Guillaume Binet

ISR + AI fusion is moving from lab demos to doctrine — it’s time to stop selling sensors and start selling decisions Recent exercises and platform tests highlight a dominant pattern: decision superiority, not sensor count, drives adoption. Sensor stacks that deliver: Actionable fused tracks Closed-loop engagements Explainable AI in the loop …are the ones clearing procurement barometers. This trend is visible both in global investments and in India’s test programs emphasizing data products and battle management. [Sources: The Times of India, YouTube] For OEMs with AI/ISR capabilities, three factors determine market success in India: Latency and resilience under contested spectrum Explainability for human-in-the-loop decision acceptance Integration rails to Indian BMS/C2 and logistics Demonstrate these, and your position shifts from vendor to capability partner. Several leadership teams are already exploring how to make AI-ISR outputs explainable and operationally trusted. Observing these patterns early can significantly shape board-level adoption strategy. #ISR #AI #SensorFusion #DecisionSuperiority #ContestedSpectrum #DefenseTech #OEMStrategy #MilitaryTech #Integration #GlobalSecurity

9

1 Comment

If intelligence is the log of compute… it starts with a lot of compute! And that’s why we’re scaling our GPU fleet faster than anyone else. Just last year, we added over 2 gigawatts of new capacity – roughly the output of 2 nuclear power plants. And today we’re going further, announcing the world's most powerful AI datacenter, located in southeastern Wisconsin. Fairwater is a seamless cluster of hundreds of thousands of NVIDIA GB200s, connected by enough fiber to circle the Earth 4.5 times. It will deliver 10x the performance of the world’s fastest supercomputer today, enabling AI training and inference workloads at a level never before seen. For AI training workloads, you need compute at exponential scale. That’s why we designed the datacenter, GPU fleet, and network together as one integrated system. This ensures a single job can run from day 1 at exponential scale across thousands of GPUs.  Fairwater uses a liquid-cooled closed-loop system for cooling GPUs that requires zero water for operations after construction. And we’re matching all of the energy that is consumed with renewable sources. And of course, it is just one of several similar sites we’re lighting up across our 70+ regions. We have multiple identical Fairwater datacenters under construction in other locations across the US, in addition to our AI infrastructure already deployed in over 100 datacenters around the world, powering model training, test-time compute, RL tuning, and real-time inference at global scale. Too often during times like this, people go with the current and only later wonder, how did we get here?  With Fairwater, we're charting a new path: doing the hard engineering work, bringing compute, network, and storage into one highly scaled cluster, and designing closed-loop energy systems to meet real-world computing needs. And partnering with local communities to ensure it's thoughtfully done in a way that is sustainable, creates new jobs, and expands opportunity.  We are thrilled to see this take hold in Wisconsin, and we are just getting started. Learn more about Fairwater here: https://lnkd.in/gpdni9gt

7,622

376 Comments

Over half a BILLION Space dollars in May!? Funding. Acquisitions. Hiring Trends. This month, $513M flowed into space companies scaling hardware, power systems & secure comms. Big rounds from True Anomaly ($260M), Apex ($200M), and mPower signal a shift to ops-ready infrastructure across LEO, GEO & cislunar. 🛰 Major M&A: - AeroVironment’s $4.1B acquisition of BlueHalo - IonQ x Capella Space ($318M) 📈 Hiring is hot in engineering – electrical, embedded, and software lead the pack – but sales & delivery roles are rising too. The space sector is scaling fast again! return of 2021-2022 season? Come on, You KNOW you want to be subscriber number 41,728 👉🏻 Subscribe on LinkedIn https://lnkd.in/eUHkyWEi #SpaceIndustry #VentureCapital #HiringTrends #EVONA

16

🚀 𝗦𝗵𝗶𝗽𝗽𝗶𝗻𝗴 𝗮𝗴𝗲𝗻𝘁 𝗶𝗺𝗽𝗿𝗼𝘃𝗲𝗺𝗲𝗻𝘁𝘀 𝘁𝘄𝗶𝗰𝗲 𝗮𝘀 𝗳𝗮𝘀𝘁 — Fieldy integrated Atla alongside LangSmith and immediately accelerated their development workflow across tens of thousands of production traces — Karolis Mariūnas Trusted by professionals with over $1.5M in sales, Fieldy's AI-powered wearable captures and transforms daily conversations into reminders, summaries, and tasks — building a comprehensive personal memory system from users' recorded conversations. By adopting Atla, Fieldy: • 𝗥𝗲𝗱𝘂𝗰𝗲𝗱 𝘁𝗼𝗼𝗹 𝗰𝗮𝗹𝗹 𝗿𝗲𝗱𝘂𝗻𝗱𝗮𝗻𝗰𝘆 from frequent occurrences to <0.5% of traces through refined system prompts • 𝗘𝗹𝗶𝗺𝗶𝗻𝗮𝘁𝗲𝗱 𝘂𝗻𝗻𝗲𝗰𝗲𝘀𝘀𝗮𝗿𝘆 𝗮𝗴𝗲𝗻𝘁 𝗵𝗮𝗻𝗱𝗼𝗳𝗳𝘀 and reduced latency by streamlining their multi-agent architecture • 𝗥𝗲𝘀𝗼𝗹𝘃𝗲𝗱 𝗰𝗼𝗻𝘁𝗲𝘅𝘁 𝗺𝗶𝘅𝗶𝗻𝗴 𝗶𝘀𝘀𝘂𝗲𝘀 that were diluting response accuracy when retrieving relevant conversation memories ⚡ 𝗧𝗵𝗲 𝗿𝗲𝘀𝘂𝗹𝘁: a faster, cleaner agent architecture that enables users to reliably query their personal conversation archive for insights and information. Read the full case study here: https://lnkd.in/eheBhZ8T

