fi), and loved that it whispered background music only she could hear — “finally, background music to my life… we live in the future.”

3.2 The dread (the exploit)

• The same week, Instagram accounts were hijacked by social-engineering Meta’s support chatbot (cheaper than human support): talk to the bot long enough that it believes you own the account and grants access — even disabling MFA and sending codes to an attacker-stated email.
• “This is not the future — this is the present” (actually the recent past — the bug was fixed a few days prior). Hackers are always a step ahead of whatever Apple/Meta build.

3.3 The no-human-support gripe

• With a bank you can still fight through a chatbot to reach a human; with Google/Meta/Netflix you never will — they’re too big and/or fired the staff (“here’s the AI chatbot that gives away your Instagram instead”).
• Aside: a former Meta moderator reported being badly understaffed, and most quit after seeing terrible content — “AI will solve that too.”

4. Enthusiasts vs. Skeptics (Charity Majors)

• Charity Majors (blog; conference speaker — likely at Compass not Craft) argues AI enthusiasts and skeptics on engineering teams are both responding to genuine existential threats but have stopped talking to each other, creating a silo.
• The line: “AI enthusiasts are in a race against time; AI skeptics are in a race against entropy.”
• The fix: sit down and talk, and share the whole story — including the bad news (people over-share AI wins). Dávid follows @ctitzen (a strong AI critic) alongside the fanboy side to keep balance.
• Meta-skill: hold two difficult/opposing truths at once — be enthusiastic about AI and know it’s imperfect; disagree with someone and still love/respect them. Dávid: this is what Hungarians most need to learn after 1.5 decades of “them vs. us → they’re stupid” polarization (cf. Android-vs-iOS fights).
• Adopts Kent Beck’s phrase: the “unpredictable genie” that “will solve everything in an unpredictable, non-deterministic way.”

5. Curiosity, the Inner Child & Intuition

5.1 Intelligence ≠ emotion ≠ humanity

• We’re teaching AI “everything we know,” forgetting intelligence isn’t emotion or humanity.
• Humans are wired for simplistic friend/foe thinking (survival: “can I trust this person or should I draw my knife?“) — the untrusting survived and reproduced, so we’re engineered for “good/bad, like/don’t like.” Dávid: simplifying is necessary given how much we face — but Fanni urges making a space in your head that understands it’s not that simple, and getting curious.

5.2 Curiosity as the driving force

• Recurring theme of their conversations: curiosity should drive software development, personal development, AI development — and personal relationships (“the death of love is when you stop being curious about your partner and start expecting them to behave a certain way”).

5.3 “Kids just know”

• Children are unbothered and in tune with intuition — they know they’re too hot for long pants, or when they’re full — and “parents make them forget those things.”
• Dávid’s realization: his belief that he “can’t tolerate cold” traces to childhood memories of parents forcing long pants (“no, it’ll be too cold”) while he walked home sweating.
• Reconnect with your inner child / sense of amazement: Fanni canceled her gym membership to jump rope at the playground and watched a snail cross the street with delighted two-year-olds (“I didn’t realize snails were so fast”). Go outside, spend time in nature.

5.4 The Roost “slow social” app

• Roost (icon: a pigeon) is a “slow social media”: you send a bird carrying a message, and it arrives at the speed that bird actually travels (based on real locations).
• From it, Dávid learned a snail travels ~2 km/h (a “rain” snail message) — so a message 10 km away takes ~5 hours, “surprisingly fast.”

6. “What to Do When You’re Not Detail-Oriented”

• A newsletter article: “Captain Obvious,” but valuable for being hands-on/specific rather than high-level.
• Method: identify the skills you’re bad at by noticing dread — e.g., the “weird pain in your stomach around Sunday 4pm” before a Monday presentation.
• Then deliberately do those uncomfortable things frequently to get out of your comfort zone, so they stop blocking your opportunities.
• Fanni: listen to your body — those symptoms signal what’s happening. Quotes: “You don’t hate Mondays, you hate your job” and “sorry for being late, I didn’t want to come.”
• Aside: they were handed a signed Hungarian edition of Gergely Orosz’s The Software Engineer’s Guidebook (“don’t buy books — become podcasters and receive them for free”).

7. Data-Center Water Use

• Google/Meta/Amazon are trying to fix (and PR-spin) data-center water consumption (Wired, The Verge, a Hacker News thread).
• A recent framing suggests the water issue was over-hyped and less severe than believed — met with skepticism (“okay, who paid for that study?”).
• Drinking water is becoming “the most sensitive resource”; desalination (making seawater drinkable) has gone from “no way to do that” 10 years ago to being energy-/cost-bound now. “Technology will solve it” — maybe.
• Slides into the recurring “capitalist section” — big tech pouring money to improve public opinion “is always what’s happening.”

People & References Cited

• Hosts: Dávid Schneitoffer (“Kispócok”), Gábor Náday (“Máté”/“Mefi”), Fanni / Karolina Tóth — ABK Podcast / Leadership Anonymous.
• Ian Thomas — the scheduled-but-absent guest (name only).
• Charity Majors — “enthusiasts race against time; skeptics race against entropy.”
• Kent Beck — the “unpredictable genie.”
• Gergely Orosz — The Software Engineer’s Guidebook (Hungarian edition, signed).
• @ctitzen — an AI-critical voice Dávid follows for balance.
• Companies/products: Apple/iPhone (anti-snatching, part binding), Meta (glasses, Instagram support-chatbot exploit), Google/Meta/Amazon (data-center water), Netflix, Instagram, ChatGPT, SimCity 3000 (an off-topic “SimCity 3K in 4K” patch), GTA San Andreas remaster (AI-mangled assets), Roost (slow-social app), Find My iPhone.
• Concepts: AI-native org / translation layer, why/what/how pyramid, people-vs-tools problems, hardware DRM / part pairing, social-engineering chatbots, enthusiast-vs-skeptic silo, holding opposing truths, curiosity, inner child / intuition, noticing dread to find weak skills, desalination, scaling (“the cake shop” analogy; the Tokyo subway NFC-chip 12ms story).

Video: https://www.youtube.com/watch?v=wPx8adZqo9o — Transcript via yt-transcript.sh; outline generated from the transcript.