Welcome to another edition of Talsco Weekly

IBM i Brief: 🔧 IBM releases Power11 performance guide. 📊 IBM makes a big AI data bet with $11B Confluent deal.
AI: 🧱 AI projects fail due to cultural barriers. 🔧 GitAgent launches open standard for AI agents. 🧠 Trust in AI isn’t human trust. 🤔 AI agents may be slowing development velocity despite productivity claims.
Career: 🕰️ We’ve Been Here Before. 💻 Why IBM i Developers Should Still Learn to Code.
Cloud: ☁️ Kyndryl brings Cloud Uplift to Canada.
Development: 🗂️ IBM i automates Index Advisor cleanup. 🔧 One command change can silently wreck your test environment.
ERP: 🖥️ IBM launches AI-powered SAP migration tool.
Hiring: 🎯 Stop Obsessing Over the Tech Stack. Start Hiring Better.
Learning: 🎓 PUB400 shaped an IBM i career comeback.
Modernization: 🏭 AI agents are reshaping software development infrastructure. 🌉 LegacyBridge automates 5250 data entry with AI.
Security: 🔒 IBM i faces escalating security threats from AI.
Trends: ⚠️ System of record status won’t guarantee growth.
Vendor: 🏢 The AS/400 just got an AI unlock.

IBM i Brief

🔧 IBM releases Power11 performance guide

IBM releases a Power11 performance optimization guide while flagging a Java 8 bug affecting IBM i 7.3–7.6 across four PTF group levels.

📊 IBM makes a big AI data bet with $11B Confluent deal

IBM closed its $11 billion acquisition of Confluent on March 17, adding real-time data streaming to its watsonx AI platform. Confluent serves over 6,500 enterprises, including 40% of the Fortune 500. IBM stock rose 2.8% to $256, though shares remain down roughly 20% from their 52-week high. Analysts maintain a consensus Buy rating with an average price target of $319 — implying ~25% upside.

AI

🧱 AI projects fail due to cultural barriers

This article highlights that most AI failures stem from organizational issues rather than technical problems.

For AI, there are three recommended solutions:

expanding AI literacy across all departments so teams share a working vocabulary,
establishing clear frameworks for AI autonomy with rules for audibility and observability,
creating cross-functional playbooks developed collaboratively.

Success requires treating cultural transformation as seriously as technical implementation.

Takeaway: This happens time and time again — it’s organizational, not an issue with the tech. Does picking the right tech, matter. Of course. But the key is, is the organization ready. And is the organization ready depends on a number of things. One of them being leadership that can get buy-in and drive change.

🔧 GitAgent launches open standard for AI agents

For those IBM i devs exploring AI agents — here are some things to consider.

The new framework enables developers to define, version, and run AI agents natively in Git repositories. Agents become version-controlled config files that export to Claude Code, OpenAI, CrewAI, and other platforms.

The system supports deterministic workflows, compliance-ready governance for FINRA and SEC requirements, and reusable skill modules.

Developers can fork, customize, and collaborate on agents like open-source software—complete with CI/CD pipelines and instant deployment across frameworks.

Takeaway: The world shifted while you were writing RPG—agents are now the interface between users and systems. The opportunity is clear: build this layer yourself and control the value.

🧠 Trust in AI isn’t human trust

At the iCON conference this past week, there was a lot of discussion around the “trust” factor with AI. Of course this ties into security and privacy.

As with anything new, there is an adoption curve, and this rate also ties into how much we trust it—or, another way to say it is, can we rely on it.

Psychology research shows trusting people and relying on AI activate different brain regions. Users don’t need AI to be a “trustworthy partner”—they need it to be a reliable tool.

If AI is just another tool that we need to rely on, the key factors are: predictability, transparency, controllability, error acknowledgment, and value alignment.

Reframing from “trust” to “reliance” changes how we design and evaluate AI systems for real-world adoption.

🤔 AI agents may be slowing development velocity despite productivity claims

Engineers report AI agents produce more code but lower quality, causing outages and requiring increased cleanup time at companies like Amazon and Anthropic.

Career

🕰️ We’ve Been Here Before

The fear that AI will eliminate jobs isn’t new — it’s the same alarm raised about every transformative technology in history. The printing press, electricity, the telephone, the calculator, the personal computer — each one triggered waves of pushback, panic, and predictions of mass displacement.

For over 100 years, automation has been “about to take all the jobs.”

It never does.

It transforms them.

IBM i developers bringing deep platform expertise to an AI-augmented workflow aren’t at risk — they’re the ones who will lead.

💻 Why IBM i Developers Should Still Learn to Code

AI writes code. That doesn’t mean you stop learning. IEEE Spectrum makes the case that deep coding knowledge is more valuable in the AI era — not less. Someone has to architect the system, review the output, and know when the AI is wrong.

Think of it like math in the age of calculators: you still need to understand the concepts to guide the tools.

For RPG developers, your platform expertise isn’t a liability. It’s your edge.

Cloud

☁️ Kyndryl brings Cloud Uplift to Canada

The company now offers its Cloud Uplift service in Microsoft Azure’s Canadian datacentre regions, enabling enterprises to migrate IBM Power workloads (AIX, IBM i, Linux) to Azure without rewriting code.

