Nick Armstrong-Crews

When it comes to autonomous driving, Tesla and Waymo couldn’t be more different in how they approach the problem, and what they believe it will take to solve it. Waymo is all-in on the “hardware-heavy” school of thought: LiDAR, radar, HD maps, redundant compute, and highly curated geo-fenced environments. It’s a masterpiece of engineering, no doubt. And it works, but only in specific cities, under tightly controlled conditions, and with a very high operational cost. Tesla, on the other hand, is betting on vision and scale. No LiDAR. No HD maps. Just neural networks trained on billions of real-world miles, constantly updated from a global fleet of cars. Less hardware. Lower cost. More scale. And most importantly, a direct path to consumer vehicles, not just robotaxi pilots. From a tech perspective, the contrast is clear: • Waymo builds autonomy like it’s building a #NASA mission. • Tesla builds it like it’s scaling a #smartphone feature. I’m leaning toward Tesla’s approach, not because it’s perfect, but because it’s data-rich, less dependent on expensive hardware, and more #economically viable in the long run. Of course, the race isn’t over. #Regulations, edge cases, and safety validation will shape the outcome. But if autonomy is ultimately a software problem at scale, I think Tesla may be closer than we think. HITEC Angeles Investors The Tech Series Porsche AG Pirelli NASA - National Aeronautics and Space Administration General Motors Rivian BYD U.S. Department of Transportation OpenAI

349

325 Comments

I can see the future of autonomous taxis as they would certainly help with peak times demand and other use cases, but I can also see how there will be many challenges of making it really work at scale. Here is one, a very ''analogue'' problem, which I am sure leaves Waymo engineers wake at night. Local fleet management operations are crucial, meaning, in every city where AV taxis are launching, the underlying infrastructure forms the backbone and ensures the fleet runs smoothly.

235

74 Comments

🚨 Two AI Legends Just Said the Same Thing: LLMs Aren't the Path to True Intelligence Richard Sutton (Father of Reinforcement Learning, 2024 Turing Award Winner) and Yann LeCun (Meta's Chief AI Scientist) are both moving beyond LLMs. Here's why this matters for everyone: 🧵 The Problem They Both See 🤔 Current AI chatbots (like ChatGPT) are basically sophisticated copy-cats. They: Mimic what humans would say ✍️ Don't actually understand consequences Can't truly learn from mistakes Have no real goals beyond predicting text Think of it like this: A parrot 🦜 can repeat "the sky is blue" perfectly, but it doesn't understand what sky, blue, or "is" means. What Real Intelligence Needs 🧠 Both experts say we need AI that: 1️⃣ Learns like a child does 👶 Not by copying, but by trying things and seeing what happens 2️⃣ Understands the physical world 🌍 Can predict: "If I push this cup, it will fall" 3️⃣ Has actual goals �� Not just "predict the next word" but "achieve this specific outcome" 4️⃣ Plans ahead ♟️ Like chess players thinking multiple moves ahead, not just reacting The Future They Envision 🔮 Instead of bigger chatbots, imagine AI that: Learns continuously from experience 📈 Updates its understanding in real-time Can actually reason about cause and effect Builds mental models of how things work Why This Matters to You 💡 We're potentially at a turning point. The current AI boom might hit a wall 🧱, and the next breakthrough will come from completely different approaches. For businesses: Don't just bet on scaling current AI - watch for fundamentally new architectures For careers: Understanding real-world interaction and planning might become more valuable than prompt engineering The Bottom Line 📝 Two of AI's most influential thinkers are saying the same thing: We've been building really impressive calculators 🖩 when we should be building systems that actually understand the world. The next phase of AI won't just be ChatGPT-5 or GPT-6. It might be something entirely different. 🚀 What do you think - are we heading toward a major shift in how we approach AI? source: https://lnkd.in/gsHZvNbE https://lnkd.in/gchsKwsE #AI #GenAI #AGI #LLM #

41

3 Comments

My latest Substack is now live: Slacker Index: Why vertical integration is key for rapid hardware development In part 2 of a series, I break down what Slacker Index is and why optimizing lead times is so critical for enabling rapid, agile hardware development. https://lnkd.in/gwZczwdf

89

11 Comments

Speculation is essential for modifying the latency/throughput curve of deployed models. One of my new favorite resources for this is Perplexity’s glossary of speculative techniques: https://lnkd.in/gKnh2pQx It’s the perfect mix of approachability and technical depth!

