🗞️ The $47B AI Bubble Just Popped (And Why That's Actually Good News)

Hey, it's Oliver, here's your AI update for this week!

In today's issue:

• AI funding crashes 67% as investors finally ask "where's the revenue?"
• Microsoft quietly kills 40% of their AI initiatives—the great winnowing begins
• New study reveals the AI productivity paradox: more tools, less actual work done
• DeepMind's breakthrough in protein folding just made drug discovery 1000x faster
• And more...

P.S. You are losing thousands of dollars a month on repetitive tasks. On Tuesday, I’ll show you how to win it back for the price of lunch.

Quick question. What do you think of the length of our articles?

Best Links This Week

My Must-Reads:

AI Trends & News

• AI funding reality check: AI startups raised $104 billion in first half of 2025, but activity outside AI is slow with fintech funding dropping 42%. The concentration is revealing winners and losers. (CNBC)

Tools & Software Finds

• Yahoo Japan goes all-in on AI: Yahoo Japan mandated generative AI use for all 11,000 employees, targeting a productivity doubling by 2028. Workers spend 30% of time on routine tasks that AI can optimize. (Tech.co)

Industry Moves

• DeepMind’s protein breakthrough: Isomorphic Labs is applying next generation AlphaFold model to therapeutic drug design, rapidly characterizing macromolecular structures for disease treatment. (Deepmind)

Worth the Scroll

• Watch how Alex Banks creates his own AI Avatar in a $5 million studio, it’s quite fascinating. (LinkedIn)

We recently launched our social media on X and LinkedIn — go give us a follow!

💰 1. AI Funding Concentration: $104B Raised But Non-AI Startups Struggle

AI startups raised $104 billion in first half of 2025, but outside of AI, fintech funding dropped 42% to $10.5 billion. While AI companies capture massive investment, the rest of the startup ecosystem is facing a funding drought.

What This Tells Us:

• Capital is concentrating in AI at the expense of everything else
• Traditional sectors are being starved of investment
• We're seeing the formation of a two-tier startup economy

The Reality Check: While everyone celebrates AI funding records, the broader startup ecosystem is quietly dying. The winners won't just be AI companies—they'll be the ones that survive the non-AI funding winter.

🧬 2. DeepMind's Next-Gen AlphaFold: Drug Discovery Gets an AI Upgrade

Isomorphic Labs is applying DeepMind's next-generation AlphaFold model to therapeutic drug design, rapidly characterizing macromolecular structures crucial for treating diseases. This isn't just faster protein prediction—it's AI becoming a drug discovery partner.

The Impact:

• Drug development timelines could compress dramatically
• Complex protein interactions become predictable
• AI moves from supporting research to leading it

The Bigger Picture: We're watching the moment AI stops being a tool and becomes a research collaborator. The next decade of medicine might be designed by algorithms, not just assisted by them.

📊 3. The AI Productivity Paradox: Study Shows AI Actually Hurts Developer Efficiency

New research from METR reveals a shocking truth: AI tools actually hampered software developer productivity despite expectations they would boost efficiency. Developers using AI tools completed tasks slower and made more errors than those working without assistance.

Why This Matters:

• The productivity promises of AI are being challenged by real data
• Context switching between AI tools and actual work creates friction
• We're seeing the first hard evidence that more AI doesn't equal better results

The Ripple Effect: This study represents a watershed moment—the first major evidence that AI productivity gains might be largely illusory. The winners won't be those with the most AI tools, but those who use them most strategically.

💸 4. The Two-Tier Economy: AI Captures 53% of All VC Dollars

AI startups received 53% of all global venture capital dollars in the first half of 2025, jumping to 64% in the U.S. Meanwhile, Q2 funding fell to $101.5 billion from Q1's record $128.4 billion, revealing a market increasingly divided between AI haves and have-nots.

Why This Matters:

• Capital allocation is creating stark winners and losers
• Non-AI startups are facing an existential funding crisis
• We're seeing the emergence of AI as the only "safe" investment category

The Reality Check: This isn't just about AI getting more funding—it's about everything else getting less. The next generation of unicorns won't just be AI companies, they'll be the non-AI survivors who found ways to thrive without venture capital.

1. 🎯 AI-Powered Decision Architecture

Smart teams use AI to map complex decisions before making them, creating decision trees that account for variables human brains can't track simultaneously.

The Process: Feed AI all decision variables, constraints, and success metrics. Have it model scenarios, identify blind spots, and create reusable decision frameworks.

Copy-Paste Prompt:

I need decision architecture for: [YOUR DECISION]

DECISION MAPPING:
- Key variables and their interconnections
- Hidden assumptions I might be making
- Downstream consequences I haven't considered

SCENARIO MODELING:
- Best/worst/likely outcomes for each option
- Resource requirements and timelines
- Risk factors and opportunity costs

Context: [paste relevant information, constraints, goals]

2. 🔄 AI-Enhanced Process Archaeology

Advanced teams treat their workflows like archaeological sites—using AI to dig up buried time-wasters and forgotten bottlenecks that familiarity blinds them to.

The Setup: Document your key processes with time estimates and handoff points. Feed this to AI for deep analysis of hidden inefficiencies.

Copy-Paste Prompt:

Analyze this workflow for optimization opportunities:

PROCESS ARCHAEOLOGY:
- Time sinks invisible to the team
- Redundant steps from historical reasons
- Handoff friction slowing everything down

OPTIMIZATION BLUEPRINT:
- Changes that would eliminate 80% of delays
- Steps that could be automated or eliminated
- Quick wins vs. strategic improvements

Current process: [paste workflow documentation]

3. 🧩 AI-Powered Knowledge Synthesis

The most effective teams use AI to connect disparate information sources and create insights no single person could generate, turning information overload into competitive intelligence.

The Workflow: Collect information from multiple sources, then use AI to find non-obvious connections and create synthesis reports that reveal invisible patterns.

Copy-Paste Prompt:

Synthesize these sources for non-obvious insights:

PATTERN RECOGNITION:
- Connections between unrelated data points
- Contradictions revealing hidden truths
- Weak signals predicting future developments

STRATEGIC IMPLICATIONS:
- Opportunities from trend intersections
- Actions based on synthesized intelligence
- Updates to our strategic assumptions

Information sources: [paste reports, feedback, data]

A quick note before you go

Thanks for reading this week’s Brain Bytes — I hope something here helped you move faster or think better.

How’d this one land?

• A ✨ Super Helpful!
• B 🤔 Some Good Stuff!
• C 😕 Could be better.

P.S. Want curated tool picks and content recs? Fill this 30-second form so I can tailor the drops to you: Fill out form →

See you next week, — Oliver