- Monday Thought
- Posts
- GPT-5, Robotaxis, and the Real Bottleneck
GPT-5, Robotaxis, and the Real Bottleneck
Why the future of AI depends less on models—and more on humans, habits, and hardware.
💭 Monday Thought
Hello friends, happy Monday!
What am I thinking about this week?
You don’t win by having the fastest AI—you win by making people trust it.
Everyone’s racing toward autonomy (ChatGPT Agent, robotaxis, enterprise agents), but trust in fully autonomous AI is dropping.
Your new coworker isn’t human, but still needs culture.
Workers are already bonding with AI more than with their colleagues. That’s weird, wonderful, and risky.
Infrastructure is the bottleneck, not ideas.
Everyone’s talking about AI breakthroughs, but your personal insight is: the bottleneck is physical.
AI is bumping into grid capacity, energy strategy, and real-world physics.
We had a special Ampersand Book Club virtual Lunch and Learn on Friday where Melanie Grubb led us through an interactive strategy session ala Blue Ocean Strategy. Thank you Melanie. It was terrific!
Sidenote:
On Saturday, Natalie, Olivia and I attended a special family fitness class. We had an amazing time.
If you know me, you know I am a lone wolf when it comes to fitness and exercise (Thank you FitBod, Runna, Strava and Apple), but I made a special exception to join this one time.
Are you a lone wolf or a wolf pack fitness person? What tech do you use?
Curious to learn about it through the poll at the bottom.
But for now, let’s get into it 👇
⚡️ Quick Hits
OpenAI’s GPT-5 Nears Release
OpenAI is poised to launch its next-generation AI model, GPT-5, in August. CEO Sam Altman has hinted at vastly improved coding skills and “reasoning” capabilities.
Robotaxi Race Accelerates
Waymo has surpassed 100 million fully driverless miles. Tesla is scaling up its own robotaxi trials and will expand to more cities by year-end.
Massive AI Infrastructure Investments
Google secured a $3B hydropower deal and plans to invest $25B in data centers. Meta is planning multi-gigawatt AI facilities under a project codenamed Prometheus.
🛰 Signals
Trust as the Key to Human-AI Collaboration
A Capgemini report finds AI agents could generate $450B in value—but trust in full autonomy is falling fast. Human-AI teams perform better when trust and oversight are baked in.
Humans Bonding with AI at Work
Upwork’s new research reveals 2/3 of heavy AI users trust AI more than coworkers. 64% feel closer to AI than teammates—hinting at massive shifts in work culture and collaboration.
AI’s Growing Energy Appetite
Generative AI could drive U.S. electricity demand from 4% to 12% by 2028. Each mega data center requires massive power—outpacing grid upgrades and energy investments.
🔨 Tool
Tana – Still my favorite thinking tool. It’s how this issue came together. If you’re building workflows, agents, or better systems of insight, it’s a must.
✍️ Prompt of the Week
“Pretend you are a founder from the year 2030. You just raised your Series A for an AI-native company in a post-agent world. What does your product do—and what assumptions are now obviously outdated?”
🧠 Final Thought
Have a great week everyone.
See you next week.
(Longer email than usual, so wrapping it up real quick!)
📊 Quick Poll – What Tech Powers Your Fitness?
Which of these best describes how you track or guide your fitness/health routines? |
🧵 Hit your answer above and I’ll share the results + insights next week!
If someone forwarded this email to you, and you like what you read, don’t be shy. Head over to MondayThought.com to read more and subscribe.
Reply