Start free trial

Five minutes.
AI, made clearer.

Ask who did it, on what, and how many Macs.

Read the briefing

Members read every story in the archive, re-verified. Sources and corrections stay open to everyone.

What the pen marks mean
  • rememberThe takeaway: what to remember
  • figureThe number that matters
  • evidenceThe evidence line
  • rulingThe ruling
  • claimThe claim that fails

Each mark is drawn once, as you reach it. Numbers are only circled when they come from the story’s own figures.

A Swedish TV season's finished edit really vanished, and the premiere moved to 2027. The rogue AI agent behind it comes from unnamed sources, and TV4 says the human factor, not AI.

Reporters point at a confused small robot while a technician behind them holds a cracked backup drive.
AI illustration

Aftonbladet, citing unnamed sources, reported that an AI agent deleted the edited second season of a TV4 show. The premiere planned for 22 November moved to 2027.

Members · 30 days free

See what the evidence actually shows.

Members read the full check on every story: what the evidence shows, why it matters to you and the one line to take with you. Every past edition, re-verified, and the Receipts Pack come with it. A$89 a year, about A$0.24 a day.

  1. At sign-up: 30 days free. Payment card required. One introductory trial per customer.
  2. Your trial ends 30 days after you start. Unless you cancel before then in Account → Manage subscription, we charge A$89 for the first year. It renews automatically at A$89/yr until cancelled.
  3. You can ask for a full refund within 14 days after any annual payment, renewals included, with no reason needed, by emailing [email protected]. This voluntary refund does not limit your rights under the Australian Consumer Law.

By starting your trial, you agree to the Terms.

BS Killer is published by Inferno Tech Pty Ltd, ABN 27 647 413 474.

Open the evidence4 source pages

The claim we checked

An unsupervised AI agent deleted a TV show's entire edited season weeks before release.

These are quoted receipts, not a count of independent investigations. Several reports may rely on the same original source.

gizmodo.com ↗
citing unnamed sources familiar with the matter, that an AI agent deleted the edited version of the show’s upcoming second season
euronews.com ↗
pushing the planned 22 November premiere back to 2027
euronews.com ↗
Mia Lindqvist denied that an AI agent was responsible
euronews.com ↗
TV4 said all the original footage of the show remains available but the season must be edited again from scratch.
gizmodo.com ↗
TV4 later told Aftonbladet that its investigation found the material disappeared because of “the human factor” at the production company and “not AI.”
futurism.com ↗
falls short of ruling out AI involvement
futurism.com ↗
their backup didn’t work and that means they have to recut everything from scratch
nyadagbladet.se ↗
Hon säger i stället att problemet var en backup som inte höll.
nyadagbladet.se ↗
Pressansvarige Ludvig Ternstedt beskriver det som ett tekniskt haveri.
Open this check in the full collection →
The idea, illustrated01

Go inside the check.

    The full explanation appears when your reading access is confirmed.
    Next: Two Macs you cannot buy

    Meta's 816B Muse Spark Open does run on two Mac Studios: two 512GB Macs that MacRumors says Apple has not released, in 4-bit precision.

    A squeezed brain balanced across two desktop computers, and a shopper at a store window facing an empty display stand.
    AI illustration

    Meta released Muse Spark Open, and Apple published a post on running it across Mac Studios with MLX.

    Members · 30 days free

    See what the evidence actually shows.

    Members read the full check on every story: what the evidence shows, why it matters to you and the one line to take with you. Every past edition, re-verified, and the Receipts Pack come with it. A$89 a year, about A$0.24 a day.

    1. At sign-up: 30 days free. Payment card required. One introductory trial per customer.
    2. Your trial ends 30 days after you start. Unless you cancel before then in Account → Manage subscription, we charge A$89 for the first year. It renews automatically at A$89/yr until cancelled.
    3. You can ask for a full refund within 14 days after any annual payment, renewals included, with no reason needed, by emailing [email protected]. This voluntary refund does not limit your rights under the Australian Consumer Law.

    By starting your trial, you agree to the Terms.

    BS Killer is published by Inferno Tech Pty Ltd, ABN 27 647 413 474.

