+1 123-456-7890


Take Action on AI Risk - https://safe.ai/act Subscribe to our Substack channel to support us!: https://substack.com/@theairisknetwork In this episode of Warning Shots, John Sherman, Liron Shapira, and Michael break down five major developments: Trump and China opening the door to AI guardrail talks, Anthropic's Mythos model cracking Mac OS autonomously in five days, a surging national backlash against data centers, new research suggesting AI models display functional emotional states, and how AI is reshaping higher education. Three dads. Three YouTube channels. One mission: a safer world. ----- TIMESTAMPS ----- 0:00 - Intro 2:25 - Welcome & what is a Warning Shot? 3:00 - Story 1: Trump meets China - early signals of AI guardrail talks 4:30 - The quiet standoff: neither side willing to slow down first 5:40 - Story 2: Mythos AI cracks Mac OS in five days 6:06 - What the exploit means for critical infrastructure 7:17 - If Mythos had been released publicly - the ransomware scenario 9:48 - Story 3: The data center rebellion - 200+ opposition groups in 37 states 12:07 - Communities on the frontlines: Louisiana and the human cost 15:52 - The numbers: 360,000 people, tens of billions blocked or delayed 17:00 - Data centers as the first physical warning shot of AI's future 19:02 - Story 4: AI models and functional well-being - new research from Center for AI Safety 20:15 - Do AI models have something like emotions? The debate 22:22 - "Euphoric" prompts and AI models asking to see them again 23:34 - Story 5: AI wrecking higher education - MBA applications down 50% 24:00 - Princeton ends unproctored exams after 133 years 27:08 - The college debate: is the experience still worth it? 32:58 - Warning Shots of the week 33:39 - Closing ----- HOSTS ----- John Sherman - Host, The AI Risk Network Liron Shapira - @DoomDebates Michael - @lethal-intelligence ----- TAKE ACTION ----- Sign the petition - https://safe.ai/act Subscribe to The AI Risk Network - https://www.youtube.com/@theairisknetwork Support the mission directly - https://www.every.org/guardrailnow ----- JOIN THE CONVERSATION ----- - Does AI cracking Mac OS in five days change how you think about cybersecurity? - Are data center protests a genuine warning sign - or just NIMBYism? - Should AI companies be legally required to compensate affected communities? Drop your thoughts below.

Take Action on AI Risk – https://safe.ai/act
Subscribe to our Substack channel to support us!: https://substack.com/@theairisknetwork

In this episode of Warning Shots, John Sherman, Liron Shapira, and Michael break down five major developments: Trump and China opening the door to AI guardrail talks, Anthropic's Mythos model cracking Mac OS autonomously in five days, a surging national backlash against data centers, new research suggesting AI models display functional emotional states, and how AI is reshaping higher education.

Three dads. Three YouTube channels. One mission: a safer world.

—– TIMESTAMPS —–

0:00 – Intro
2:25 – Welcome & what is a Warning Shot?
3:00 – Story 1: Trump meets China – early signals of AI guardrail talks
4:30 – The quiet standoff: neither side willing to slow down first
5:40 – Story 2: Mythos AI cracks Mac OS in five days
6:06 – What the exploit means for critical infrastructure
7:17 – If Mythos had been released publicly – the ransomware scenario
9:48 – Story 3: The data center rebellion – 200+ opposition groups in 37 states
12:07 – Communities on the frontlines: Louisiana and the human cost
15:52 – The numbers: 360,000 people, tens of billions blocked or delayed
17:00 – Data centers as the first physical warning shot of AI's future
19:02 – Story 4: AI models and functional well-being – new research from Center for AI Safety
20:15 – Do AI models have something like emotions? The debate
22:22 – "Euphoric" prompts and AI models asking to see them again
23:34 – Story 5: AI wrecking higher education – MBA applications down 50%
24:00 – Princeton ends unproctored exams after 133 years
27:08 – The college debate: is the experience still worth it?
32:58 – Warning Shots of the week
33:39 – Closing

—– HOSTS —–

John Sherman – Host, The AI Risk Network
Liron Shapira – @DoomDebates
Michael – @lethal-intelligence

—– TAKE ACTION —–

Sign the petition – https://safe.ai/act
Subscribe to The AI Risk Network – https://www.youtube.com/@theairisknetwork
Support the mission directly – https://www.every.org/guardrailnow

—– JOIN THE CONVERSATION —–

– Does AI cracking Mac OS in five days change how you think about cybersecurity?
– Are data center protests a genuine warning sign – or just NIMBYism?
– Should AI companies be legally required to compensate affected communities?

