🚀 FP! Week In Review, Briefly #33

Also: Key Up Wing and Down Wing news from the week that was

In Case You Missed It ...

✨💪 AI can do the work. Companies still aren’t sure they trust it (Monday)

✨🏁 The AI race that may matter most: 3 scenarios (Wednesday)

📱 How social media is saving the AI revolution (Friday)

---

⤵ Up Wing/Down Wing

A selection of pro-progress and anti-progress news items from the past week.

⤴ Up Wing Things

• Anthropic agrees terms of $30bn funding deal at $900bn valuation - FT

• Wall Street Hearts AI, as Blockbuster IPO Shows - WSJ

• Notable Researchers Join $4 Billion Effort to Build Self-Improving A.I. - NYT

• Mira Murati Wants Her AI to ‘Keep Humans in the Loop’ - Wired

• Anduril Raises $5 Billion, as Push to Modernize the Military Accelerates - NYT

• The jobs apocalypse: a (very) short history - The Economist

• How AI Could Benefit Workers, Even If It Displaces Most Jobs - AI Frontiers

• AI Can Lead to a Fix of This Broken Government Program - Bberg Opinion

• Will AI turn us all into hipsters and artisans? - FT

• The School Trying to Rebuild Education for an AI World - TFP

• RIP, Liberal Arts Colleges. Long Live the Liberal Arts. - Bberg

• The newest AI boom pitch: Host a mini data center at your home - Ars Technica

• Can floating data centres meet AI’s huge energy demand? - New Scientist

• Data Centers in Space: A Pipe Dream, or AI’s Next Big Thing? - WSJ

• SpaceX and Google Are in Talks to Launch Data Centers in Orbit - WSJ

• Data Centers and Local Economies in the Age of AI: A Shift-Share Approach - SSRN

• The AI Boom Needs Carbon Removal - Heatmap

• Once again, SpaceX has set a new record for the tallest rocket ever built - Ars Technica

• Jeff Bezos’s Blue Origin weighs first external fundraising - FT

• A plan to make drugs in orbit is going commercial - MIT

• Italy prepares for return of nuclear power - WNN

• USA to examine SMRs for commercial shipping - WNN

• Fervo’s Hot Rocks Are Now a Hot Stock - Heatmap

• Regulatory Reform Is Headed for the Nation’s Largest Grid - Heatmap

• Routine vaccines may cut dementia risk—experts have startling hypothesis on how - Ars Technica

• How an ‘Impossible’ Idea Led to a Pancreatic Cancer Breakthrough - NYT

• A New Hantavirus Vaccine Is in the Works - Wired

• A Single Infusion Could Suppress H.I.V. for Years, Study Suggests - NYT

• The world of 2050: What’s actually possible - Big Think

• Frictionless Security and Supersonic Flights: What Travel Might Look Like in 20 Years - WSJ

• The World’s Most Surprising Capitalist Makeover Is Under Way in Sweden - WSJ

• Why personalised pricing could be a good deal for shoppers - FT

---

⤵ Down Wing Things

• Why A.I. Safety Controls Are Not Very Effective - NYT

• Overworked AI Agents Turn Marxist, Researchers Find - Wired

• Elon Musk’s Grok Is Losing Ground in AI Race - WSJ

• Peter Thiel-Backed Startup Bets Its AI Can Boost Hiring - WSJ

• Meta’s Embrace of A.I. Is Making Its Employees Miserable - NYT

• White-collar workers report growing feelings of ‘AI brain fry’ - FT

• This Is Why You’re Drowning in Busywork - NYT Opinion

• Women at the sharp end as AI takes over administrative roles - FT

• Korea’s AI ‘People’s Dividend’ Will Ensure Common Poverty - Bberg Opinion

• Why Americans dread AI - FT

• The AI Backlash Could Get Very Ugly - The Atlantic

• 7 in 10 Americans oppose data centers being built in their communities - Wapo

• Meet the Sad Wives of AI - Wired

• China’s Best and Brightest Tech Talent Is Going Back to China - WSJ

• Xi Jinping Is Planning for China’s Final Victory Over the U.S. - NYT Opinion

• A new US military wargame series began by simulating a nuclear weapon in orbit - Ars Technica

• Why birth rates are falling everywhere all at once - FT

• Peak Human Is Coming Sooner Than You Think - TFP

• Why U.S. Test Scores Are in a ‘Generation-Long Decline’ - NYT

• The real reason Americans hate the economy so much - Vox

• How American Cool Dies - NYT Opinion

• Funding for Early-Stage Climate Tech Is Drying Up - Heatmap

---

↕️ Which Wing Things?

• The Quadrillion-Dollar Disagreement on AI and the Economy - AI Frontiers

• Nvidia Is Buying the Chip Supply Chain - WSJ

• AI’s Next Phase Plays Into TSMC’s Hands - WSJ

• Taiwan Is the Key to AI Dominance - WSJ Opinion

• China Seeks A.I. Independence, Weakening Trump’s Leverage - NYT

• Life without US tech - FT

• The Shared Feeling of Being Harvested by the Future - NYT Opinion

• AI desperately needs more adult supervision - FT

• What happens when AI starts building itself? - TechCrunch

• The Democratic approach to AI is not all about bans - The Economist

• A.I. and Humans Battle It Out in a Cybersecurity Showdown - NYT

• The Tech Jobs That Are Safe From AI - WSJ

• AI Can’t Agree on Which Jobs AI Might Destroy - WSJ

• Does an LLM Hear Your Prayers? - WSJ Opinion

• Parenting Teens in the Age of AI Means Choosing Trust Over Control - Bberg Opinion

• Typing Is Being Replaced by Whispering—and It’s Way More Annoying - WSJ

• NASA provides some details about Artemis III, but hard decisions remain - Ars Technica

---

Essays and Q&As

✨💪 AI can do the work. Companies still aren’t sure they trust it

Up to the Task: New findings from research group METR suggests Anthropic’s Claude Mythos Preview can now autonomously handle some complex tasks that would take skilled humans significant amounts of time, up to around 16 hours, to complete. This news adds to growing speculation that AI systems are advancing toward more sustained, independent problem-solving.

Identifying AGI: Defining AGI remains an elusive task. To different companies and researchers, AGI means very different things:

To OpenAI, for example, AGI means outperforming humans at economically valuable work. To Google, it is mastery of most cognitive tasks. Anthropic co-founder Ben Mann prefers an ‘economic Turing test,’ where AI can perform a job without revealing itself as AI

The debate, however, feels increasingly less philosophical and more practical as these tools find their place in the market.

The Reality: In practice, companies care less about abstract AGI benchmarks than dependable performance. Enterprise AI systems must stay on task, avoid dangerous mistakes, explain decisions to regulators, and function reliably inside real organizations. For now, AI “agents” remain heavily supervised assistants, not autonomous digital workers.

Up Wing Up Shot: Goldman Sachs sees AI adoption accelerating as computing costs fall sharply, potentially driving massive growth in enterprise AI use over the next 15 years. Still, the transformation is likely to be uneven and incremental, spreading via individual tasks and workflows rather than arriving in a single dramatic breakthrough moment.

---

✨🏁 The AI race that may matter most: 3 scenarios