    Open the evidence4 source pages

    The claim we checked

    Meta's 816-billion-parameter Muse Spark Open model runs locally on a pair of Mac Studios.

    These are quoted receipts, not a count of independent investigations. Several reports may rely on the same original source.

    macrumors.com ↗
    most capable U.S. open-weight model to date
    macrumors.com ↗
    two 512GB Apple M5 Ultra Mac Studios running the latest MLX framework
    macrumors.com ↗
    as of yet, but it's supposed to be coming later this month
    macrumors.com ↗
    It won't be cheap, because the 256GB
    machinelearning.apple.com ↗
    On a cluster of four Mac Studios, Muse Spark Open generates over 36 tokens per second
    machinelearning.apple.com ↗
    Muse Spark Open in 4-bit (NVFP4) precision
    letsdatascience.com ↗
    15.74 generated tokens per second for a two-Mac pipeline setup
    letsdatascience.com ↗
    Meta's claim that this is the most capable U.S. open-weight model is its own assessment
    huggingface.co ↗
    requires a GPU with at least 700GB of VRAM for NVFP4 and 2TB for BF16
    Open this check in the full collection →
    The idea, illustrated02

    Go inside the check.

      The full explanation appears when your reading access is confirmed.
      Next: A map of the tests

      Mercor says 90% of US work hours sit on tasks AI benchmarks barely test. It is a fact about 29 tests, not about what AI can do, and Mercor calls it provisional.

      A flashlight beam lighting a single office tower in a dark city of workplaces, an inspector pointing at the dark.
      AI illustration

      Mercor published a benchmark map of the US economy, and its 90% figure reached the Hacker News front page.

      Members · 30 days free

      See what the evidence actually shows.

      Members read the full check on every story: what the evidence shows, why it matters to you and the one line to take with you. Every past edition, re-verified, and the Receipts Pack come with it. A$89 a year, about A$0.24 a day.

      1. At sign-up: 30 days free. Payment card required. One introductory trial per customer.
      2. Your trial ends 30 days after you start. Unless you cancel before then in Account → Manage subscription, we charge A$89 for the first year. It renews automatically at A$89/yr until cancelled.
      3. You can ask for a full refund within 14 days after any annual payment, renewals included, with no reason needed, by emailing [email protected]. This voluntary refund does not limit your rights under the Australian Consumer Law.

      By starting your trial, you agree to the Terms.

      BS Killer is published by Inferno Tech Pty Ltd, ABN 27 647 413 474.

      Open the evidence3 source pages

      The claim we checked

      90% of US working hours are spent on tasks not covered well by AI benchmarks.

      These are quoted receipts, not a count of independent investigations. Several reports may rely on the same original source.

      mercor.com ↗
      90% of American working hours are spent on tasks with little or no coverage from existing benchmarks
      mercor.com ↗
      we start with a list of 29 benchmarks selected for their quality and economic relevance
      mercor.com ↗
      we then ask an LLM judge to tag each benchmark prompt with as many O*NET tasks as it thinks correspond to that prompt
      mercor.com ↗
      only 208 of 923 occupations have a task with at least 5 benchmark prompts
      mercor.com ↗
      Benchmarks are not yet saturated by frontier models.
      mercor.com ↗
      Any incomplete list will necessarily bias down estimates of economic coverage
      mercor.com ↗
      We validate our LLM tagging with human expert tags on Mercor’s APEX benchmarks, a subset of this list of 29.
      runtimewire.com ↗
      Mercor says the coverage estimate is provisional and depends on its selection rules.
      the-decoder.com ↗
      a field dominated by programming that makes up just 7.6 percent of total US employment
      Open this check in the full collection →
      The idea, illustrated03

      Go inside the check.

        The full explanation appears when your reading access is confirmed.
        The finish line

        Edition complete

        You’re up to speed.

        That’s the 17 Oct 2026 briefing. Keep the useful bits. Leave the noise.

        Reading estimate: 546 words at 200 words per minute. Source quotes and the optional sections below add reading time.

        Get tomorrow's check in your inbox.

        One AI claim a night, checked against independent sources, receipts attached. Free.

        Free email updates. Unsubscribe any time.