69

I visited Santa Monica earlier this year and was surprised by how many self-driving cars I saw—especially Waymo’s. It felt like the future had already landed. But beyond the cool factor, I started noticing the smaller systems, needs, and business gaps forming around them. In this post, I explore not only the major infrastructure shifts but also the hidden micro-opportunities—some so accessible that even solo builders and small investors can get involved. https://lnkd.in/g2vbivA3

10

On September 29th Anthropic Killed a quarter of YC companies and almost no one is talking about it... They renamed Claude Code SDK to "Agent SDK" and gave everyone access to the same engine that powers Claude Code. Memory, context management, subagents, the whole system. What this actually means: you can now build Claude Code but for your specific workflow. Want to build: -Claude Code for legal (reviews contracts, flags compliance issues) -Claude Code for finance (analyzes reports, builds forecasts) -Claude Code for customer support (resolves tickets, searches knowledge bases) -Claude Code for research (reads documents, synthesizes findings) -Claude Code for marketing (maintains brand voice, creates content) -Claude Code for productivity (Preps meetings, provides insights, keeps you on taks) The infrastructure that makes Claude Code actually work is now a toolkit you can point at any domain. If you've been thinking "I wish Claude Code existed for [my industry] or [my business]" you can just build it now.

21

2 Comments

Big news that Waymo is expanding to more cities including Seattle and Denver 🚗🤖 It’s easy to forget that Waymo is an early example of an advanced AI agents in production, operating at scale, in a high-risk environments and it works reliably! Why? Because of countless hours of training and hardening in simulation before hitting real roads. Enterprise AI agents face the same reality. If they’re expected to operate in risky or high-stakes environments, they need robust simulation to prepare. At Veris AI, we build high-fidelity simulations so AI agents can learn and expand safely, like a Waymo.

53

2 Comments

The Power of Data: More Data. More Compute. Better Models. Waymo's recent blog post (https://lnkd.in/gydNP68D) highlights a powerful truth: performance in robot autonomy scales with data and compute. They show the impact of real data, how larger models, trained on vast diverse datasets, are yielding breakthrough capabilities in autonomous driving. This echoes a theme we’ve seen in the robotics community. Covariant, in their RFM-1 update last year (https://lnkd.in/gVrHx2Sz), made the case that the difference between good and great performance in robotic manipulation comes down to real-world data and massive amounts of it. At Diligent Robotics, we have always placed a high value on experience our robots can collect out in the WILD in the hands of real users. Our robots have logged hundreds of thousands of hours operating inside complex, dynamic indoor environments like hospitals where the stakes are high and the variability is endless. This hard-earned experience is what allows us to deliver reliable and scalable autonomy, coexisting gracefully with our human co-workers and adapting fluidly to real-world challenges. The future of robotics won’t just be built in simulation or labs. It will be forged in the noisy, unpredictable, people-filled places where robots work alongside us. I'm extremely proud to say that this is where Team Diligent shines. We've taken years to hone our operational execution muscle and this team is ready to go, exciting times! #Robotics #AI #Autonomy #DataDriven #PhysicalAI