36

A couple of weeks ago, Nick Horob proposed the idea of building a tool that would analyze the sentiment of the Market Talk forum of AgTalk (for non-ag folks, it's an old school forum that's a bastion of poorly documented knowledge). It just so happens that I've been scraping data from AgTalk for nearly a year now, so I already had the data. https://agtalkpulse.com/ My biggest surprise from building it was how neutral the average sentiments are! Here's a quick breakdown of how it works: 1) A crawler consistently checks the active pages of AgTalk's Market Talk forum, scraping and persisting the posts. 2) When a new post is persisted, the Anthropic structured data API is used to assess a sentiment score and extract crop mentions and category tags. 3) A simple NextJS dashboard exposes the data for exploration. If you have any suggestions or feedback, feel free to share it! This was really just a small toy project so I could use my NAS a bit more! I'm happy to share the data with anyone who is interested.

34

8 Comments

Innovation … SpaceX style: In an earlier post I spoke about Qualcomm. Here I will attempt to describe how innovation works at SpaceX - at least from the vantage point I had. Instead of a narrative I will simply enumerate. 1) It all starts, in SpaceX, with incredibly ambitious missions and multi-year or decades long consistency centered on executing to those missions. This requires a world view, a point of view and boldness 2) Once picked - using deletion as a way to zero in on the essence of what needs to be built. Big hardware programs are complex to start with. Hence relentless simplification is the only way to both zero in what needs building and then having a shot at actually succeeding (most people can add; very few can delete) 3) At SpaceX there are no system engineering teams or architect roles. Not because architecture or systems engineering are not important. Rather they are too important to be left in the hands of one team or a small set of people. All engineers are architects and systems engineers. This means no one is waiting for the grand truth to be synthesized and disseminated. Architecture work, system engineering and building are widely disseminated into the daily act of building - hence innovation is constant, pervasive and a daily act 4) Constant iteration with fast feedback loops 5) Very little philosophizing (Feynman’s word). Mostly daily doing 6) No separation between engineering and manufacturing. Many things can be prototyped. Few can be built at rate. Rate manufacturing culture is embedded in the engineering culture at SpaceX. There are no “technology development” and “manufacturing” phases. The deletion culture also helps here because engineers who learn to delete end up with the smallest bill of materials (BOM). And all engineers are trained to work towards “buildable designs” and to own the full business outcomes of their designs 7) SpaceX innovation also came from a healthy disregard for expertise and experience and a heavy tilt to first principles reasoning (sometimes this was disconcerting for mid career people such as myself … but it eliminated hierarchies and positional power etc) 8) Reasoning risk at the right level. When building systems, not every component needs to be perfect. This means risks need to be reasoned at the right level with the intent to simplify the designs. The final thing has to work to some function and performance. 9) Avoiding margin stacking. Whether in building schedules (“green light”) or in building bill-of-materials. If some part goes through five partner steps - it lends itself to insourcing. Thus avoiding four margin layers 10) Finally, the culture is set up to entirely avoid overthinking, navel gazing, etc. It’s about relentless doing and problem solving in that context I had not experienced this kind of innovation culture before. But once in it, I fully appreciated and enjoyed it.

104

6 Comments

“But what about NATS?” - I hear this all the time—especially after my recent posts on MQTT. 𝐅𝐢𝐫𝐬𝐭: No single message broker or protocol is a silver bullet. Neither MQTT nor Kafka nor AMQP nor NATS. Never! 𝐒𝐞𝐜𝐨𝐧𝐝—𝐚𝐧𝐝 𝐭𝐡𝐢𝐬 𝐢𝐬 𝐰𝐡𝐲 𝐈’𝐦 𝐰𝐫𝐢𝐭𝐢𝐧𝐠: Synadia, the original donor of NATS, has told the CNCF it plans to withdraw the project, re-license it under the Business Source License (BUSL), and reclaim community assets such as nats.io and the GitHub organization. Industry experts (e.g., Alex Kritikos) have warned me for years that NATS wasn’t truly open source / had open standards. Now I see their point. 🔗 Full CNCF statement: https://lnkd.in/eAwxEdBe

