Sukanta Ganguly, PhD, MBA

YellowDog.ai just set a 10x benchmark uplift in scale computing, delivering 40,000 tasks per second (TPS) and managing 100,000 compute nodes in the cloud. What's even more interesting, that's 2x IBM Symphony and opens an intriguing pathway for these until-know closed/captive systems.

11

We chose a good subject for our march Autotech Council meeting, and after this year's CES, it seems even more obvious. Our meeting on the AI Defined Vehicle will show how it's a natural next step after the Software Defined Vehicle phase that we've been in Tesla launched the Model S in 2012. Now, most carmakers have SDVs. It's time to layer on the AI in user-facing roles, and in invisible functions managing the vehicle.

1

Barbarians at the Gate (https://lnkd.in/gkEBX79s) is a very interesting new paper from some great folks at Berkeley, with some exciting results showing the potential for AI in systems research. But I think the authors aren't quite asking the hardest problem about where this takes systems as a field. I wrote a new blog post about it: https://lnkd.in/gxS3s4v4 The core point of the paper is that AI can be effective at automatically driving the 'number go up' part of systems research. Given an evaluator, AI can automate progress on implementation. I believe this is 100% correct. The paper also talks about how more focus will have to go into developing evaluators. Again, I 100% agree, but think they're understating their case. Building evaluators (basically tests) is, I believe, going to become the focus of systems and software over the next decade. Which leaves researchers with the hard problem of choosing which problems to go after. That's never been easy in academic systems, and is only going to become more of a challenge. But the best systems research is driven by choosing problems well! This vision for AI could lead to lots more of the least interesting systems research. The "number go up" kind. Which could make it harder than ever to identify the truly innovative and visionary work. We're already seeing that elsewhere in science. So, mostly, I think the paper is right, but not yet asking the biggest question: what should systems research be?

276

12 Comments

AI has entered its systems phase. Outcomes are no longer driven by model size alone, but by how well models are integrated into context, tools, routing, and feedback loops. A small set of primitives is defining this shift: - Tool use turns models into operators, not just assistants. - Dynamic context (e.g. MCPs) makes context portable, composable, and governed at runtime. - Intelligent routing allocates the right model to the right task under real-world constraints. - Telemetry closes the loop, turning production behavior into learning. On top of these primitives, new alignment strategies – synthetic data, verifiable rewards, and multi-turn objectives – are allowing systems to learn from real outcomes rather than static labels. The implication: As capabilities commoditize, differentiation moves up the stack to system design, domain grounding, and alignment with measurable value. Pricing follows the same path, shifting from labor proxies to outcomes, and in some cases even risk-sharing. For builders: think more about control planes that govern routing, context, and accountability. For buyers: evaluate AI as an operator, one that can be audited, improved, and ultimately trusted with responsibility 🔗 More here: https://lnkd.in/gYfF7cWu

23

1 Comment

Vertical AI feels like a horserace 🏇 but after stepping back and thinking about the current state of market, I think what's equally unique is you have three different models competing at once. I think about it through this lens: 1️⃣ AI-enabled SaaS: the familiar path Take a proven vertical category and supercharge it with AI. Clear ROI, legible GTM, easy for VCs to understand. But real questions around differentiation and long term moats. 2️⃣ AI Roll-Ups: PE gets a facelift Acquire businesses, deploy AI to improve unit economics, and own the data + operations. Powerful if executed well, but operationally and financially complex. Big open questions around end state valuation. 3️⃣ AI-native industry disruptors: Become the thing Don’t sell software. Don’t buy incumbents. Build a fundamentally new version of the industry from scratch. Highest upside, but probably the highest casualty rate. You have to choose the model where the physics match the market. Each model has different capital needs, metrics, risk profiles, and founder skill requirements. I did a deep dive on each model 🤿 : ▪️where each model wins (and fails) ▪️how data dynamics actually differ across them ▪️why valuation expectations may diverge more than people expect ▪️and where I think the biggest outcomes will come from... As always, if you are building, investing, or just trying to make sense of vertical AI landscape, I always love to jam. Full post here: https://lnkd.in/g9a_9rtj

50

8 Comments

⚠️ Founders - consider the below reasons to not use Forward Deployed Engineers (FDEs) to acquire your *first* customers. This signals (1) lack of PMF; (2) miscalculation b/w development and sales timeframes; and (3) mischaracterization of Palantir Technologies timing when they deployed their FDE model. 💡 At the early-stage, it's crucial to demonstrate ability to repeatedly acquire customers with limited resources - capital, team, & time - that supports the narrative of 'more funding will amplify my traction.' Bespoke solutions for one company demonstrates zero repeatability. These custom solutions will not carry over to other customers within the same cohort. 💡 AI accelerates engineering but sales, system integration, and stakeholder buy-ins are perpetually slow. This imbalance results in concentrating all your engineering resources into one company, ignoring other opportunities, and still not going live on an accelerated timeline. 💡 Palantir didn't start using FDEs until year 3 when they got their first *gov't* customer, not their first customer. Gotham was already built and had strong PMF. Gotham allowed Palantir to take GTM risk by using FDEs b/c if it didn't work, Gotham was humming in the background. TLDR: At the early-stage, repeatable customer acquisition is the holy grail. Don't let AI deceive you that it can accelerate human behavior like it has with development cycles. It C.A.N.N.O.T.

33

1 Comment

In the latest episode of the OpenAI podcast I got talk to OpenAI CFO Sarah Friar and legend himself Vinod Khosla about the AI industry. Sarah explained the relation between compute and revenue (OpenAI's revenue keeps 3x'ing each year) and Vinod provided an insightful way to tell if we're in a bubble: API calls. Are they going up or down? (Spoiler: Up, up, up!) https://lnkd.in/g2tTQBdp

24

1 Comment

A recent arXiv paper proposes a rigorous rethink of how we model intelligence in multi-agent systems from the Google European scientific team. Its core idea is simple and radical: intelligent agents are not external decision-makers acting on a stable environment; they are embedded parts of a shared world that includes other learning agents. Why this matters: - Classical reinforcement learning struggles with non-stationarity once multiple agents learn simultaneously - “Theory of mind” reasoning leads to infinite regress (“I predict you predicting me…”) - Many paradoxes in coordination, alignment, and governance stem from treating agents as outside the system The paper offers a coherent alternative: - agents as predictive systems, not just optimizers - actions treated as random variables to be predicted, including one’s own - multi-agent reasoning handled at the level of shared universe models, not ad-hoc recursion This is not a practical algorithm. It’s a foundational framework, a new “north star” comparable to what AIXI once was for single-agent intelligence, which was introduced by Marcus Hutter in his book Universal Artificial Intelligence in 2005. For anyone thinking seriously about: agentic AI, AI-AI interaction, autonomous organizations, alignment and governance beyond toy models, this paper is an important signal of where the theoretical frontier is moving. Alexander Meulemans Rajai Nasser Maciej Wołczyk Marissa A. Weis, Seijin Kobayashi, Blake Richards, Guillaume Lajoie, Angelika Steger, Marcus Hutter, James Manyika, Rif A. Saurous, João Sacramento, Blaise Agüera y Arcas https://lnkd.in/e4-vd6VR

35

1 Comment