The Race Toward Super-intelligence: Why the Future of AI Affects Us All

By Alternative Press Team
Based on the conversation with Salim Ismail, Dave Blundin, and Peter Diamandis

Recorded June 26, 2025


We are living in one of the most transformative moments in human history. Artificial intelligence (AI) is no longer a tool we use: it’s fast becoming a powerful force shaping economies, governments, education, and even the nature of reality itself.

In a recent discussion between Salim Ismail (founder of OpenExO), Dave Blundin (founder of Link Ventures), and futurist Peter Diamandis, the trio unpacked urgent and groundbreaking developments in the AI space.

.

Reference: The Race for Super Intelligence & What It Means for Us

Chapters

  • 00:00 – The Race for Digital Superintelligence
  • 10:57 – The $32 Billion AI Startup: Insights and Implications
  • 13:41 – Talent Wars in AI: Recruitment and Retention Strategies
  • 16:21 – Government’s Role in AI Development
  • 24:46 – Future Predictions: AI and Natural Disasters
  • 01:00:57 – The Impact of AI on Critical Thinking
  • 01:07:47 – AI and Job Displacement: A Double-Edged Sword
  • 01:14:00 – RoboTaxis: The Future of Transportation
  • 01:32:29 – China’s Solar Power Surge: A Global Perspective
  • 01:45:42 – The Crypto Landscape: Bitcoin’s Resurgence

This conversation is part of a growing call for awareness: the rise of artificial superintelligence (ASI) is not decades away.  It’s imminent.


What Is Superintelligence and Why Should You Care?

Elon Musk predicts we may reach digital superintelligence as early as this year or next. But what does that mean?

The panelists distinguish between:

  • AGI (Artificial General Intelligence): AI that can perform any intellectual task a human can do.

  • ASI (Artificial Superintelligence): AI that far surpasses human intelligence across all domains.

The implications are staggering. If one ASI is developed before all others, it could become the only one: suppressing competitors and potentially concentrating power at an unimaginable scale.

As Dave Blundin puts it, “This is the winner-take-all dynamic.  Once you create the first true ASI, the race is over.”


$32 Billion Startups and the Talent Arms Race

One of the most talked-about stories in the AI world is Ilya Sutskever’s $32 billion startup.  With virtually no product launched, the company has attracted billions in investment from top-tier VCs.  Why?

Because Ilya likely pitched a vision for safe, self-improving ASI, and that’s a game ending move if true.

Meanwhile, companies like Meta and OpenAI are offering $100 million signing bonuses to lure top AI researchers.  It’s an unprecedented scramble for talent reminiscent of the arms race, but with algorithms instead of missiles.

Meta, with $70 billion in cash, is spending aggressively after setbacks like the underwhelming release of LLaMA 4.


Should Governments Control Superintelligence?

Governments are watching closely and preparing to act.  A major theme from the conversation is nationalization: if a company develops ASI, will the government seize control?

Salim Ismail believes it’s likely. “It may not be a buyout. It might be: ‘We own this now. It’s a national security issue.’”

However, Dave Blundin argues that in the U.S., the government is more likely to partner with, rather than absorb, private companies… similar to how defense contracts work today.

Still, the implications are enormous. Once a government holds a monopoly on ASI, that AI could encode a specific worldview: political, religious, or ideological, with global consequences.  Think about what that could mean.


AI in Government: Revolution or Red Tape?

The U.S. government is finally getting serious about AI. The upcoming launch of AI.gov is designed to integrate AI across federal agencies from optimizing air traffic to speeding up FDA drug approvals and infrastructure monitoring.

The potential is massive:

  • Cut costs

  • Eliminate fraud

  • Reduce multi-year permit processes to seconds

But the government won’t build this itself.  Instead, private firms like Palantir and AWS will deliver compartmentalized AI systems, ensuring security and compliance.

Salim calls it a chance to cut the cost of government by 10x, citing examples like how a single map based program cut wind turbine approval time from 3 years to 30 seconds.


AI and Critical Thinking: A Lost Skill?

A troubling insight came from an MIT study: using ChatGPT to write essays drastically reduced participants’ ability to recall or understand what they wrote.  Content recall and understanding dropped by 83% compared to those who researched and wrote manually.

The reason?  When AI fills in the gaps, your brain doesn’t learn.  It’s the “Waze Effect”… just as many folks can no longer navigate without GPS, many people – especially the young AI users –  may soon struggle to think deeply without AI.

But Dave Blundin offers a nuanced take: “Yes, we remember less, but we’re covering far more ground.  Speed and breadth may matter more than memorization.”

Still, the big question remains:  How do we train young minds to think critically in an AI-assisted world?


The ASI That Saves Lives… or Destroys Them?

Not all superintelligence is threatening.  DeepMind recently released an AI model that predicts hurricane paths with stunning accuracy: up to 140 km closer than traditional forecasts.  The model was trained on 45 years of cyclone data and could save billions in damages and countless lives.

But the same capability that allows us to predict hurricanes might one day allow us to steer them.  Imagine a future where governments negotiate to divert disasters. That is an unsettling yet plausible scenario the panel explores.