39

12 Comments

Agentic AI - The Next Engineering Frontier The shift from reactive AI to proactive agents is accelerating faster than most anticipated. We're moving beyond single-task models to autonomous systems that can plan, execute, and adapt—fundamentally changing how we architect intelligent systems. Key trends shaping 2025: - Multi-agent orchestration replacing monolithic deployments - Tool-calling capabilities becoming standard across LLMs - Reasoning chains enabling complex problem decomposition - Human-AI collaboration patterns emerging as the sweet spot The roadmap ahead involves solving coordination complexity, establishing safety guardrails, and building robust evaluation frameworks. Benchmarks like GAIA remind us how challenging true general intelligence remains—even for seasoned engineers. As we scale these systems, the focus shifts from "can we build it?" to "how do we govern it responsibly?" The engineering challenges are immense, but so is the potential impact. Recently completed Hugging Face excellent agentic AI course—their practical approach to complex concepts is impressive. The hands-on framework really crystallizes the architectural patterns we're seeing in production #AgenticAI #Engineering #MachineLearning #TechLeadership #AIStrategy

64

8 Comments

Why we wrote this story: Robotaxies are a hot topic. GM agrees GM Repurposes Robotaxi Fleet to Test Tech It Plans for Your Car In an unexpected but strategic move, General Motors (GM) is repurposing its robotaxi fleet (formerly the Cruise Project) to test autonomous driving technology for personal vehicles. This decision comes after the automaker made the bold choice to discontinue its ambitious robotaxi project under the Cruise division, a venture that once promised to be the future of autonomous ride-hailing. The company is now focusing on evolving its advanced driver-assistance systems (ADAS) to work within the framework of everyday consumer vehicles. Read More: https://lnkd.in/evtPZHFP Subscribe to our newsletter! www.chargeduppro.com #ElectricTrucks #GreenTech #CleanTransportation #SustainableFleets #ChargingInfrastructure #ZeroEmissions #FutureOfTransport #EnvironmentalImpact #ClimateAction #RenewableEnergy #ElectricVehicleCharging #InfrastructureChallenges #CleanEnergyFuture #EVAdoption #GreenLogistics #TransportationNews #ClimateChangeSolutions #ChargingNetwork #FleetManagement #DrayageTrucks

9

2 Comments

Waymo’s “closed door” issue is silly and not a technology problem. It’s an observation as groundbreaking as “Is it going to charge itself?”  I’m over these nonsensical articles to rile up people who only read headlines and then become subject matter experts. As with any technology the hardware is never 💯 in the beginning. Let’s be clear: Adding powered doors to a vehicle is not hard. Minivans have had automatic sliding doors for 20+ years. The #Tesla #ModelX opens and closes every door in every direction with the touch of a “button.”  Sensors. Motors. Control logic. Done. When headlines frame this as some unsolved robotics challenge, it misses the point. This isn’t cutting-edge autonomy. Here’s what’s really happening: When you remove a human driver, every small friction point becomes visible.    •  Who closes the door?   •  What happens if a passenger forgets?   •  What if the car can’t move until it’s sealed?   •  Who owns that responsibility: hardware or software? In traditional vehicles, humans solve this. In autonomy, It’s door actuators. And they can solve the problem quicker than a human can exit the car, shut the door, reenter the vehicle and refasten their seatbelt. If doors can be automated reliably on a minivan, they can be automated on a robotaxi. The real question is: Are we solving root system design problems or overreacting to symptoms? #EVinfrastructure #robotaxi #FSD

33

7 Comments

Went deep with Heatmap News about the latest tariff fun times -- it's super important to realize that the US manufacturing ecosystem has atrophied enough that medium+ volume, non-defense startups are presently wise to start offshore until something changes. Have heard horror stories about cost structures simply not working unless you have sufficient volume to fully vertically integrate -- and that's before we increased tariffs on manufacturing inputs.

70

6 Comments

DoD’s new “Big Six” priorities are Applied AI, Biomanufacturing, Contested Logistics, Quantum & Battlefield Information Dominance, Scaled Hypersonics, and Scaled Directed Energy. They provide welcome clarity on where the department is concentrating its technology focus. I’m particularly pleased to see Applied AI elevated. From my experience driving AI adoption across SOCOM and the U.S. Space Force, AI needs to be operationalized across all DoD networks. The logical next step is clear: the DoD CIO should be tasked by SECWAR to make enterprise-wide Applied AI deployment a reality, leveraging a diverse ecosystem of AI companies to maximize speed, resilience, and innovation. Emil Michael, Doug Matty, Katie Arrington John Weiler, CoFounder IT-AAC Daniel "Rags" Ragsdale, Ph.D. Charleen L.

50

4 Comments

How do you maintain rigorous functional safety standards while scaling autonomous fleets across different global markets? 🌍 In this livestream, Riccardo Mariani (Head of Industry Safety), Richard Bramley (Platform Safety Architecture Lead), and Felix Friedmann (Product/Engineering Lead), explore how NVIDIA DRIVE Hyperion and DriveOS provide the high-performance, safety-critical architecture necessary for the next generation of transport. 🗓️ Add to your calendar: https://bit.ly/4biIbWL 📺 Tune in: https://bit.ly/4rjYHLR

12