This Week in Startups – E2037 Summary

Episode Title: Grape-picking robots, startup investing for all, and the earnings numbers you need to know
Release Date: November 2, 2024
Host: Jason Calacanis
Guests/Co-Hosts: Alex Wilhelm, Dr. Chang Lu (Extend Robotics)

Overview

This episode of This Week in Startups dives into some of the hottest trends at the intersection of startups, technology, and markets. Jason Calacanis and Alex Wilhelm cover:

• Advances in robotics and AI, particularly grape-picking robots,
• US legislative efforts to democratize private market investing,
• The latest major tech company earnings and what they signal,
• Innovations in global EVs, battery swapping, and cloud/AI infrastructure,
• The societal and labor impacts of automation and AI.

Their wide-ranging, energetic discussion is grounded in practical insights for founders, investors, and tech enthusiasts, with plenty of real-world examples and predictions.

Key Discussion Points & Insights

1. Robotics, AI, and Grape-Picking Robots (Extend Robotics)

Guest: Dr. Chang Lu, Founder/CEO, Extend Robotics
[03:44–21:59]

• Automation’s Next Frontier:

  • Jason introduces the concept of general vs. vertical AI using examples from past investments (Root AI for berries, CafeX for coffee).
  • The goal: Robots moving from narrow, repetitive tasks to more generalized roles in dynamic environments.

• Extend Robotics’ Approach:

  • Dr. Lu explains Extend creates software—rather than hardware—to enable teleoperation, VR control, and human/AI collaboration (human-in-the-loop robots).
  • Their software upgrades commercial robots, allowing real-time remote operation via VR and gesture control, enabling robots to "learn" from their human operators ([07:49]).
  • Threefold purpose:
    1. Real-time teleoperation; 2) Data collection for AI training; 3) Autonomous inference with human supervision ([08:46]).
  • Key Quote:
    “We’re proposing a more smooth transition...amplify labor, we don’t replace labor, we amplify labor with AI.” – Dr. Lu [09:48]

• Grape-Picking in Action:

  • Robots currently require human teleoperation, but will progressively automate as AI learns.
  • Training challenge: Handling delicate grapes requires advanced force sensors and precision to prevent fruit damage ([13:22]).
  • Collaboration: Working with Queen Mary University and Saffron Grange vineyard on custom grippers (end-effectors).
  • Timeline for commercial, fully autonomous grape harvests: 3–5 years ([14:57]).
  • Societal Impact: Discussion on skilled labor, migration, and the notion that the world’s best operator could “virtually” oversee global tasks through robots and AI ([15:24], [17:14]).
  • Quote (Calacanis):
    “If you had the greatest pilot of all time, you could have that pilot in every single plane. You could have the person who manages vines the most perfectly manage every vine in the world.” [15:24]

• Technical Details:

  • Their system streams 3D data at ~150ms latency, compatible with Wi-Fi, 5G, and potentially Starlink, making global remote operation possible ([17:44–18:40]).

• Other Applications and UK Startup Climate:

  • Extend’s tech also applies to nuclear, space, and manufacturing robotics; agriculture is a recent focus ([19:25]).
  • Dr. Lu describes the UK as a strong environment for software/AI startups due to government grants, innovation support, and VC presence ([21:05]).

2. Second-Order Effects of AI & Robotics

[21:59–28:08]

• Elevating Global Standards with AI:

  • The “perfect practitioner” in any field could be replicated everywhere via AI/robotics—raising quality and reducing cost, e.g., in winemaking ([22:33]).
  • Jason and Alex discuss how automation plus global broadband enables work to be done remotely, e.g., vineyard workers in Mexico operating robots in Napa via VR ([17:14]).

• Labor & Salary Convergence:

  • Trend toward a global standard for tech and admin job salaries as automation and teleoperation enable cross-border work ([28:08–30:44]).
  • Quote (Calacanis): “You’ll just see that the salaries converge on a global average for work. And then people will decide if they want the luxury item of having the person in person.” [28:49]

• Potential Negative Impacts:

  • Challenge for “B” level workers in developed countries to compete when “A+” talent can be virtually everywhere ([30:44–31:03]).