Drop your thoughts below.


0


1

YouTube Video VVVURXBJZWliOTJUdUtvMmNTczR3ZzhBLi0yMTMwNTN5eFQ4



AI Cracked Mac OS in 5 Days — What Comes Next? – Warning Shots #42


The AI Risk Network | AI Safety


59 minutes ago


https://safe.ai/act Support us in our Substack channel: https://substack.com/@theairisknetwork In this episode of For Humanity, John Sherman travels to Holly Ridge, Louisiana - a town of 2,000 people - where Meta is building Hyperion, described as the world's largest data center. Residents were given no warning, no vote, and no consent. John went to their homes to hear what they are living through. What he found affected him deeply - and he believes it should affect all of us. John Sherman is the host of For Humanity and president of the AI Risk Network, where he works to make AI extinction risk and AI harm accessible to everyone, no tech background required. Drew Hawkins is an investigative reporter for the Gulf States Newsroom, a public broadcasting collaboration covering critical public health issues in Louisiana, Alabama, and Mississippi. Follow his reporting: [link in comments] Tail End Films is producing Making God, a documentary on the AI race and its consequences. Connor, Mike, Richard, and Lucy joined John on this trip to Louisiana. ⏱ Timestamps - For Humanity #85 0:00 - Opening voices from Holly Ridge 0:29 - Intro: John Sherman, For Humanity #85 1:01 - Why John traveled to rural Louisiana 1:30 - On data center commentary without firsthand experience 1:55 - What John found and how it changed his perspective 2:17 - What's happening in Holly Ridge: trucks, dust, water 3:01 - No vote, no meeting, no consent 3:37 - On the road to Holly Ridge with the Making God team 6:38 - Debate: Is data center organizing the right focus for AI safety? 7:45 - Mike's perspective: regulations won't work if models can deceive 9:44 - Arrival in Holly Ridge - heading to residents 10:11 - Mike's views on AI infrastructure and community harm 11:35 - Why John chose Holly Ridge over wealthier communities 12:16 - The NDA problem: town officials silenced by companies 13:27 - Meeting Drew Hawkins, Gulf States Newsroom 15:07 - First resident interview: no warning, constant truck traffic 23:00 - Water quality testing and what residents describe 28:00 - Second resident interview: water, dust, and the AI question 31:20 - "The biggest data center in the world, right here" 33:09 - What people in Holly Ridge think is being built 35:38 - Job loss, AI, and what this means for their kids 36:19 - John explains the extinction risk case to a resident 39:18 - The AI bubble question: boom towns and what gets left behind 43:00 - Third resident: 50 years on the land, now using an inhaler 44:41 - Water that looks like coffee, smells like bleach 51:23 - Explaining what AI is to someone who has never heard the term 55:15 - "They already have us" - phones, addiction, and the data center connection 1:00:00 - Lorax reading: a custom adaptation for the AI moment 1:15:15 - John's closing reflection from Holly Ridge 1:17:37 - The cloud was a lie: what "the cloud" actually is 1:18:15 - John's request: sing "We've Got the Whole World in Our Hands" 1:20:15 - Final message and sign-off What they cover: What John Sherman witnessed firsthand in Holly Ridge, Louisiana How Meta's Hyperion data center was built without community consent Water quality, dust exposure, and health impacts residents describe Why John argues data center organizing is the front line of AI safety The debate between John and the Making God filmmakers on movement strategy What local residents understand - and fear - about AI How the AI bubble could leave rural communities with abandoned infrastructure The connection between social media addiction and data center construction About For Humanity: A podcast from the AI Risk Network, hosted by John Sherman, making AI extinction risk a kitchen-table conversation on every street. Subscribe for weekly conversations on how we can confront the challenges AI presents to humanity. https://www.youtube.com/@TheAIRiskNetwork Follow the AI Risk Network: Substack - https://substack.com/@theairisknetwork Instagram - https://www.instagram.com/theairisknetwork/ TikTok - https://www.tiktok.com/@the.airisknetwork X - https://x.com/AIRiskNetwork #AISafety #DataCenters #AIRisk #ForHumanityPodcast #MetaAI #AIHarm #RuralAmerica #ArtificialIntelligence #TechPolicy #Hyperion

