IBM i memory‑bandwidth advantage and Open Memory Interface; IBM’s partnership with OpenAI and AI integration; open vs. closed AI models; developer tools and open‑source repositories for IBM i; AI‑driven hiring trends and remote work considerations; learning resources and self‑directed query techniques; security patches covering 40 vulnerabilities; Nvidia’s reliance on Micron memory and its relevance to Power systems.
Welcome to another edition of Talsco Weekly
- IBM i Brief: 🤝 IBM goes all in on OpenAI for the enterprise.
- AI: ⚔️ The open vs. closed AI war is an old story.
- Development: 🧰 A toolbox for IBM i developers who tinker.
- Hiring: 📈 Heavy AI Spenders Are Hiring, Not Cutting. ⚖️ Companies can hold out for on-site talent. For how long?
- Learning: 💡 Great teachers build the conditions. Students do the work. 🔑 One query unlocks every IBM i Service.
- Security: 🔒 Forty Vulnerabilities Land in One PTF Guide.
- Did You Know: 💡 Nvidia needs Micron. So does your Power box.
IBM i Brief
🤝 IBM goes all in on OpenAI for the enterprise
IBM bets its consulting arm on OpenAI.
IBM just joined OpenAI’s Elite partner tier, embedding GPT-5.6, Codex, and ChatGPT Work into IBM Consulting Advantage.
That is the cost-of-change thesis at enterprise scale. Anyone running mature systems knows the weight of that legacy. IBM is betting the way forward is AI-led, secure, and delivered through consulting. The direction is set.
AI
⚔️ The open vs. closed AI war is an old story
We have fought the open versus closed war before.
Open source, open weight, or closed AI models. The industry is picking sides again, and if it feels familiar, it should. Operating systems collapsed from dozens of platforms down to a handful, leaving IBM i as the last supported proprietary system in the midrange.
Application software ran the same play. Companies once owned their code, then traded control for speed when Y2K and SaaS pushed them toward commercial ERP suites.
AI now poses the identical choice. Own your stack, or wait for your vendor to add the features. Either way, you pay. The names change. The tradeoff does not.
Development
🧰 A toolbox for IBM i developers who tinker
Open the hood and read the code.
Bob Cozzi wrote the book on RPG IV. His GitHub is a working example of learning by doing (see Learning below) .
Thirty-five open-source repositories, most built for IBM i and VS Code. rpgiv2free converts fixed-format RPG IV to free format. SQLTools and open-UDTF add high-performance SQL functions. db2json turns query results into JSON. Clone one, read how it works, and adapt it.
This is the autodidactic path in practice, not theory.
Hiring
📈 Heavy AI Spenders Are Hiring, Not Cutting
Heavy AI spenders are hiring, not cutting.
Ramp and Revelio Labs tracked AI spending across more than 21,000 U.S. firms. The heaviest adopters grew headcount 10 percent over two years. Entry-level roles grew 12 percent. Light adopters saw no change. The study is white-collar and correlational, so read it as a signal, not a rule.
Takeaway: For IBM i shops staring down retirements, it reframes the question: AI adoption and hiring juniors are not opposites.
⚖️ Companies can hold out for on-site talent. For how long?
The best people often want remote only.
Companies keep tightening return-to-office rules, with full-time mandates climbing fast. Some of that makes sense. At a small shop the IBM i developer also owns the hardware and the help desk, and presence matters. But here is the field reality. If the talent you need will not come on-site, remote is the only door left. And some of the very best people only want remote. A company can hold out for someone willing to come in, and that is a fair choice. But every company sits in a different spot, and there is no clean answer.
Learning
💡 Great teachers build the conditions. Students do the work.
Don’t steal the revelation from your junior developers.
The next generation of IBM i talent will not learn from a data dump. This short piece makes the case that real learning is autodidactic. The teacher sets up the trust and the tension. The student finds the aha and owns it.
For anyone mentoring at COMMON or inside a shop, the lesson is simple. Stop handing over answers.
Start engineering the moment where a developer discovers one for themselves.
🔑 One query unlocks every IBM i Service
Stop waiting for an expert to hand you queries.
The piece above was about self-directed learning. IBM Champion Pete Massiello hands administrators the SQL version.
QSYS2.SERVICES_INFO catalogs every IBM i Service on your system, complete with a working example for each one. Filter on ‘%JOURNAL%’ or ‘%SECURITY%’ and IBM shows you what it already built. No waiting for someone to teach you.
Run the catalog, copy an example, explore.
That habit is what turns a good administrator into a great one.
Security
🔒 Forty Vulnerabilities Land in One PTF Guide
Patching is not optional. It is the cost.
Volume 28, Number 28 dropped 40 security vulnerabilities on IBM i shops, plus 22 HIPER firmware patches spanning Power9, Power10, and Power11. Some fixes are still pending. A number that large is not noise. It is a signal about how much attack surface a modern IBM i stack now carries. Skipping a patch window does not save you the change. It defers it, with interest. Schedule the work before someone else finds the opening.
Forty patches. Who is applying them?
If that question does not have a clear answer at your shop, we should talk. I place IBM i security talent, and when an outside vendor is the better fit, I will tell you that too. My job is getting you to the right resource, not only my resource. Reach out and I will point you in the right direction.
Did You Know
💡 Nvidia needs Micron. So does your Power box.
Micron just posted record numbers. Nvidia can’t get enough of its chips. The AI boom turned memory, not processors, into the hottest thing in the data center. 🔥 Micron’s most advanced memory is now built specifically for Nvidia’s AI platforms →
Here’s why.
The GPU does the computing — the actual math for AI, graphics, all of it. Memory (like Micron’s HBM) stores the data and feeds it to the GPU fast enough to keep it working.
👨🍳 Think of it as a chef and the ingredients on the counter. A fast chef starves without ingredients delivered quickly. That’s why memory bandwidth, not just the GPU, is the real constraint.
Now the part the headlines skip. 👇
IBM saw this coming years ago. Starting with Power9, IBM pulled memory off the old on-chip DDR controllers and onto a serial interface called OMI — the Open Memory Interface — a deliberate move to raise bandwidth. The modules in your Power box aren’t standard DIMMs. They’re DDIMMs, engineered to break the exact bandwidth ceiling the industry is panicking about now. Power11 doubled it again.
And the memory makers behind those DDIMMs? Micron was one of the founding suppliers, the same company now feeding Nvidia’s AI accelerators.
So while the market treats memory bandwidth like a 2026 discovery, your platform was engineered around it before the hype had a name. ⚡
The AI world is just now catching up to something IBM i has quietly run on for years. It won’t be the last time.
Join
Sign up for Talsco Weekly to get the latest news, insight and job openings for the IBM i professional.
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.
Share
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!