3. Broader Societal Themes – Immigration, Labor Shortages, Cultural Shifts

[31:59–33:02]

• Tech adoption as the inevitable response to labor shortages and restrictive immigration policies (e.g., cashiers → kiosks).
• How old industries (alcohol, agriculture) will be “dragged forward” by labor scarcity and competition—sometimes reluctantly ([31:03–31:59]).

4. Startup Investing for All: The Accredited Investor Debate

[34:30–48:33]

• The Problem:

  • Current US laws restrict startup/private market investment to “accredited” investors—those earning $200k+ or having $1M net worth ([34:57]).
  • This rules out most Americans from high-growth investments and wealth creation (e.g., getting early in Uber, LinkedIn).

• Legislative Changes:

  • The “Empowering Main Street in America Act” proposes letting more people invest—possibly via a test of knowledge, not just net worth ([38:41]).
  • The accredited standard has already “diluted” over time due to inflation (now ~18% of US households qualify, up from <2% in 1983) ([38:41]).

• Calacanis’s Perspective:

  • Wealth barrier does little to actually protect people (especially compared to sports betting, crypto, etc.).
  • Advocates for a robust test—like a driver's license—if needed, to broaden access and close the wealth gap ([41:54], [43:03]).
  • Quote:
    “The amount of appreciation in companies ultimately happens in the private market more than it happens in the public markets today.” [43:03]
  • Practical advice: Only bet a small share of net worth and diversify across 30–50 startups to have a shot at decent returns ([45:34–46:33]).

5. AI, Compute Constraints, and Cloud Infrastructure

[48:33–53:33]

• OpenAI’s Compute Bottleneck:
  • Even with Microsoft Azure, OpenAI is compute-constrained. Their operational costs for infrastructure reach $5–10B/year ([00:00–00:14]; [49:12]).
  • ChatGPT’s UI lauded for usability, but Google’s underlying models may still win on results ([00:14]; [49:12]).
  • Emergence of new interfaces (e.g., ChatGPT Canvas, AI voice mode) enabling new workflows (in writing, learning, tutoring) ([51:15–53:33]).

6. Big Tech Earnings—Google, Amazon, AI Trends

Google / Alphabet: [53:33–59:10]

• Q3 revenue: $88.3B (+15%); Cloud revenue up 35% YoY (to $11.4B), now profitable at nearly $2B/quarter ([53:33]).
• YouTube: $8.9B in Q3 ad revenue; $50B annual, $15B of which is now subscriptions ([51:41]).
• AI/LLMs increasingly driving search for mainstream users, even as early adopters like Jason/Alex shift queries to ChatGPT et al. ([54:32–54:50]).

Key Quote (Alex Wilhelm):
“Sundar Pichar during the earnings call said today, more than a quarter of all new code at Google is generated by AI, then reviewed and accepted by engineers.” [66:20]

Founders’ Lessons:
Hedge by developing multiple revenue streams (e.g., Google’s YouTube and Cloud arms could “save” the company even as search faces replacement threat from LLMs and AI assistants) ([58:04]).

Amazon / AWS: [68:10–72:06]

• AWS: $27B in Q3 (annual run rate $110B+); “re-accelerating” growth and 10.5B in quarterly operating income ([68:10]).
• AI cloud business is now multi-billion dollar run rate, tripling year-over-year. Potentially 30%+ of AWS revenue in 2–3 years ([69:59–70:36]).
• Increasing competition from Google Cloud and Oracle; startups benefit from “credit economy” but should mind long-term costs ([72:06–73:57]).

7. AI Business Models, Monetization, and Content

[59:10–63:39]

• OpenAI’s web search interface is up, compressing and summarizing third-party content. Sources cited, but links are “buried,” drawing criticism and likely legal challenge from publishers ([60:41–61:53]).
• Larger debate over “fair use” and responsibility to content creators versus aggregators/AI platforms.
• Voice-mode AI agents reaching desktop, changing user interaction style ([63:11–63:39]).