55

1 Comment

If you've been to San Francisco recently, you'll notice something unusual.... Thousands of Waymo autonomous vehicle roaming the streets. When I started at Waymo back in 2019 it was more an R&D Project, than a bonafide commercial product. That's changed. Reports are indicating that Waymo now officially surpassed Lyft in ride-hailing market share in SF, with Uber potentially next in its sights. This isn't just about one company winning; it's a powerful signal about the shifting landscape of disruption, particularly how technology is now decisively trumping incumbency. For years, when we talked about competitive moats in tech, the "network effect" was king. Think Facebook: "People are on Facebook because other people are on Facebook." It created an almost impenetrable barrier. For rideshare companies like Uber and Lyft, their network effect was built on a two-sided marketplace: riders <-> platform <-> drivers. The more riders, the more drivers; the more drivers, the better service for riders, creating a virtuous cycle. But what Waymo is demonstrating is a profound disruption to this established idea, especially when AI and end-to-end automation enter the picture. Here's why it's different for self-driving: Waymo has fundamentally removed the marketplace element from the equation. Consider the traditional rideshare model: - Uber/Lyft: Riders <--- Platform ---> Drivers (a complex marketplace to manage, balance, and incentivize) Now look at Waymo: - Waymo: Riders <--- Platform ---> .... Nothing (no drivers to manage, no human supply/demand dynamics, just pure autonomous service delivery) They've replaced one side of the network with pure technology. The network effect, in its traditional sense, relied on the human element, the supply side of the marketplace. When you automate that supply side entirely with AI-driven vehicles, the competitive advantage shifts dramatically from a "people network" to a "technology and operational efficiency network." This isn't just a niche observation; it's a blueprint for disruption across industries. When you can automate an entire end-to-end process – removing the complex, variable human element from one side of a traditional marketplace – you redefine what a competitive moat looks like. It's no longer just about who has the most users or suppliers; it's about who has the most robust, scalable, and efficient autonomous system. For those of us in procurement and supply chain, this is a profound case study. Are we truly leveraging AI to remove complexity and build end-to-end automation, or are we just optimizing existing marketplaces? The Waymo story highlights that AI isn't just an efficiency tool; it's a fundamental re-architecting force that can dismantle long-held competitive advantages. What are your thoughts on this? Where else do you see AI dismantling traditional network effects? #AI #Automation #Waymo #NetworkEffect #Disruption #Procurement #Innovation

40

5 Comments

Nothing surprising in this article—except the LiDAR cost. There’s plenty to unpack in Tesla’s ongoing robotaxi saga—and the continued flawed insistence on a vision-only approach—but the idea that LiDAR costs $20,000 per vehicle is wildly outdated. Unless you’re building a high-end terrestrial mapping rig or surveying from a drone, that number just doesn’t hold up. In reality: Automotive-grade LiDAR now runs $500–$1,200 per unit, even lower in volume. Full autonomy typically uses 5–6 LiDARs, plus a sensor fusion and compute module ($3K–$5K). Total cost? Roughly $7K–$10K—at low scale. Companies like Hesai and Luminar are already delivering sub-$500 units in production. Even early-stage OEMs are sourcing solid sensors around $1,000 or less. And then there’s Moore’s Law—costs haven’t just dropped 80% since 2015, they’re still in free fall. #LiDAR #FutureOfMobility #SelfDrivingCars #MobilityTech #SensorFusion #Innovation #DeepTech #ComputerVision #AI #AutonomousVehicles #ArtificialIntelligence https://lnkd.in/gp-eHVUa

15

2 Comments

Robotics is at a turning point. The fundamentals—chips, connectivity, compute—have finally caught up, and with Viam, the tools to integrate them with hardware are widely accessible. That means engineers can stop asking, “Can I even get this to work?” and start asking, “What’s the right problem to solve?” We’re already seeing incredible answers emerging on Viam from startups like Gambit Robotics, Inc. building intelligent kitchen robots to enterprise teams driving new automation across industries. Thanks to Patrick Kulp for visiting the lab to see some of it firsthand and for the thoughtful feature in Tech Brew. Read the full article here: https://lnkd.in/ekKuNaCJ

76

1 Comment

Tesla releases detailed safety report after Waymo co-CEO called for more data Most founders think transparency is optional. It's not. Here's what just happened: Waymo's co-CEO stood on stage at TechCrunch Disrupt and called out competitors for hiding their safety data. She didn't name Tesla. She didn't have to. Everyone knew who she meant. A few weeks later? Tesla released the most detailed safety report they've ever published. 5 million miles between major collisions using FSD. 699,000 miles for regular drivers. That's 7x safer. Here's the problem you're facing: Your competitor is telling a story about your industry. They're setting the standard for transparency. They're defining what "responsible" looks like. They're making you look like you're hiding something. Even if you're not. → In PR, this is called losing control of the narrative. The best way to respond when a competitor attacks? Instead of playing defense, own a different metric entirely. Here's what Waymo did brilliantly: They made transparency the competitive weapon. Not better technology. Not lower prices. Not faster deployment. Transparency. And it worked. Because now every article about Tesla's data mentions that Waymo called for it publicly. ✅ Define the standards before your competitors do ✅ Make transparency your competitive advantage ✅ Force competitors to respond on your terms ✅ Turn data releases into PR moments Not: ❌ Waiting until you're called out ❌ Releasing data reactively ❌ Letting competitors set the narrative ❌ Assuming your product speaks for itself Tesla's numbers are impressive. But Waymo already won the PR battle by being first to demand transparency. Your competitors aren't just building better products. They're building narratives that force you to respond. PS: Are you setting the transparency standards in your industry? I put together a free guide with: • Full Forbes pitching playbook • Forbes pitch template you can copy • Proven viral content framework (100k+ views) 📌 Get free instant access: https://lnkd.in/dSuJb9Dr