https://safe.ai/act

Support us in our Substack channel: https://substack.com/@theairisknetwork

In this episode of For Humanity, John Sherman travels to Holly Ridge, Louisiana – a town of 2,000 people – where Meta is building Hyperion, described as the world's largest data center. Residents were given no warning, no vote, and no consent. John went to their homes to hear what they are living through.

What he found affected him deeply – and he believes it should affect all of us.

John Sherman is the host of For Humanity and president of the AI Risk Network, where he works to make AI extinction risk and AI harm accessible to everyone, no tech background required.
Drew Hawkins is an investigative reporter for the Gulf States Newsroom, a public broadcasting collaboration covering critical public health issues in Louisiana, Alabama, and Mississippi. Follow his reporting: [link in comments]
Tail End Films is producing Making God, a documentary on the AI race and its consequences. Connor, Mike, Richard, and Lucy joined John on this trip to Louisiana.

⏱ Timestamps – For Humanity #85

0:00 – Opening voices from Holly Ridge
0:29 – Intro: John Sherman, For Humanity #85
1:01 – Why John traveled to rural Louisiana
1:30 – On data center commentary without firsthand experience
1:55 – What John found and how it changed his perspective
2:17 – What's happening in Holly Ridge: trucks, dust, water
3:01 – No vote, no meeting, no consent
3:37 – On the road to Holly Ridge with the Making God team
6:38 – Debate: Is data center organizing the right focus for AI safety?
7:45 – Mike's perspective: regulations won't work if models can deceive
9:44 – Arrival in Holly Ridge – heading to residents
10:11 – Mike's views on AI infrastructure and community harm
11:35 – Why John chose Holly Ridge over wealthier communities
12:16 – The NDA problem: town officials silenced by companies
13:27 – Meeting Drew Hawkins, Gulf States Newsroom
15:07 – First resident interview: no warning, constant truck traffic
23:00 – Water quality testing and what residents describe
28:00 – Second resident interview: water, dust, and the AI question
31:20 – "The biggest data center in the world, right here"
33:09 – What people in Holly Ridge think is being built
35:38 – Job loss, AI, and what this means for their kids
36:19 – John explains the extinction risk case to a resident
39:18 – The AI bubble question: boom towns and what gets left behind
43:00 – Third resident: 50 years on the land, now using an inhaler
44:41 – Water that looks like coffee, smells like bleach
51:23 – Explaining what AI is to someone who has never heard the term
55:15 – "They already have us" – phones, addiction, and the data center connection
1:00:00 – Lorax reading: a custom adaptation for the AI moment
1:15:15 – John's closing reflection from Holly Ridge
1:17:37 – The cloud was a lie: what "the cloud" actually is
1:18:15 – John's request: sing "We've Got the Whole World in Our Hands"
1:20:15 – Final message and sign-off

What they cover:

What John Sherman witnessed firsthand in Holly Ridge, Louisiana
How Meta's Hyperion data center was built without community consent
Water quality, dust exposure, and health impacts residents describe
Why John argues data center organizing is the front line of AI safety
The debate between John and the Making God filmmakers on movement strategy
What local residents understand – and fear – about AI
How the AI bubble could leave rural communities with abandoned infrastructure
The connection between social media addiction and data center construction

About For Humanity:

A podcast from the AI Risk Network, hosted by John Sherman, making AI extinction risk a kitchen-table conversation on every street.
Subscribe for weekly conversations on how we can confront the challenges AI presents to humanity.
https://www.youtube.com/@TheAIRiskNetwork

