Talsco Weekly: The AI Decision Tree Just Got Messier

Welcome to another edition of Talsco Weekly

IBM i Brief: 📉 IBM’s worst day since 1968 wasn’t random. 📦 IBM Champions are already unboxing the Power11.
AI: 🌳 The AI decision tree is getting messier. 💻 Bob 2.0 arrives with an IBM i-specific upgrade.
Career: 🔔 The AI job panic letter doesn’t hold up. 📈 Technology doesn’t kill jobs. It moves them..
Database: 🗄️ Your AS/400 database might already be modern.
Hiring: 🧩 External talent just became a permanent strategy.
Modernization: 🛠️ Legacy systems aren’t the problem. 📖 AI can read legacy code. It can’t modernize it.
Open Source: 🔀 Why AI is quietly pushing IBM i toward Git.
Trends: 🤝 AI strips busywork. What fills the void?

IBM i Brief

📉 IBM’s worst day since 1968 wasn’t random

IBM shares plunged 25%, its biggest one-day drop by market cap ever, after a Q2 revenue miss. The real driver: AI-fueled data center spending pulled IT budgets toward servers and chips, leaving less for IBM’s mainframes and software. Investors are now pricing in a future where AI tools replace IBM’s software too. One analyst called it “a triple whammy.”

🔄 Update:

Shares clawed back 3.9% Thursday to near $219, though IBM remains well below its 50 and 200 day averages. Securities fraud investigations are piling on, but IBM is countering with new Power systems and AI focused software, insisting its software business isn’t being disrupted by AI. Markets aren’t sold yet. (More here)

🔍 Our Take:

We’re not outside observers on this one. IBM i shops run on IBM’s software, hardware and OS. When Big Blue has its worst day in decades, that’s home turf, not background noise.

Inside the IBM i world, IBM Bob has been the headline: an AI assistant built for RPG, and a real signal IBM is investing here. But the financial markets are telling a different story.

Investors are pricing in doubt about whether IBM’s software, AI included, can hold ground against outside competition.

Two things are true at once. IBM Bob is real progress for i shops. Wall Street isn’t convinced it’s enough yet.

Disclaimer: The market news on IBM will likely change by time of publishing.

📦 IBM Champions are already unboxing the Power11

Longtime IBM i admin Pete Massiello just unboxed the new Power11 S1112, calling it a must-have upgrade for every P05 shop, high praise from someone who has installed more machines than he can count.

AI

🌳 The AI decision tree is getting messier

Enterprises are weighing a flood of fast-changing AI tools while trying to avoid costly missteps. Licensing models are shifting, new platforms keep emerging, and leaders are still deciding whether to buy, build, or wait.

The risk: by the time a choice is finalized, today’s frontrunner may already look outdated. The pressure isn’t just to move fast, it’s to do enough diligence to avoid backing the wrong stack.

It’s not just buy vs. build vs. wait anymore, it’s also frontier model vs. open source vs. vendor platform, and each path carries a different risk profile. A 2026 enterprise survey covered by The Register found 81% of leaders are concerned about AI vendor dependency, but only 6% believe they could switch primary providers without real operational disruption, a gap that’s making teams slower to commit.

Part of the problem is pricing itself: vendors often win deals with generous pilot credits, then usage-based repricing and shifting license terms turn out to cost far more at scale than teams budgeted for.

See Bob 2.0 and the IBM i Premium Packages below.

That’s why more leaders are treating build-vs-buy-vs-wait, and now model-vs-vendor, as a per-decision framework rather than a single company-wide call — buy when AI slots into an existing workflow, build (or self-host open source) when it’s core to the differentiation, wait when the ROI or the technology itself is still unproven. For anyone who’s watched platform cycles play out before, the instinct holds: the tools will keep changing.

Our Take: While IBM Bob is a no brainer option to consider for IBM i centric shops. What about shops with a mix of tech stacks, and countless number of 3rd party packages that might rest on the IBM i or off?

💻 Bob 2.0 arrives with an IBM i-specific upgrade

IBM’s agentic coding tool got two updates June 24: a faster, unified Bob 2.0 core, and a Premium Package built for IBM i. The package lets Bob read and write source directly on QSYS or the IFS, adds Developer and Database Modes for RPG, COBOL, and SQL, and includes roughly 30 IBM i-specific tools. Jack Henry has used it to speed up RPG modernization. Pricing runs $20 to $40 monthly on top of base Bob costs.

