Tim Poschel

Excited to share that Capital One has entered into an agreement to acquire Brex, a modern, AI-native software platform offering intelligent finance solutions that make it easy for businesses to issue corporate cards, automate expense management and make secure, real-time payments. https://bit.ly/3ZuzRgU

18

Enterprise #AI success requires strong data management. Capital One Tech leaders shared their approach for producing high-quality data that fuels AI innovation with VentureBeat. Check it out for their tips and insights. https://bit.ly/3FQzKpG

2

The ill effects of a modern Ai technology being used in unethical ways is real and the consequences are indeed unimaginable.

1

AI Is Not Coming for Your Job—But Someone Using AI Might Be We’re at a pivotal moment in the evolution of artificial intelligence. For the first time, advanced models like GPT-4, Claude, and Gemini are no longer just tools for researchers—they’re assistants for writers, analysts, coders, designers, and decision-makers. The narrative that "AI will replace humans" is outdated. The more accurate story is: AI will amplify those who are ready to work with it. Here’s what that means: AI is a force multiplier: It’s not replacing creativity, empathy, or critical thinking—it’s enhancing them. A junior marketer using AI can produce strategy decks, analyze campaigns, and write at the level of someone years ahead in experience. A solo founder with AI can punch above their weight, building prototypes, websites, and customer support flows that used to take a team. The new literacy is prompt literacy: Knowing how to speak the language of AI—how to prompt clearly, iterate rapidly, and refine outputs—is becoming as important as knowing how to use a spreadsheet or write an email. It's a skill that cuts across industries. There’s a growing gap between the curious and the complacent: Those experimenting with AI are discovering shortcuts, new ways of thinking, and entire workflows they didn’t know were possible. Those ignoring it may find their skillsets quietly becoming obsolete. Human + AI > AI alone: AI doesn't understand context, emotion, or consequence like we do. It’s a pattern engine, not a moral compass. The most powerful outcomes still require human insight, judgment, and intent. The question isn't whether AI will change your job. It’s how you’ll change with AI. Start small. Play. Experiment. Build. The future belongs to those who are willing to collaborate—with machines, and with each other.

1

Brilliant piece by Will Larson - https://lnkd.in/gkyRE-Q8 I find myself frequently describing LLMs as just another component in our toolkit, and agents as just standard software wrapped around API calls to LLMs that may or may not have access to “tools” - yup, more standard software. Of course, as this is all just software, you can combine it to your heart’s content, creating larger software systems including ones with an agent as the main orchestrator. The compounding non-determinism of systems like these will either feel like magic or a nightmare. Once you see it that way, it’s much easier to reason about.

12

Who remembers Doc Brown in 𝘉𝘢𝘤𝘬 𝘵𝘰 𝘵𝘩𝘦 𝘍𝘶𝘵𝘶𝘳𝘦, waving his arms around and declaring that the “flux capacitor” made time travel possible? That’s how many in tech view the mainframe: a mythical box where you send data, mysterious forces swirl, and results magically appear. But they are wrong - the mainframe is just a huge, scalable, very secure server. Yes, it runs a different operating system: not Linux, not MacOS, not Windows, but z/OS. Yes, its file system is quirky: with uppercase names, and hierarchies defined with “.”’s instead of “/”’s. The language of choice? COBOL. Which is verbose and readable. If you can read any programming language, I promise you can read COBOL. The mainframe is not the "flux capacitor" - it’s just another computing platform—different, yes, but approachable. Don’t fear it. Embrace it. You’ve got this. #MainframeModernization #zOS #COBOL #TechDemystified 

33

7 Comments

Most IT transformations focus on tools. The real shift is structural. In my conversation with Alex Gutman, CTO of Gilbane, Inc., a different model emerges; one where IT is no longer organized around projects, but around value streams tied directly to client outcomes. A few themes stood out: 1. Value streams over projects By aligning teams to how the business delivers value, Gilbane moves beyond fragmented initiatives to intentional, outcome-based execution. 2. Speed through accountability Persistent ownership and cross-functional alignment create faster learning loops and reduce friction between IT and the field. 3. Architecture that enables agility A flexible, modern stack allows the organization to adapt quickly to changing client and market demands. Listen to the full episode to explore how leading enterprises are making this shift here: https://lnkd.in/gSfpyFvn #Podcast #TechStrategy #TeamSetup #ValueStreams #Construction #CTO #CIO

35

5 Comments

🩺 AI can’t save lives if our data stay broken. In this article for Medical Economics, Errol Rodericks, Director of Product Marketing at Denodo, shares a wake-up call for U.S. primary care physicians: “AI can’t help if the data it relies on are fragmented, outdated, or locked away.” Every day, clinicians are being asked to do more with less. AI holds incredible promise to reduce administrative burden and deliver better care — but without trusted, connected, and context-rich data, those benefits will remain out of reach. 💡 In this insightful piece, Errol explores: - Why data fragmentation is undermining #AI in #healthcare - How modern data platforms unify information without rebuilding systems - The power of #DataProducts to turn insights into action - Why #DataGovernance and trust are the real foundation for AI success 📖 Read the full article here: https://okt.to/C8sJwQ #HealthcareAI #DataManagement #AIinHealthcare #DataForALL #NoDataLeftBehind #NoAIWithoutData

4

One thing I'm always advocating is CDC. Nobody needs it until they do. Eventually someone's going to ask for history and if you implement immediately when asked, you've got no historical view except for data *after* implementation. If you run a data system, it needs CDC. Put it in place before you need it. And even if a data vendor sends you "the whole history every time".... change track it. It will eventually save your bacon 🥓

11

1 Comment

𝐃𝐚𝐭𝐚 𝐂𝐨𝐧𝐭𝐫𝐚𝐜𝐭𝐬: Dashboard shows wrong numbers. Pipeline failed again. Why? Someone upstream changed a column name without telling consumers. We're all running on handshake agreements about our data until those agreements break at the worst possible time. 𝐃𝐚𝐭𝐚 𝐜𝐨𝐧𝐭𝐫𝐚𝐜𝐭𝐬 𝐚𝐫𝐞 𝐭𝐡𝐞 𝐟𝐢𝐱. They're like API contracts but for datasets. You explicitly define what you expect: - Schema and data types - Business definitions - Quality expectations - When data arrives and how fresh it should be 𝐖𝐡𝐲 𝐭𝐡𝐢𝐬 𝐦𝐚𝐭𝐭𝐞𝐫𝐬 𝐧𝐨𝐰: Teams own different data products (data mesh is real). AI models break when data drifts. Tools like dbt and Soda finally make this stuff practical. 𝐇𝐨𝐰 𝐭𝐨 𝐬𝐭𝐚𝐫𝐭: Pick one important dataset. Write down the rules. Set up alerts when rules break. Expand from there. It won't stop change from happening, but it makes change predictable instead of surprising. Full guide on my Substack: https://lnkd.in/gPHPxF9m #DataEngineering #DataQuality #DataOps #DataStrategy

19