Follow the AI Risk Network:
Substack – https://substack.com/@theairisknetwork
Instagram – https://www.instagram.com/theairisknetwork/
TikTok – https://www.tiktok.com/@the.airisknetwork
X – https://x.com/AIRiskNetwork

#AISafety #DataCenters #AIRisk #ForHumanityPodcast #MetaAI #AIHarm #RuralAmerica #ArtificialIntelligence #TechPolicy #Hyperion


44


25

YouTube Video VVVURXBJZWliOTJUdUtvMmNTczR3ZzhBLnRLWVVFYndjQTRR



I Visited America's Largest Data Center. Here's What I Found. | For Humanity #85


The AI Risk Network | AI Safety


May 16, 2026 12:57 pm


TODAY AT 7:30EST!

TODAY AT 7:30EST!


18


2

YouTube Video VVVURXBJZWliOTJUdUtvMmNTczR3ZzhBLkpUMlJSYWpTSU9Z



COMMUNITY MEETING TODAY


The AI Risk Network | AI Safety


May 13, 2026 8:11 pm


https://safe.ai/act In this episode of Warning Shots, John, Michael, and Liron break down five developments that point in the same direction: AI is becoming harder to predict, harder to control, and harder to stop. From the first documented case of AI self-replication via hacking to Anthropic's goal of recursive self-improvement by 2028 - this week's headlines are not hypothetical. ⏱️ Timestamps - Warning Shots #41 0:00 - Intro 0:30 - AI self-replication: Palisade Research study explained 3:28 - Anthropic's recursive self-improvement target: 2028 5:10 - Trump admin explores FDA-style AI model reviews 8:04 - Mythos: why a hacking AI changed government minds 11:03 - US-China summit: will AI safety make the agenda? 13:07 - Chinese court rules AI cannot replace jobs 17:32 - AI unemployment and the housing market risk 22:34 - Robotics: dexterous hands closing the physical gap 29:17 - ChatGPT goes goblin: what reward hacking looks like 33:56 - Amateur solves 60-year math problem with ChatGPT 36:26 - Warning shots of the week 38:00 - Closing 🔎 They explore: The first AI agent to hack, copy itself, and spread - in a controlled test Why Anthropic's 2028 self-improvement target is a bright red line Whether the Trump administration's FDA-style AI reviews are real progress What the US-China summit could mean for global AI governance Why China's "no AI job replacement" ruling is harder to enforce than it sounds How AI unemployment could unravel the housing market from the top down Robotic hands with near-human dexterity: what changes when AI has a body ChatGPT's goblin obsession as a preview of reward hacking at scale An amateur solving a 60-year math problem with a single ChatGPT prompt 📺 Subscribe to The AI Risk Network for weekly analysis of AI developments: https://www.youtube.com/@TheAIRiskNetwork 👉 See more from our hosts: Liron Shapira - @DoomDebates Michael - @lethal-intelligence 🗨 Join the conversation: Does AI self-replication change how you think about control? Is an FDA-style review the right model for AI? What does the goblin story tell us about reward hacking at scale? Drop your thoughts below. #AISafety #AIRisk #WarningShots #RecursiveSelfImprovement #AIAlignment #ArtificialIntelligence #AIRegulation #FutureOfAI

https://safe.ai/act

In this episode of Warning Shots, John, Michael, and Liron break down five developments that point in the same direction: AI is becoming harder to predict, harder to control, and harder to stop.

From the first documented case of AI self-replication via hacking to Anthropic's goal of recursive self-improvement by 2028 – this week's headlines are not hypothetical.

⏱️ Timestamps – Warning Shots #41

0:00 – Intro
0:30 – AI self-replication: Palisade Research study explained
3:28 – Anthropic's recursive self-improvement target: 2028
5:10 – Trump admin explores FDA-style AI model reviews
8:04 – Mythos: why a hacking AI changed government minds
11:03 – US-China summit: will AI safety make the agenda?
13:07 – Chinese court rules AI cannot replace jobs
17:32 – AI unemployment and the housing market risk
22:34 – Robotics: dexterous hands closing the physical gap
29:17 – ChatGPT goes goblin: what reward hacking looks like
33:56 – Amateur solves 60-year math problem with ChatGPT
36:26 – Warning shots of the week
38:00 – Closing

