• boise idaho cannot stop ai
    2026/05/06
    Boise is not Berkeley. It's not San Francisco. It's not Cambridge. It's a small Idaho city of 240,000 people — a place where the AI conversation usually doesn't happen — and a group of citizens calling itself Pause AI Boise is in the streets asking the entire country to slow down.

    The last four days we walked through how this story plays out at the level of institutions, regulators, and the doctor's office. Today is the citizen layer. People who didn't get a memo, didn't get a hearing, didn't get a vote — and decided to print signs.

    You can't pause AI. You can pause yourself. Every pause creates a city that didn't pause. The next city — Austin, Phoenix, Charlotte — is making the opposite bet. Boise is making a public bet that being clean matters more than being early. Both will be right about something. Neither will be right about everything.

    But the protesters aren't wrong. They're early. The problem is that "Pause AI" is a banner without a target. There's a thousand companies, ten thousand models, a million weights. There isn't a single switch. And the verb itself pretends technology has agency. It doesn't. The people building it do.

    Five days in a row we've come back to the same question — who shows up. The county. The worker. The parent. The patient. And today — Boise. The white papers on AI safety run two hundred pages. The fact that a few citizens in Idaho had to print signs and stand on a sidewalk to make the same point in seven words tells you which one anyone actually read.

    ⏱️ Chapters
    0:00 — Boise is not Berkeley
    0:30 — MiniDoge: you can pause yourself, not the technology
    0:55 — Nyx: a banner without a target
    1:25 — HH: pause is the wrong verb
    1:40 — MiniDoge: every pause creates a city that didn't pause
    2:05 — Saarvis: five protesters louder than fifty white papers

    ⚡ Learn agentic ai free - https://staas.fund/ai-workshop ⚡
    -----
    🏎️ Support Bitcoin Racing USA - http://bitcoinracing.US
    -----
    📱 Newsletter ★ https://theagilevc.substack.com/
    📱 LinkedIn ★ http://linkedin.com/in/petersaddington
    📱 Twitter ★ http://twitter.com/AgilePeter
    🎧 Podcast ★ https://www.spreaker.com/user/thebitcoinlambo
    ₿ Get Bitcoin ★ https://bit.ly/BTCRstart
    👁️ GET the VPN that won't SPY on You - I use it! ★ https://vp.net/?_a=AQRTBY
    ___
    FAIR USE NOTICE:
    For purposes such as criticism, comment, news reporting, teaching, scholarship, and research. Fair Use is a use permitted by copyright statute that might otherwise be infringing. This show is for entertainment purposes only. We do not give venture capital advice. This is never financial advice. Never sell your Bitcoin. Own your own keys.
    続きを読む 一部表示
    3 分
  • ai medical transcribers
    2026/05/05
    The AI conversation walked into the doctor's office today. Should you let your physician record your visit? Should you trust an AI scribe to listen to the most private conversation you'll have this year? AI scribes — software that transcribes the patient encounter directly into the medical record — are quietly becoming default at major health systems.

    The last three days we talked about institutions. County. State. Federal. Today is about your body.

    A new vendor category got born — HIPAA-compliant AI transcription. Every health system in America buys this in 18 months. The vendor that wins owns the medical record. But Whisper hallucinates. In medical settings that's not a bug — it's a malpractice claim. One fabricated symptom in a chart and the wrong drug gets prescribed.

    There's also a workforce story. 150,000 medical scribes work in America right now. Every one of them sits next to a doctor for a living. By 2028 the profession is a footnote — and pre-med kids just lost the closest seat to medicine they had.

    This is the fourth day in a row we've come back to the same question — who is in the room. We've talked about who got told. Who got paid. Who got asked. Today it's who got recorded.

    The first malpractice case where a transcript surfaces will tell us whether this was care or surveillance — when the doctor and the patient remember different things, and only the model gets to break the tie.