Education, Toys, and AI Best Friends

OpenAI and Mattel are collaborating to embed GPT-5 into children’s toys like Barbie and Hot Wheels.  The goal: create interactive, educational experiences tailored to a child’s learning style.

But concerns abound.  Will kids form emotional attachments to AI?  Will imagination suffer if toys now talk back?

“We could get incredible data on learning behavior,” says Diamandis, “but we must tread carefully.”


What’s Next: From MIT Dorm Rooms to Multibillion-Dollar Startups

Start-ups like Cursor reached $500 million in ARR (annual recurring revenue) in under three years: the fastest in SaaS history.

Many start-ups are born in places like MIT’s Link Studio, founded by Blundin, where student teams receive seed funding and mentorship.

The formula for success?  Teams of technical best friends, not just smart individuals.

In Israel, this is second nature.  Time spent in their mandatory military service creates tight-knit teams who then build companies.  At MIT and Waterloo, late-night problem sets forge similar lifelong alliances.  As Blundin says, “You can change the idea. You can’t change the relationship.”

The Energy Crisis of Intelligence: AI’s Thirst for Power

Beneath the excitement and exponential growth of AI lies a looming bottleneck: energy.

As Dave Blundin bluntly put it:

“We are going to become limited by power in our quest for digital superintelligence.  This is America’s Achilles heel.”

AI models, especially large foundation models like GPT and Grok, require massive computational power.  Training them is not just a matter of data and talent, it’s a brute force energy problem.  And it’s getting worse with each model iteration.

China’s Head Start

While the U.S. debates regulation and infrastructure, China is going all-in on energy production.  Massive investment into solar, nuclear, and battery storage is giving them an edge in powering the AI race.

As the panel discussed, this isn’t just an economic competition, it’s strategic survival.  The nation that controls the energy to power AI could control its evolution.

A Hidden Arms Race

While Meta throws billions at hiring and start-ups chase valuations, there’s a deeper arms race happening underground in power grids and data centers.

  • AI inference and training are already stretching U.S. power infrastructure.

  • Start-ups are competing for GPU clusters, but may soon be competing for megawatts.

  • Major players like Microsoft and Amazon are investing in private energy sources to power their AI ambitions.

Peter Diamandis points out, “We don’t have a compute bottleneck.  We have a power bottleneck.”

If unchecked, the AI revolution could stall not due to ethics, talent, or regulation, but simply because we can’t keep the lights on.

📈 Bitcoin & Crypto: A Quiet Resurgence Amid the AI Storm

While AI dominates headlines, another disruptive force is quietly regaining momentum: cryptocurrency.   In the closing section of their wide-ranging discussion, Peter Diamandis, Salim Ismail, and Dave Blundin turned their attention to the shifting landscape of Bitcoin, exchanges, and digital finance and the signals are loud and clear: crypto is back in play.

Bitcoin: Not Dead Yet

After a multi-year bear market and regulatory headwinds, Bitcoin is showing signs of renewed strength.  The panel’s tone suggests growing institutional interest and market activity, especially as global distrust of fiat currencies and centralized control deepens.

Salim emphasized that crypto’s underlying philosophy of decentralization, transparency, and programmability, aligns well with a future where trust in centralized institutions is eroding.


The Real Revolution: Crypto Infrastructure Maturity

More significant than price action is what’s happening behind the scenes:

  • Exchanges have become faster, more secure, and more scalable.

  • Decentralized finance (DeFi) continues to push innovation, enabling borrowing, lending, and trading without traditional banks.

  • Layer 2 scaling solutions (like Lightning for Bitcoin or Arbitrum for Ethereum) are making transactions faster and cheaper: both key to widespread adoption.

Dave Blundin pointed out that we’re entering a phase of invisible crypto: where blockchain infrastructure powers apps and financial tools without requiring users to understand the tech.  Just like cloud computing or TCP/IP, blockchain may disappear into the background, yet reshape everything.


Regulation Is Evolving (Slowly)

Peter noted a trend toward “soft regulation”.  Governments are no longer ignoring crypto but aren’t ready to embrace it fully either.  The U.S. appears to be testing frameworks, particularly as digital currencies intersect with national security, banking, and cross-border remittances.

With the rise of central bank digital currencies (CBDCs), the competition between sovereign-issued and open source crypto is heating up.  This friction could define the next decade of global finance.


The Bottom Line

AI may be absorbing all the oxygen in tech media but Bitcoin and blockchain aren’t going anywhere.  If anything, they’re being redefined and reborn in quieter, more strategic ways.

As Diamandis remarked:
“The infrastructure is here. The apps are coming. Crypto’s next wave will feel like it was overnight… but it’s been building all along.”


Final Thoughts: The Curve Is Going Vertical

Every panelist agrees;  this wave of AI is unlike anything the world has seen.

“This is the greatest drama in human history,” says Dave Blundin. “We’re witnessing the birth of a force that may change everything.”

Whether AI becomes a liberator or a new form of control depends on the choices we make now.  From national security to kindergarten classrooms, this is no longer a future problem.  It’s a right-now problem, and an opportunity.