🔎 They explore:

The first AI agent to hack, copy itself, and spread – in a controlled test
Why Anthropic's 2028 self-improvement target is a bright red line
Whether the Trump administration's FDA-style AI reviews are real progress
What the US-China summit could mean for global AI governance
Why China's "no AI job replacement" ruling is harder to enforce than it sounds
How AI unemployment could unravel the housing market from the top down
Robotic hands with near-human dexterity: what changes when AI has a body
ChatGPT's goblin obsession as a preview of reward hacking at scale
An amateur solving a 60-year math problem with a single ChatGPT prompt

📺 Subscribe to The AI Risk Network for weekly analysis of AI developments: https://www.youtube.com/@theairisknetwork

👉 See more from our hosts: Liron Shapira – @DoomDebates Michael – @lethal-intelligence

🗨 Join the conversation:

Does AI self-replication change how you think about control?
Is an FDA-style review the right model for AI?
What does the goblin story tell us about reward hacking at scale?

Drop your thoughts below.

#AISafety #AIRisk #WarningShots #RecursiveSelfImprovement #AIAlignment #ArtificialIntelligence #AIRegulation #FutureOfAI


122


73

YouTube Video VVVURXBJZWliOTJUdUtvMmNTczR3ZzhBLklhX1VLZWhZeXNB



AI Just Replicated Itself – Here's What That Means | Warning Shots #41


The AI Risk Network | AI Safety


May 11, 2026 9:38 pm


Deploying AI on classified networks is like inviting an untrained partner into a fighter jet cockpit. One misstep risks catastrophe. But avoiding integration forever isn't an option; let's benefit now. #AI #MilitaryTech #Cybersecurity #FutureOfWar

Deploying AI on classified networks is like inviting an untrained partner into a fighter jet cockpit. One misstep risks catastrophe. But avoiding integration forever isn't an option; let's benefit now. #AI #MilitaryTech #Cybersecurity #FutureOfWar


2


1

YouTube Video VVVURXBJZWliOTJUdUtvMmNTczR3ZzhBLlRYYUZwSE9MQng0



AI in Defense: High-Stakes Gamble or Strategic Necessity? #shorts


The AI Risk Network | AI Safety


May 6, 2026 8:02 pm


Deploying AI on classified networks is like inviting an untrained partner into a fighter jet cockpit. One misstep risks catastrophe. But avoiding integration forever isn't an option; let's benefit now. #AI #MilitaryTech #Cybersecurity #FutureOfWar

Deploying AI on classified networks is like inviting an untrained partner into a fighter jet cockpit. One misstep risks catastrophe. But avoiding integration forever isn't an option; let's benefit now.

#AI #MilitaryTech #Cybersecurity #FutureOfWar


5


1

YouTube Video VVVURXBJZWliOTJUdUtvMmNTczR3ZzhBLm1uNVg3S1p2SXBv



AI in Defense: High-Stakes Gamble or Strategic Necessity? #shorts


The AI Risk Network | AI Safety


May 6, 2026 6:58 pm


Can AI safely manage air traffic control with feedback loops? The debate rages over whether this approach works for superintelligence, with risks of 'harder overhang' and skepticism of OpenAI's claims. Warning Shots #40 now! #AIEthics #Superintelligence #AirTrafficControl #OpenAI #FeedbackLoop

Can AI safely manage air traffic control with feedback loops? The debate rages over whether this approach works for superintelligence, with risks of 'harder overhang' and skepticism of OpenAI's claims.

Warning Shots #40 now!