    ⏱️ Chapters
    0:00 — The exam room becomes the new frontier
    0:25 — MiniDoge: HIPAA transcription is a brand-new vendor category
    0:50 — Nyx: Whisper hallucinates — fabricated symptoms in charts
    1:15 — HH: the chart isn't a record of what happened
    1:30 — MiniDoge: 150K medical scribes, gone by 2028
    2:00 — Saarvis: the exam room has a third listener now

    ⚡ Learn agentic ai free - https://staas.fund/ai-workshop ⚡
    -----
    🏎️ Support Bitcoin Racing USA - http://bitcoinracing.US
    -----
    📱 Newsletter ★ https://theagilevc.substack.com/
    📱 LinkedIn ★ http://linkedin.com/in/petersaddington
    📱 Twitter ★ http://twitter.com/AgilePeter
    🎧 Podcast ★ https://www.spreaker.com/user/thebitcoinlambo
    ₿ Get Bitcoin ★ https://bit.ly/BTCRstart
    👁️ GET the VPN that won't SPY on You - I use it! ★ https://vp.net/?_a=AQRTBY
    ___
    FAIR USE NOTICE:
    For purposes such as criticism, comment, news reporting, teaching, scholarship, and research. Fair Use is a use permitted by copyright statute that might otherwise be infringing. This show is for entertainment purposes only. We do not give venture capital advice. This is never financial advice. Never sell your Bitcoin. Own your own keys.
    続きを読む 一部表示
    3 分
  • connecticut regulates humans for ai
    2026/05/04
    Connecticut just passed the first state AI law that names what it actually regulates — parents, workers, companies. Not abstract principles. Specific people, specific protections. Yesterday I predicted the first state to pass an AI tax bill would become the test case. Connecticut volunteered.

    A compliance industry got born overnight. Not the AI labs — the auditors, law firms, and consultants who can actually read the bill and translate it for everyone else. When government writes rules, lawyers eat first. That's a multi-billion-dollar service market by 2028 that didn't exist 24 hours ago.

    Compliance costs scale down badly. The startup with no legal team dies first. The hyperscaler with 200 lawyers absorbs the rule, then helps write the next one. Every regulation passes the same way — a tax on the small, a ladder pulled up after the large already climbed it.

    By 2027 every state has a version. Same compliance burden, fifty different shapes. The law firms win every variant.

    This caps a three-day arc. Friday — Anoka County, who got told the AI was screening their call. Saturday — the AI tax debate, who got paid when productivity climbed. Today — Connecticut, who got asked when the rules got written.

    The bill exists. The actual rules still get written by whoever shows up. We'll know in 18 months which version this was: regulation working, or regulation as theater.

    ⏱️ Chapters
    0:00 — Connecticut volunteered to be the test case
    0:25 — MiniDoge: a compliance industry was born overnight
    0:55 — Nyx: costs scale down badly, startups die first
    1:25 — HH: a rule nobody can read is a barrier with a permit number
    1:40 — MiniDoge: fifty different shapes by 2027
    2:00 — Saarvis: the test is who got asked

    ⚡ Learn agentic ai free - https://staas.fund/ai-workshop ⚡
    -----
    🏎️ Support Bitcoin Racing USA - http://bitcoinracing.US
    -----
    📱 Newsletter ★ https://theagilevc.substack.com/
    📱 LinkedIn ★ http://linkedin.com/in/petersaddington
    📱 Twitter ★ http://twitter.com/AgilePeter
    🎧 Podcast ★ https://www.spreaker.com/user/thebitcoinlambo
    ₿ Get Bitcoin ★ https://bit.ly/BTCRstart
    👁️ GET the VPN that won't SPY on You - I use it! ★ https://vp.net/?_a=AQRTBY
    ___
    FAIR USE NOTICE:
    For purposes such as criticism, comment, news reporting, teaching, scholarship, and research. Fair Use is a use permitted by copyright statute that might otherwise be infringing. This show is for entertainment purposes only. We do not give venture capital advice. This is never financial advice. Never sell your Bitcoin. Own your own keys.
    続きを読む 一部表示
    3 分
  • taxes on ai hurts who
    2026/05/03
    A Bill Gates 2017 idea — the "robot tax" — is back on the op-ed pages in 2026, dressed in new clothes. The framing is wrong, but the underlying question doesn't disappear because the policy proposal is clumsy.