8. Societal Change—AI Tutors, Demystifying Opportunity

[51:15–53:33]

• Custom AI tutors now broadly accessible, democratizing what was previously a privilege of affluent families.
• Jason describes using ChatGPT to quiz his daughters on American history ([51:15]).

9. Global EVs, Battery Swap Tech, and Geopolitics

[81:05–87:32]

• Chinese auto innovation:
  • Xiaomi’s SU7 Ultra (1,500hp EV) tackles Nurburgring at supercar-speeds ([81:16]).
  • NIO’s 4th-gen battery swap station—fully autonomous swap in 3 minutes ([82:35–83:51]).
• Jason sees battery swapping as most viable for robo-taxi fleets, less so for private cars ([83:51–85:57]).
• Battery-as-a-Service future, and VTOL (air taxi) applications discussed ([85:57–87:19]).
• US/China tensions: US will likely block Chinese EV imports over security, tech, and competitive concerns ([87:32]).

10. Audience Q&A: Hottest AI Sectors for Startups

[74:39–77:25]

• Jason: Verticals where “perfect pilot” can be codified—tax, legal, healthcare, repair—are ripe for AI-driven disruption.
• Alex: AI companionship will be a massive opportunity as societies age and people seek digital relationships ([76:23–77:25]).

Notable Quotes & Moments

• “We amplify labor, we don’t replace labor, we amplify labor with AI.”
  — Dr. Chang Lu, Extend Robotics [09:48]

• “If you had the greatest pilot of all time, you could have that pilot in every single plane…. The most perfect practitioner for any behavior in the world.”
  — Jason Calacanis [15:24]

• “The amount of appreciation in companies ultimately happens in the private market more than it happens in the public markets today.”
  — Jason Calacanis [43:03]

• "More than a quarter of all new code at Google is generated by AI, then reviewed and accepted by engineers.”
  — Alex Wilhelm [66:20]

• "The perfect pilot for every profession is coming. The perfect professional for everything.”
  — Jason Calacanis [26:49]

• On OpenAI's compute bottleneck:
  “They're compute-constrained even with Azure in their pocket. That blew my mind.”
  — Alex Wilhelm [00:00], [49:12]

Important Timestamps

• 03:44 – Discussion on general vs. vertical AI, Extend Robotics intro
• 07:49 – How grape-picking robotics teleoperation/training works
• 09:48 – The “human-in-the-loop” transition in automation
• 14:57 – Timeline for commercial autonomous grape harvesting
• 17:44 – Global VR teleop: latency, Wi-Fi/5G/Starlink
• 21:05 – UK as a startup (software/AI) hub
• 22:33 – “Perfect pilot” theory, labor globalization
• 28:49 – Labor cost convergence & global salary implications
• 34:57 – What’s an accredited investor; barriers to private investing
• 41:54 – Tests for sophisticated investing, analogy to gambling
• 43:03 – Wealth creation happens in the private markets
• 51:15 – AI tutors, democratization of learning
• 53:33 – Google Q3 earnings breakdown
• 58:04 – Rationale for big tech’s revenue diversification
• 68:10 – Amazon/AWS earnings and the AI cloud business
• 72:06 – Cloud credits for startups; cloud infrastructure economics
• 74:39 – Audience Q: AI startup verticals
• 81:16 – Chinese EVs: Xiaomi SU7 Ultra / Nio battery swap
• 87:32 – US-China tech standoff, import implications

Tone and Style Notes

Jason and Alex keep the conversation fast-paced, skeptical-yet-optimistic, and full of industry war-stories and actionable insights for founders and investors. There is a practical, sometimes irreverent, "in the trenches" tone—often challenging mainstream regulatory logic or tech hype, but genuinely excited by innovation and the democratizing potential of AI.

Closing Thoughts

This episode bridges deep technical developments (robotics, cloud AI) with the practical realities of building, investing in, and competing in startups today. It weaves in geopolitical context, regulatory reform, and social impact—making it a must-listen for anyone interested in where tech and business are headed.

Next episode drops Monday, Wednesday, Friday. Send topic requests or guest recommendations to Jason and Alex!