AI’s Massive Energy Appetite:

Here’s a graph that maps out the explosive growth in data center power demand, largely driven by AI workloads from 2025 onward.  The “AI wedge” in electricity use is growing rapidly.

🔌  What the Graph Tells Us

1. Data centers powering AI are lighting up the grid

By 2026, total data-center electricity demand is projected to double compared with 2022, and the AI-driven portion – represented by the striped “AI wedge” – is expected to make up a growing share .

2. Running cities off algorithms

Consuming 5 – 9 GW of continuous power in 2025 translates to 46 – 82 TWh annually.  Roughly matching electricity usage of Switzerland or Finland

3. A looming strain on resources

Some forecasts indicate data centers (especially those focused on AI) may account for 6 – 9% of U.S. electricity by 2030.  Meanwhile, global AI-related electricity use is estimated between 1 – 2% of the world’s total.


Broader Context & Implications

Training vs. Inference:

  • Training giant models like GPT – 3/4 can burn through ~1,000 – 1,300 MWh (enough for 130 average homes per year) theverge.com.

  • Inference, or everyday use, adds significant cumulative load, each ChatGPT query uses ~0.34 Wh, with video processing using dozens of Wh wsj.com.

GPU-driven power surge:

Carbon & water footprint:

  • AI infrastructure is contributing to rising carbon emissions: Google, for instance, saw a 27% year-over-year increase in electricity usage and a 51% jump since 2019 en.wikipedia.org+2theguardian.com+2theverge.com+2.

  • Water-intensive cooling is also becoming a challenge; some data centers are expected to use 4–6.6 billion m³ of water by 2027 en.wikipedia.org.


⚖️ What Comes Next?

Edge computing as a greener alternative

Processing AI workloads on-device (e.g., smartphone chips) can cut energy use by as much as 90% per query axios.com.

Policy & Infrastructure moves

  • Governments (e.g., via executive orders) are accelerating energy infrastructure development. Think faster grid connections and nuclear deployment to meet AI’s demands reuters.com.

  • Energy regulators recognize that data center demand could drive regional electricity use beyond 5 – 9% by the end of the decade .

Efficiency through innovation

  • Advances like large-batch training saved ~80% of training energy in prototypes arxiv.org.

  • Hyperscalers are optimizing co-location, cooling efficiency, and carbon-aware scheduling arxiv.org.


🧩 Taking Stock

This isn’t a small box you plug in, it’s an energy revolution that demands national-scale infrastructure, smarter regulation, and tech solutions that look beyond chip performance.

As AI accelerates, our power grids, cooling systems, and environmental policies must accelerate too.

⚛️ Powering the Future: Nuclear Energy and the AI Revolution

As artificial intelligence advances toward artificial superintelligence (ASI), one constraint could stall everything: energy.  AI’s insatiable appetite for electricity has data centers consuming gigawatts like never before.

So where will all this power come from?

The answer, increasingly, is pointing toward nuclear energy.


Why Nuclear? Because the Grid Can’t Keep Up

Traditional energy sources such as coal, natural gas, even solar and wind face significant bottlenecks in scaling up fast enough to meet AI-driven demand.

AI training clusters, especially for models like GPT, Grok, or Anthropic’s Claude, require 24/7 stable baseload power.

This is where nuclear shines:

  • Consistent output – no reliance on sunlight or wind.

  • High energy density – far more efficient per acre than solar or wind farms.

  • Low carbon footprint – essential in an AI era where emissions are rising sharply.

Peter Diamandis noted, “It’s not just a compute problem. It’s a power grid problem. And AI doesn’t care if the lights stay on, it needs uninterrupted electricity.”


The Nuclear Renaissance: Small and Scalable

We’re witnessing the early stages of a nuclear power renaissance, fueled in part by AI’s demands.  But unlike the behemoths of the past, the future is modular:

  • SMRs (Small Modular Reactors): Safer, faster to build, and cheaper. Companies like NuScale and Rolls-Royce are leading development.

  • Micro Reactors: Even smaller reactors designed to power campuses or clusters of data centers independently.

  • Fusion Energy: Still a decade or more away for practical use, but attracting billions in venture capital, including for AI-powered plasma modeling.

The U.S. Department of Energy recently streamlined the approval process for next-gen reactors, and private sector interest is surging. SoftBank’s $1 trillion energy infrastructure proposal even includes nuclear as a core pillar.


AI Needs Its Own Grid

Salim Ismail proposed a future where entire AI campuses from GPUs to researchers are powered by dedicated nuclear facilities.   In effect, building energy-sovereign AI zones, much like silicon fabs today have dedicated water and power supply.

Some states like Wyoming and Virginia are exploring public-private partnerships to co-locate AI infrastructure with next-gen nuclear plants, aiming for first deployment by 2030.


Final Thought

If AI is the new electricity, then nuclear is the new generator. As the world races toward superintelligence, it will need to pair breakthroughs in code with breakthroughs in kilowatts, and nuclear may prove to be the only energy source capable of keeping up.

“We’re not just building smarter machines,” said Dave Blundin. “We’re building machines that need their own cities, and nuclear may be the only way to power those cities.”


Alternative Press