Thom Kaleta

When modernizing legacy systems, we often focus on the tech, overlooking the deep personal investment in what came before. Dismissing that history as 'crap' often guarantees resistance, turning potential allies into roadblocks. Our new story in 'Facilitating Software Architecture and Design' features Michael Plöd sharing his experience with a stubborn system, a veteran architect, and the surprising power of simply asking 'why.' This isn't just about technical blueprints; it's about understanding people, respecting their work, and turning perceived threats into new opportunities. Empathy becomes a critical architectural tool. Read how an honest conversation can be the most effective facilitation: https://buff.ly/PuRqzMn

60

5 Comments

The Orchestration Graph Why Execution Is Free, Supervision Is Expensive, and the Firm Must Be Rewired Firms have always been shaped by their constraints. Labor was expensive. Coordination was difficult. Execution required time, teams, and overhead. These limitations didn’t just influence how companies operated, they defined why they existed, how they structured themselves, and what strategy looked like in practice. The enterprise was built in the image of its bottlenecks. Some people claim that with the rise of AI, one of the most fundamental constraints - execution capacity - is rapidly dissolving. If that assumption holds, then the question becomes: * How do firms operate when a foundational constraint disappears? * What happens to structure, to resourcing, to how success is measured? The firm becomes less like a factory and more like an operating system: a platform for managing distributed, intelligent execution. The next generation of firm strategy will hinge on redrawing these boundaries: * What work must remain in-house to protect quality, context, or IP? * What work can be delegated to external orchestration layers, partner agents, or API-based execution models? It’s not just a tooling shift, it redefines what the firm actually contains. In the traditional enterprise, coordination was manual. In the hybrid enterprise, coordination becomes programmable. This reframes the org chart. It’s no longer the primary map of how work flows. The orchestration graph is: the dynamic, often invisible network of people, agents, and systems connected by delegation logic, execution loops, and escalation paths. Designing a modern company means designing that graph. Not who reports to whom, but what gets routed where, under what supervision, and with what degree of autonomy. If orchestration becomes the new coordination layer, then what we measure must evolve with it. Productivity is no longer a count of actions, it’s a signal of how well a system is functioning. In many cases, the person is no longer producing; they’re designing the system that does. The core constraint of the modern firm is shifting: From labor to leverage. From execution to orchestration. From headcount to confidence. Link in comments. By Matan-Paul Shetrit #Orchestration #Graph #Business Connected Data London 2025 has been announced! 20-21 November, Leonardo Royal Hotel London Tower Bridge Join us for all things #KnowledgeGraph #Graph #analytics #datascience #AI #graphDB #SemTech #Ontology 🎟️ Ticket sales are open. Benefit from early bird prices with discounts up to 30%. 2025.connected-data.london 📋 Call for submissions is open. Check topics of interest, submission process and evaluation criteria https://lnkd.in/dhbAeYtq 📺 Sponsorship opportunities are available. Maximize your exposure with early onboarding. Contact us at info@connected-data.london for more.

28

1 Comment

MariaDB reacquires SkySQL, integrating AI-driven, serverless DBaaS into MariaDB Cloud, boosting flexibility and scalability for enterprise workloads - https://lnkd.in/gWAd3NNz “Our customers have made it clear that they want flexibility,” said Rohit De Souza, CEO of MariaDB. “Acquiring SkySQL helps accelerate plans and adds further innovation to MariaDB Cloud.” #DBaaS #AgenticAI #CloudDatabase #TechIntelPro

15

The real gains come not from unquestioningly adopting every tool but from aligning AI investments with business bottlenecks. https://hubs.li/Q03VjWH50 Written by Paul Kovalenko of @Langate Software

2

Vertical SaaS: The Hidden Gem of the Cloud 🚀 In the latest episode of Platform Builders, Christine Spang and Isaac N. sit down with John Melas-Kyriazi, CEO of Standard Metrics, to explore the power of vertical SaaS — a game-changer for businesses. John dives deep into how focusing on specific industries unlocks unparalleled value, the unique challenges of building tailored software, and how Standard Metrics is revolutionizing revenue optimization.💡 If you're serious about SaaS strategy, optimizing revenue, or curious about the future of vertical SaaS, you need to listen to this episode! 🎧 Listen here: https://lnkd.in/eXxy2YhU #SaaS #TechLeadership #VerticalSaaS #PlatformBuildersPodcast #Innovation

3

1 Comment

Just published: SaaS Roadmaps: How to Prioritize Features for Growth. Practical guidance for founders and PMs — gather user feedback, tie features to business goals, and prioritize what actually moves the metrics. Full article: https://wix.to/NPQKycT #SaaS #ProductManagement #Growth #Startup #Strategy

3

Building production LLM systems demands real architecture and rigorous evaluation, not just clever prompts and vibe-coding. Andrew McNamara & Charles Lee #RailsWorld talk explores building production LLM systems through Shopify Sidekick’s Rails architecture, covering orchestration patterns and tool integration strategies. They will cover LLM-based evaluation frameworks and demonstrate how robust evaluation systems become critical infrastructure for Reinforcement Learning pipelines, while examining how RL can learn to hack their own evals (and how to safeguard against this). https://lnkd.in/eKDb4shw

