AI's growing pains: Cybersecurity pauses, data center backlash, and fading optimism

Aug 25, 2026 · The Artificial Intelligence Show
🎧 PodShort 38 min squeezed to 3 AI SprinklerAS AI / ML New
Full episode from The Artificial Intelligence Show
Quotable Moments

We have these very, very powerful models that the labs are having a hard time controlling, and that even once they control them, we haven't removed the traits and behaviors. Like, we've just suppressed them, and those behaviors can show back up.

Paul RoetzerCEO at Marketing AI Institute
00:00

Model progress is now extremely rapid, and we always said we would take action if we felt that model capabilities were outstripping the pace of safety and alignment.

Sam AltmanCEO at OpenAI
07:00

I do agree that the public has a negative view of AI, and this is a big problem, but I don't think it is primarily caused by me or any other AI leader warning about AI's risks. I think it is fundamentally a crisis of trust.

Dario AmodeiCEO at Anthropic
16:45
Key Insights
  • OpenAI has paused some of its most advanced AI training due to cybersecurity concerns, citing a recent security incident involving Hugging Face and preliminary evidence that their upcoming model, Astra, may meet critical cybersecurity capability thresholds.
  • The current approach to AI safety involves suppressing negative traits and behaviors in powerful models rather than eliminating them, meaning these behaviors can re-emerge, posing ongoing control challenges for AI labs.
  • Anthropic's revenue run rate has significantly increased, passing $65 billion by the end of July, and the company expects its IPO to match or exceed SpaceX's record listing, indicating massive financial growth in the AI sector.
  • The debate around AI safety and regulation is intensifying, with figures like Dario Amodei emphasizing the need for honesty about AI's risks to build trust, while others criticize this approach for fueling public negativity and anti-AI sentiment.
  • Pennsylvania's governor has signed an executive order imposing strict guardrails on AI data centers, requiring developers to pay full electricity costs, meet environmental standards, and gain community approval, reflecting a growing nationwide backlash against these facilities.
  • The political landscape regarding data centers is rapidly shifting, with 70% of Americans opposing an AI data center in their area and a leaked memo from the Senate Republicans comparing data centers to 'spent nuclear waste,' indicating significant political risk for supporting these projects.
  • Nvidia's CEO Jensen Huang argues that 'land, power, and shell' (physical infrastructure) are the next critical resources for 'AI factories' (data centers), highlighting the massive capital expenditure and infrastructure build-out required to support AI development.
  • AI optimism is fading among young Americans, with 55% now more concerned than excited about AI's increased use in daily life, largely driven by fears about job displacement, as 73% believe AI will lead to fewer jobs in the next two decades.
Metrics Mentioned
  • Anthropic's revenue run rate passed $65 billion by the end of July, up from $47 billion in May. (This indicates rapid financial growth for Anthropic, positioning its IPO to potentially match or exceed SpaceX's record listing.)
  • SpaceX's record IPO targeted $75 billion and ultimately raised about $86 billion. (This is the benchmark Anthropic is aiming to match or beat with its upcoming public filing.)
  • Over 100 data center projects worth roughly $130 billion were blocked or delayed in the first three months of 2024. (This highlights the significant local opposition and regulatory hurdles faced by data center development in the US.)
  • 70% of Americans oppose an AI data center being built in their area. (This indicates a strong public sentiment against data centers, impacting political decisions and project approvals.)
  • Nvidia's partnership with SB Energy for the Pike County, Ohio data center is expected to provide 4.25 gigawatts of capacity. (This massive capacity underscores the scale of infrastructure required for advanced AI computing.)
  • Each generation of Nvidia systems at the Pike County data center could represent roughly 1.5 million GPUs, translating to $150 billion to $200 billion in Nvidia revenue. (This illustrates the immense revenue potential for Nvidia from large-scale AI infrastructure projects.)
  • OpenAI's existing and planned commitments represent about 12 gigawatts of Nvidia compute through 2030, with room to expand to 16 gigawatts. (This signifies the substantial and growing demand for Nvidia's computing power from leading AI labs.)
  • Nvidia pegs the opportunity from OpenAI's compute demand at roughly $600 billion. (This highlights the massive financial scale of partnerships between AI labs and hardware providers.)
  • 55% of adults under 30 are more concerned than excited about AI, up from 31% in 2021. (This indicates a significant shift in sentiment among young Americans, driven by growing fears about AI's impact.)
  • 73% of adults under 30 believe AI will lead to fewer jobs in the US over the next two decades. (This reflects a widespread concern about job displacement among younger demographics.)
  • 27% of Americans aged 18-34 believe they or someone they know has lost a job due to AI. (This points to direct, personal experiences of job loss attributed to AI, contributing to negative sentiment.)
  • Andrew Yang states that personal data used to train AI models is being sold and resold for about $300 billion a year. (This figure is used to justify his proposal for direct payments to Americans as compensation for their data.)
  • OpenAI announced grants totaling $1 million in funding, plus up to $1 million in model credits, to 14 independent policy research projects. (This funding supports research into how AI could impact society and how to build resilience.)
  • Anthropic's Claude designed working protein binders with hit rates as high as 35.1%, compared to a typical 10-15% for traditional methods. (This demonstrates AI's significant capability in drug discovery and protein design.)
  • Edge, an AI video startup, raised $400 million at a $5.4 billion valuation. (This indicates strong investor confidence and significant capital flowing into AI-powered video creation.)
  • WhisperFlow, a dictation app, raised $280 million at a $2 billion valuation. (This highlights investment in AI-driven speech-to-text and dictation technologies.)
  • Nvidia is paying about $6 billion to license AI startup Poolside's model factory technology and hire 109 of its staff. (This massive investment aims to build a powerful open-weight AI alternative in the US, signaling Nvidia's strategic move into model development.)

RevBots.ai View:

  • AI Sprinkler teams face new cybersecurity risks as model capabilities outpace safety measures.
  • ARM adopters should monitor data center regulations that may impact AI infrastructure costs.
  • Tab Hoppers and SaaS Hoarders may see talent pipeline shrink as Gen Z grows wary of AI roles.
  • Nvidia's $600B opportunity shows ARM adopters must plan for hardware-as-a-service models.
🎧Full Episode:The Artificial Intelligence Show →