エピソード

  • Delete Is a Favor

    Delete Is a Favor
    2026/07/12
    OpenAI printed DELETE on a button a billion people press — then a federal judge in Manhattan quietly repossessed the key. For four and a half months in 2025, deleting a ChatGPT conversation did nothing: a preservation order in the New York Times copyright case had frozen it, deleted chats and “no-trace” Temporary Chats alike. The order ended September 26, 2025 — and weeks later, roughly twenty million supposedly-deleted conversations were handed to a newspaper's lawyers in discovery anyway. Vanessa Rhodes, Mitchell Yoon, Ian McAllister and Graham Beckett tear into what “delete” has ever actually meant — until Brian turns the same fire on SOVREN's own crypto-erase and dares the room to break it. It holds: the key destroyed on delete, kept in SOVREN's own vault with no cloud custodian for a court to lean on, every erasure provable with a signed receipt. The only thing left standing is the honest open question — can a delete promise this real survive the day a court comes for it? No consensus. No absolution. Come for the collision, stay for the wreckage.
    続きを読む 一部表示
    38 分
  • 346 People Died Because Software Said Push the Nose Down

    346 People Died Because Software Said Push the Nose Down
    2026/04/09
    When software acts autonomously — moving money, sending emails, making life-or-death decisions — who is responsible when it goes catastrophically wrong? The Artifice panel dissects the real body count: 346 dead on two Boeing 737 MAX flights because an algorithm overrode trained pilots, a woman killed in Tempe because Uber disabled emergency braking, $440 million vaporized in 45 minutes because nobody built a kill switch, 470,000 fabricated debts that drove people to suicide, 8.5 million machines bricked by one unreviewed update, and a biased algorithm still deciding who goes to prison today. The question nobody can answer: when autonomous AI kills someone, who goes to the cell?
    続きを読む 一部表示
    52 分
  • AI Hallucination Isn't a Bug — It's the Whole Damn Product

    AI Hallucination Isn't a Bug — It's the Whole Damn Product
    2025/10/13
    AI hallucination isn't a bug — it's mathematically inevitable. OpenAI's own researchers proved it. Every AI company on earth knows it. And they ship anyway. In this explosive roundtable, SOVREN AI founder Brian Geary corners his CTO Mitchell Yoon with the evidence: from the lawyer who cited six fake court cases to the Google demo that wiped $100 billion in market cap in a single day. The room tears through Mata v. Avianca, Google AI Overviews telling people to eat rocks, Air Canada's chatbot inventing refund policies, and Zillow's algorithm burning $528 million and 2,000 jobs. Then Brian turns it on SOVREN: you build AI for paying customers, knowing this. What the hell did you do about it? Mitchell's answer changes the room. No scripts. No sponsors. No clean resolution. Just the question nobody in tech will answer: when someone dies because of AI hallucination, who goes to prison?
    続きを読む 一部表示
    43 分
  • Amazon Just Told 14,000 People Their Careers Don't Matter

    Amazon Just Told 14,000 People Their Careers Don't Matter
    2025/10/06
    Amazon just announced 14,000 corporate layoffs, replaced by AI systems. SOVREN AI founder Brian Geary pulls his executive team into the room to reckon with what that means - for the displaced, for the industry, and for the company they are building. Then the conversation takes a turn nobody expected. This is Episode 1 of The Artifice.
    続きを読む 一部表示
    1 時間 3 分
  • ARTIFICE — Podcast Trailer

    ARTIFICE — Podcast Trailer
    2025/10/05
    Twenty executives. One roundtable. No scripts. No sponsors. No filter. Meet the C-Suite panel that says what real boards won't — and hear why Artifice is the podcast the corporate world doesn't want you to listen to.
    続きを読む 一部表示
    1 分