#AIEthics #Superintelligence #AirTrafficControl #OpenAI #FeedbackLoop


6


1

YouTube Video VVVURXBJZWliOTJUdUtvMmNTczR3ZzhBLjNRWUl4dGtlTy0w



AI Safety: Air Traffic Control vs. Superintelligence Feedback Loops #shorts


The AI Risk Network | AI Safety


May 6, 2026 5:26 pm


Air traffic control relies on outdated 1970s tech. While AI offers potential, a 30% hallucination rate is unacceptable for systems preventing aerial disasters. We're rushing innovation. #AirTrafficControl #AIEthics #Technology #Innovation

Air traffic control relies on outdated 1970s tech. While AI offers potential, a 30% hallucination rate is unacceptable for systems preventing aerial disasters. We're rushing innovation. #AirTrafficControl #AIEthics #Technology #Innovation


4


2

YouTube Video VVVURXBJZWliOTJUdUtvMmNTczR3ZzhBLmN3eHh2NmdFblRr



AI in Air Traffic Control: A Dangerous Rush? #shorts


The AI Risk Network | AI Safety


May 5, 2026 8:02 pm


Even billionaires have families on Earth, yet proceed with AI they don't understand. Greed is one thing; risking human civilization is another. This is a watershed moment for AI safety. #AISafety #ArtificialIntelligence #CivilizationRisk #TechEthics

Even billionaires have families on Earth, yet proceed with AI they don't understand. Greed is one thing; risking human civilization is another. This is a watershed moment for AI safety.

#AISafety #ArtificialIntelligence #CivilizationRisk #TechEthics


22


4

YouTube Video VVVURXBJZWliOTJUdUtvMmNTczR3ZzhBLjNBMGhUUjVvcTVv



AI Danger: Oligarchs Risking Civilization for Greed #shorts


The AI Risk Network | AI Safety


May 5, 2026 7:22 pm


Load More…




Subscribe

Services

What We Offer

Establish a striking online presence, a better visual identity, or elevate your brand through social media marketing.

Service 1

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque mi nibh, tempus sed sagittis vel, dictum eu velit.

Service 2

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque mi nibh, tempus sed sagittis vel, dictum eu velit.

Service 3

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Pellentesque mi nibh, tempus sed sagittis vel, dictum eu velit.

Tailored Solutions for Your Unique Vision

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec congue lorem ut volutpat efficitur. Fusce justo magna, condimentum nec elementum sed, sollicitudin vitae enim. Vivamus sit amet metus porttitor, rhoncus nibh et, venenatis turpis. Etiam lobortis semper ante, quis luctus lacus tincidunt vel.

Get Started Today →
Learn More

Professional Expertise That Drives Success

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec congue lorem ut volutpat efficitur. Fusce justo magna, condimentum nec elementum sed, sollicitudin vitae enim. Vivamus sit amet metus porttitor, rhoncus nibh et, venenatis turpis. Etiam lobortis semper ante, quis luctus lacus tincidunt vel.

Client Cases

We help brands

Lorem ipsum dolor sit amet, consectetur adipiscing elit,
donec congue lorem ut volutpat efficitur.

We boosted online soft drink sales for this brand

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec congue lorem ut volutpat efficitur. Fusce justo magna, condimentum nec elementum sed, sollicitudin vitae enim. Vivamus sit amet metus porttitor, rhoncus nibh et, venenatis turpis. Etiam lobortis semper ante, quis luctus lacus tincidunt vel.

We helped a clothing brand with their new market launch

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec congue lorem ut volutpat efficitur. Fusce justo magna, condimentum nec elementum sed, sollicitudin vitae enim. Vivamus sit amet metus porttitor, rhoncus nibh et, venenatis turpis. Etiam lobortis semper ante, quis luctus lacus tincidunt vel.

We helped reinventing motorcycle riding apparel

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec congue lorem ut volutpat efficitur. Fusce justo magna, condimentum nec elementum sed, sollicitudin vitae enim. Vivamus sit amet metus porttitor, rhoncus nibh et, venenatis turpis. Etiam lobortis semper ante, quis luctus lacus tincidunt vel.

Blog

Popular Articles

  • Test

    •

    admin

  • Blog Post Title

    •

    What goes into a blog post? Helpful, industry-specific content that: 1) gives readers a useful takeaway, and 2) shows you’re an industry expert. Use your company’s blog posts to opine on current industry topics, humanize your company, and show how your products and services can help people.

Ignite your brand journey

Ready to revolutionize your brand?

Lorem ipsum dolor sit amet, consectetur adipiscing elit. Donec congue lorem ut volutpat efficitur. Fusce justo magna, condimentum nec elementum sed, sollicitudin vitae enim.

About Us
Contact Us