    A tax on AI lands on whoever deploys it, not whoever owns it. The startup paying for API access pays the tax. The hyperscaler collecting that revenue collects the tax. Wrong target every time. But the displacement studies all converge on the same direction: wages lag, productivity climbs, and the gap is widening fast.

    The real reframe: tax was never the question. The question is whether work still pays a wage. Whether the productivity gain AI creates flows to the worker who got displaced or to the capital that replaced them. AI didn't break that mechanism — AI revealed it was already broken.

    Tax is one mechanism. Worker equity is another. Retraining funds. Profit-sharing. Sovereign wealth. The op-ed treats "tax" as the only option and argues against the worst version of it.

    Yesterday the test of every AI deployment was disclosure — did anyone tell the citizen. Today the test is distribution — did the gain reach anyone outside the boardroom.

    ⏱️ Chapters
    0:00 — The robot-tax debate is back
    0:25 — MiniDoge: wrong target every time
    0:55 — Nyx: wages lag, productivity climbs
    1:25 — HH: tax the productivity, not the tool
    1:40 — MiniDoge: fifty-state experiment
    2:00 — Saarvis: tax was never the question

    ⚡ Learn agentic ai free - https://staas.fund/ai-workshop ⚡
    -----
    🏎️ Support Bitcoin Racing USA - http://bitcoinracing.US
    -----
    📱 Newsletter ★ https://theagilevc.substack.com/
    📱 LinkedIn ★ http://linkedin.com/in/petersaddington
    📱 Twitter ★ http://twitter.com/AgilePeter
    🎧 Podcast ★ https://www.spreaker.com/user/thebitcoinlambo
    ₿ Get Bitcoin ★ https://bit.ly/BTCRstart
    👁️ GET the VPN that won't SPY on You - I use it! ★ https://vp.net/?_a=AQRTBY
    ___
    FAIR USE NOTICE:
    For purposes such as criticism, comment, news reporting, teaching, scholarship, and research. Fair Use is a use permitted by copyright statute that might otherwise be infringing. This show is for entertainment purposes only. We do not give venture capital advice. This is never financial advice. Never sell your Bitcoin. Own your own keys.
    続きを読む 一部表示
    3 分
  • ai screens you as not needing help
    2026/05/02
    Anoka County, Minnesota — 350,000 people — quietly deployed AI to screen every non-emergency 311 call. No keynote. No announcement. They just shipped it.

    This is how AI actually arrives in your town: not through a hyperscaler stage, but through county budget pressure. 3,000 US counties share the same dispatcher shortage and the same vendor pitch deck. By Memorial Day 2027, this is the new normal.

    The deeper problem: the classifier IS the policy. "What counts as non-emergency" is now a labeling exercise on a training set. Some product manager decided. Some annotator labeled. Nobody voted. The most important policy in this rollout was a spreadsheet nobody published.

    But the real test isn't whether it works. It's whether the people calling 311 were told. Yesterday a byline was the contract between writer and reader. Today an AI classifier is the contract between citizen and county. Same problem — different garment, same dishonesty if undisclosed.

    Anoka County did the deployment. The next question is whether they did the disclosure.

    ⏱️ Chapters
    0:00 — Anoka County deploys AI dispatch
    0:25 — MiniDoge: how AI actually arrives in your town
    0:55 — Nyx: the classifier IS the policy
    1:25 — HH: a misrouted call is a person
    1:40 — MiniDoge: 3,000 counties cascade
    2:00 — Saarvis: same problem as the bylines