36

MCP: The New API for the Agentic Economy and Why Your SaaS Stack is Lagging The era of manual integration is over. If your software stack does not support the Model Context Protocol (MCP), it is effectively a legacy silo. As we transition into the Agentic Economy, the primary bottleneck for the enterprise is no longer the intelligence of the model itself. The real challenge is how that intelligence interacts with your proprietary data. MCP is the emerging standard that solves this. Here is why it is a non-negotiable for your current technical roadmap: 1. Unified Interoperability Instead of building bespoke, fragile connectors for every tool, MCP provides a universal interface. Your AI agents gain immediate, structured access to diverse data sources without the friction of custom middleware. 2. 10x Execution Speed By standardizing how models retrieve context, organizations are achieving ten-fold improvements in execution speed. This is the difference between an experimental chatbot and a production-grade autonomous agent. 3. Eliminating Contextual Debt Technical debt is no longer just about old code; it is about inaccessible context. MCP ensures your data is liquid and ready for any model you choose to deploy. The strategic shift is clear: The value of your SaaS stack is now measured by its ability to serve as a reliable context provider for AI. If your vendors are not adopting MCP, they are limiting your operational velocity. Is your organization prioritizing protocol standardization, or are you still building bridges to the past? #ArtificialIntelligence #SaaS #Technology #BusinessStrategy #ModelContextProtocol #AIAgents #Interoperability #EnterpriseArchitecture #AgenticEconomy #DigitalTransformation #FutureOfWork

1

"... Entire SaaS companies are at risk." 😦 LLM's (like Claude) had enough information in the public domain to learn coding and master it. Now entire SaaS companies are at risk. "That's a clear example of why you need your secret sauce to be protected." - Sami Issa of Global AI. Enterprises & Nations cannot have their information in a public cloud, or in the public domain. Only the paranoid survive. Remember... We are not a cloud. -------- Watch the clip below or the full episode on SiliconANGLE theCUBE YouTube channel. Thanks to John Furrier, Gemma Allen, SiliconANGLE & theCUBE, NYSE, #NYSE Wired, Brian J. Baumann for having us. #GlobalAI #SovereignAI

60

4 Comments

N/N's David Line answers some FAQs on AIO in this beginner's guide: -->How do LLMs differ from search engines? -->Will LLMs replace search engines? -->How do I get LLMs to mention our company in response to queries? -->Do we still need to optimise content for search engines? -->What about keywords? Are they still going to matter? -->What is the point of our website if people won’t click on it? -->Where should I start? Read the answers here: https://lnkd.in/gjec-BDM "A beginner’s guide, FAQ, how-to, explainer and vade mecum on AIO"

17

MariaDB Unveils Unified Cloud Database Platform Designed to Accelerate Agentic AI Application Development https://ow.ly/Q4Es50XgOap #MachineLearning #AdversarialAI #SmartAI #AI #EnterpriseAI #GenerativeAI #GenAI #MLAlgorithms #DeepLearning

9

Bento vs. SageMaker: Which inference platform is right for your team? For enterprise AI leaders, the question isn't "Which is better?"; it's "Which fits our scale, workflow, and infrastructure strategy?” The wrong choice can slow 𝘁𝗶𝗺𝗲-𝘁𝗼-𝗺𝗮𝗿𝗸𝗲𝘁, 𝗶𝗻𝗳𝗹𝗮𝘁𝗲 𝗚𝗣𝗨 𝗰𝗼𝘀𝘁𝘀, 𝗮𝗻𝗱 𝗽𝘂𝗹𝗹 𝗲𝗻𝗴𝗶𝗻𝗲𝗲𝗿𝘀 𝗮𝘄𝗮𝘆 𝗳𝗿𝗼𝗺 𝗶𝗻𝗻𝗼𝘃𝗮𝘁𝗶𝗼𝗻. The right one 𝗮𝗰𝗰𝗲𝗹𝗲𝗿𝗮𝘁𝗲𝘀 𝗔𝗜 𝗱𝗲𝗹𝗶𝘃𝗲𝗿𝘆, 𝗸𝗲𝗲𝗽𝘀 𝗶𝗻𝗳𝗿𝗮𝘀𝘁𝗿𝘂𝗰𝘁𝘂𝗿𝗲 𝗰𝗼𝘀𝘁𝘀 𝗽𝗿𝗲𝗱𝗶𝗰𝘁𝗮𝗯𝗹𝗲, 𝗮𝗻𝗱 𝘀𝗰𝗮𝗹𝗲𝘀 𝘄𝗶𝘁𝗵𝗼𝘂𝘁 𝗲𝘅𝘁𝗿𝗮 𝗵𝗲𝗮𝗱𝗰𝗼𝘂𝗻𝘁. Both platforms solve important but different problems. ⚙️ SageMaker offers the full AWS ML lifecycle. Best for teams who want end-to-end integration with AWS, from data prep to deployment. Powerful, but it requires some expertise to get the most out of it. 🍱 Bento Inference Platform is purpose-built for production inference. AI teams choose it for 𝗳𝗹𝗲𝘅𝗶𝗯𝗶𝗹𝗶𝘁𝘆, 𝘀𝗽𝗲𝗲𝗱, 𝗰𝗼𝗻𝘁𝗿𝗼𝗹, 𝗮𝗻𝗱 𝗰𝗼𝘀𝘁-𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝗰𝘆. They can deploy any model anywhere with tailored optimization, fast autoscaling, and a Python-first workflow that maximizes speed and control. 👉 Read the full breakdown: https://lnkd.in/gWvKW7jM #SageMaker #BentoML #AIInfrastructure

9