44

25 Comments

The aerospace industry hasn't exactly been the poster child for reinvention. 50 years of tweaks, paint jobs, and jargon. Then Astro Mechanica shows up in SF and suddenly incumbents look uncomfortable. Founded in 2021 by Ian Brooke, a full-stack mechanical engineer with a pilot license and an obsession with building machines that move faster, Astro Mechanica is going all in on one of the boldest #propulsionsystems in aerospace. The Duality #turboelectric adaptive engine isn't a sketch, it's a test-fired reality. Built in months, not decades, on half a million dollars of off-the-shelf hardware. Astro Mechanica calls itself a vertically integrated engineering collective. They're designing propulsion, #airframes, and systems the industry hasn’t reimagined in half a century. This isn't incrementalism, it's an architecture shift. The Duality runs like a shapeshifter: turbofan at takeoff, turbojet near the sound barrier, ramjet past Mach 2. Electric motors decouple fan and compressor from the #turbine, unlocking efficiency across all regimes. That's physics, finally hacked to serve speed without burning fuel like it's 1973. Validation is stacking up. First hot-fire test in October 2024 at 30% power. A proof-of-concept in two months. Flight-ready in three. Partnerships with #AFRL, #INDOPACOM, and the #RapidCapabilitiesOffice already in play. Projections that it could power an unmanned aircraft the size of Anduril Industries’s Fury nonstop across the Pacific at Mach 2–3. Investors aren’t just watching, they’re betting. In April 2025, Astro Mechanica closed a $27.1M Series A co-led by Andreessen Horowitz, Lowercarbon Capital, and Giant Step Capital, joined by Alumni Ventures, Voyager Capital, Y Combinator, Pioneer Fund, and Dolby Family Ventures. 4 months later, United Airlines Ventures added an undisclosed check. For a global airline, this isn’t a trophy, it’s a hedge on the future of long-haul flight. Ian Brooke is steering the vision with COO Ashley Pelzel, VP Defense Systems Richard Brisbin, and Director of Ops Rohan Trivedi. On the technical side, Patrick McGuire drives propulsion, Matthew Perkins leads electrical/software, and engineers like Tyler Sandberg, Scott Hightower, and Alex Kramer are pushing hardware past limits. Shirley Lee is carrying the story, because companies this bold need sharp voices too. Astro Mechanica isn’t chasing speed for speed’s sake. It’s after a new standard of mobility: trans-Pacific supersonic demos in 2-3 years, affordable Mach 2+ travel in a decade, defense contracts now, #commercialservices next, and maybe space #launch plays that make rockets look overpriced. If they make high-speed flight flexible, accessible, and sustainable, the way people and goods move will never look the same. #Startups #StartupFunding #VentureCapital #Aerospace #Defense #DefenseTech #DeepTech #Mobility #Technology #Innovation #TechEcosystem #StartupEcosystem If software engineering peace of mind is what you crave, Vention is your zen.

7

1 Comment

The autonomous vehicle industry is at a crossroads. Waymo's co-CEO Tekedra Mawakana recently warned that robotaxi companies must go beyond marketing and deliver concrete safety proofs to regulators and the public. With recent incidents involving competitors like Cruise, the pressure is on to establish standardized safety metrics and transparent validation processes. Key challenges include: 1. Data transparency: Companies must share safety test results and incident reports openly. 2. Human-in-the-loop verification: Autonomous systems need rigorous human oversight during edge-case scenarios. 3. Regulatory alignment: Collaborating with agencies to create measurable safety benchmarks. For tech founders building mobility solutions, here's what to prioritize: - Invest in real-world safety testing with third-party validation. - Develop public dashboards to track safety KPIs in real time. - Advocate for industry-wide safety protocols that regulators can audit. The path to mass adoption isn't just about innovation—it's about trust. How will your team balance speed with safety in autonomous tech development? Let's discuss.

9