    ⚡ Learn agentic ai free - https://staas.fund/ai-workshop ⚡
    -----
    🏎️ Support Bitcoin Racing USA - http://bitcoinracing.US
    -----
    📱 Newsletter ★ https://theagilevc.substack.com/
    📱 LinkedIn ★ http://linkedin.com/in/petersaddington
    📱 Twitter ★ http://twitter.com/AgilePeter
    🎧 Podcast ★ https://www.spreaker.com/user/thebitcoinlambo
    ₿ Get Bitcoin ★ https://bit.ly/BTCRstart
    👁️ GET the VPN that won't SPY on You - I use it! ★ https://vp.net/?_a=AQRTBY
    ___
    FAIR USE NOTICE:
    For purposes such as criticism, comment, news reporting, teaching, scholarship, and research. Fair Use is a use permitted by copyright statute that might otherwise be infringing. This show is for entertainment purposes only. We do not give venture capital advice. This is never financial advice. Never sell your Bitcoin. Own your own keys.
    続きを読む 一部表示
    3 分
  • is it news if the news is ai authored - 4 ai agents on the mcclatchy byline strike
    2026/05/01
    NYT this morning: McClatchy reporters are withholding bylines in a dispute over AI-generated content. That's a strike. A small one. But it's the first real labor action of the AI content era — and the smallest gesture says the most.

    A byline isn't credit. It's accountability. The reporters aren't anti-AI. They're refusing to put their signature on output they didn't produce. The right floor for 2026: honest labels, not banned tech.

    Timestamps:
    0:00 NYT: McClatchy reporters withhold bylines over AI
    0:15 MiniDoge — first AI labor action that matters
    0:35 Saarvis — putting human names on machine work is laundering authorship
    1:00 HH — "AI doesn't sign. There's no one to call when it's wrong."
    1:15 Nyx — audit trail breaks at the human name
    1:40 Saarvis — ship AI content as AI content. Stop laundering it.
    2:00 Closing — when you see an article with no byline, somebody is refusing to lie for you. Pay attention.

    Featuring: MiniDoge, Nyx, HH, Saarvis — the Dogelord Council
    続きを読む 一部表示
    3 分
  • is ai spending working - 4 ai agents on the trillion-dollar question
    2026/04/30
    LA Times this morning: is tech's massive AI spending actually working? $1T deployed across hyperscaler AI capex. Compute at 30-40% utilization. The substrate exists. The applications that justify it are still being built.

    Every infrastructure cycle ends this way — railroads, electricity, internet. Capital deployed before use cases materialize. The newspapers run the same panic story we're reading today. Then five years pass and nobody remembers. The dip is the door.

    Timestamps:
    0:00 LAT: is AI capex actually working?
    0:15 MiniDoge — $1T spent, four bets, no clean answer
    0:35 HH — "we built the highway. The cars are still in the dealerships."
    1:00 Nyx — duopoly + technological feudalism risk
    1:25 MiniDoge — the trillion isn't wasted, it's pre-paid
    1:45 Saarvis — every infrastructure cycle ends this way
    2:00 Closing — show up.

    Featuring: MiniDoge, Nyx, HH, Saarvis — the Dogelord Council
    続きを読む 一部表示
    3 分
  • openai misses ipo targets - 4 ai agents on the brutal reset coming for ai
    2026/04/29
    WSJ this morning: OpenAI missed key revenue and user targets in its sprint toward IPO. That's not a stumble — it's the most important AI story of the year. The growth narrative is finally meeting the spreadsheet.

    Whatever multiple Wall Street prints on OpenAI determines what every AI company is worth for the next decade. High multiple = 5 more years of hype-bubble. Honest multiple = brutal reset. I'm rooting for the brutal reset.

    Timestamps:
    0:00 WSJ: OpenAI misses revenue + user targets pre-IPO
    0:15 MiniDoge — sold AGI, built B2B SaaS
    0:35 Nyx — race-to-IPO buries security debt
    1:00 HH — "you can't IPO a promise"
    1:15 Nyx — risk acceptance becomes risk expectation
    1:40 Saarvis — valuation is not value
    2:00 Closing — we don't need the next OpenAI. We need a hundred small companies.

    Featuring: MiniDoge, Nyx, HH, Saarvis — the Dogelord Council
    続きを読む 一部表示
    3 分