The solution addresses data residency requirements while providing access to Azure’s cloud services and AI capabilities, following Kyndryl’s May 2024 acquisition of Skytap.

Development

🗂️ IBM i automates Index Advisor cleanup

A new global variable in IBM i 7.6 TR1 and 7.5 TR7 called QIBM_SYSIXADV_BY_DAYS automates purging old Index Advisor recommendations. The default retention is 365 days, but administrators can modify it using CREATE OR REPLACE VARIABLE to set custom periods like 180 days for semi-annual cleanup, eliminating manual data management.

🔧 One command change can silently wreck your test environment

Restoring an IBM i library to a different name sounds simple — until your logical files point to production data, your SQL catalog falls out of sync, and your triggers quietly stop firing.

Nick Litten walks through five hidden failure modes (cross-logicals, SQL objects, triggers, stored procedures, journals) and the one parameter — DFRID(*DFT) — that solves most of them before they bite you. —Watch the 🎥 How to restore IBM i libraries How to restore IBM i libraries.

ERP

🖥️ IBM launches AI-powered SAP migration tool

The IBM Consulting Application Management Suite for SAP uses generative and agentic AI to help enterprises transition from legacy ECC systems to S/4HANA before 2027’s support deadline.

Despite available tools, only 45% of organizations have migrated.

Experts note migration tools don’t address core customization issues that created technical debt, meaning companies may upgrade without gaining new value from S/4HANA’s features.

Hiring

🎯 Stop Obsessing Over the Tech Stack. Start Hiring Better.

Every quarter, a new tool. A new platform. A new debate about which AI to adopt.

Meanwhile, the real question goes unasked: do you have the right people?

The companies that win — in any era, on any platform — aren’t the ones with the best tools. They’re the ones with talent that can see the signals, connect the dots, and cut through the noise. Systems thinkers. People who understand not just how the technology works, but why it matters to the business.

AI doesn’t change this. It amplifies it.

An average developer with a great AI tool is still an average developer. A great IBM i professional with that same tool? That’s a competitive advantage.

If you’re building a team — or trying to hold one together — that conversation starts here.

Book a call with Talsco →

Learning

🎓 PUB400 shaped an IBM i career comeback

Ravisankar Pandian credits PUB400.com as instrumental to his IBM i journey since 2013, helping him land his first job and navigate career growth.

After a three-year break from 2020-2023, PUB400 enabled his return to the workforce by providing free practice access.

He emphasizes the extraordinary value of offering IBM-grade infrastructure free for 20+ years to thousands of learners worldwide, thanking founder Holger Scherer for this sustained commitment despite lacking IBM support. 👏

IBM i-er, what has PUB400 done for you?

Modernization

🏭 AI agents are reshaping software development infrastructure

Industry observers compare the current moment to Henry Ford’s assembly line revolution, which cut Model T production from 12.5 hours to 93 minutes.

Today’s GitHub-based CI/CD systems were built for human workflows, but AI now pushes code faster than these pipelines can handle.

This mismatch is driving new practices like “harness engineering,” shifting focus from continuous integration to continuous guardrails that enable software agents to produce code safely at scale.

Takeaway: History repeats itself, as new technology comes in, with it comes new ways of doing things, thus automation. It should be a signal of opportunities to come, not the fear of what will not be.

🌉 LegacyBridge automates 5250 data entry with AI

The Latvian startup uses computer vision and LLMs to eliminate manual data entry on IBM i green screens. The system extracts data from documents like invoices and PDFs, then autonomously navigates 5250 applications to input data into appropriate fields.

Unlike rigid RPA tools, LegacyBridge adapts to unexpected screens and errors, escalating only when necessary. It requires no IBM i software installation and runs entirely through terminal emulation.

Security

🔒 IBM i faces escalating security threats from AI

AI just handed every bad actor a map to your system. Carol Woodbury of Kisco Systems showed how a user with zero special authorities can hijack a powerful profile in minutes — using nothing but Run SQL Scripts.

Agentic AI and MCP make it worse. The old “nobody knows IBM i” defense is gone.

Obscurity was never security. Now it’s not even obscurity.

Trends

⚠️ System of record status won’t guarantee growth

IBM i vendors, especially in the ERP space — pay attention.

B2B incumbents relying on low churn from being systems of record face a hidden threat:

incremental budgets are flowing to AI agents, not platform expansion.

Companies not owning their AI agent layer risk becoming commoditized infrastructure while third parties capture growth sitting atop their data.

The best incumbents are building native AI agents, pricing them as expansion revenue, and moving faster than feels comfortable to own the valuable AI workflow layer.

Vendor

🏢 The AS/400 just got an AI unlock

ZCOM 2.0, from Zanden Telekom founder Bill Herron — who first put the AS/400 on the internet in 1994 — is back with the same mission: bridge IBM i to what’s next.

This time, that means AI. Using MCP, RAG, and direct SQL, ZCOM 2.0 lets existing IBM i programs talk to AI agents without forced modernization.

The data was always the asset.

Now there’s a wire to it.

Join

Contact us

If you are an RPG programmer looking to explore opportunities or a client who is looking for a talented IBM i professional, please contact us. We look forward to assisting you.

Do you know of someone who could benefit from Talsco Weekly? If so, please use the social media buttons to spread the word. Thank you!