Breaking It Down: Bob 2.0 vs. the IBM i Premium Package

Two upgrades. Two different jobs. Easy to conflate them.

Bob 2.0 is the engine. It’s a core upgrade available to every Bob user, on every platform. Faster performance, a bigger context window, smarter architecture under the hood. Everyone gets it.

The IBM i Premium Package is the specialization. It’s a separate, paid add-on built for IBM i shops specifically. It gives Bob direct access to QSYS and IFS source, fluency in RPG, COBOL, DDS, SQL, and CL, and roughly 30 IBM i-specific tools like running CL commands or launching 5250 sessions.

Simple version: Bob 2.0 makes Bob better for everyone. The Premium Package makes Bob useful for IBM i. You need the first either way. The second is what makes it worth paying for if you live on this platform.

Career

🔔 The AI job panic letter doesn’t hold up

This week’s open letter warning policymakers to prepare for AI-driven job losses leans heavily on hedge words like “could” and “may.” Critics note many layoffs already blamed on AI look more like post-pandemic over-hiring corrections. Building policy guardrails now, before the transformation even proves out, may be premature.

So if the alarm is overblown, what does history actually say?

📈 Technology doesn’t kill jobs. It moves them.

Time and time again throughout history, technological innovation has led to job growth, not job loss. Jobs certainly shift, with the appearance of new roles nobody had a name for yet.

The printing press didn’t end careers, it created publishers, editors, and translators. The assembly line didn’t erase labor, it built a middle class. The computer didn’t wipe out office work, it multiplied it.

AI won’t write a different story. The jobs that disappear will be replaced by ones we haven’t named yet. For IBM i shops, that likely means RPG developers become AI-assisted RPG developers, and operators become automation managers. The skill isn’t disappearing. It’s evolving.

The panic letters get the headlines. History tells a calmer story.

Database

🗄️ Your AS/400 database might already be modern

Moving DDS-defined physical files to modern SQL tables and views pays off: faster queries via the SQL optimizer, easier integration with REST, Node.js, and modern dev tools, and built-in referential integrity and security. Best part: no big-bang rewrite required. Hybrid setups, keeping old RPG programs running against new SQL objects, get you there gradually.

Hiring

🧩 External talent just became a permanent strategy

SAP’s latest research says the line between permanent and contract talent has broken down. With 74% of employers struggling to find skilled talent, leading companies now treat contractors and consultants as a standing part of workforce planning, not a stopgap.

IBM i shops face the same skills gap, from succession planning to modernization projects. Bringing in specialized RPG and Power talent isn’t a fallback plan. It’s essential.

Feeling the same talent crunch?

Book a quick call and let’s talk about getting the right IBM i talent in place, permanent or project-based. Our bench is 25 years deep.

Modernization

🛠️ Legacy systems aren’t the problem

A new analysis argues AI agents don’t expose outdated technology, they expose weak integrations, messy data, and loose access controls. Google’s own research found 83% of IT leaders think their infrastructure needs modernizing for AI, yet only 17% feel confident their current stack can handle it. For IBM i shops, that’s validating: the platform was never the liability.

“An older system is not always a problem. The narrative surrounding agentic AI often leads to the conclusion that legacy environments must be replaced with new, integrated platforms. In practice, a stable transactional system can still perform its function well.”

“The most important change, therefore, does not involve replacing the entire IT stack, but rather measuring and restructuring those pathways through which the agent actually creates business value.”

🔍 Our Take:

This is the argument IBM i shops have made for years, now validated from outside the community.

The platform was never the bottleneck. Integration, data quality, and access control were always the real work.

IBM i isn’t exempt from messy integrations either, but the database, OS, and hardware were built together, not bolted together after the fact.

It’s a head start, not a free pass.

Rip-and-replace was never the answer. Fixing the pathways an agent uses is.

📖 AI can read legacy code. It can’t modernize it

One example from the piece: AI coding assistants recently analyzed six million lines of RPG code running on IBM Power systems, handling roughly 30 million requests a day, surfacing technical debt in weeks. But visibility isn’t modernization. Deciding what to restructure, move, or leave alone is still a business judgment call, not something a model can make for you.

💭 Takeaway:

This is worth sitting with for IBM i shops leaning on AI tools like IBM Bob. Fast code analysis is genuinely useful, but it only tells you what exists, not what to do about it. Deciding whether to leave a program alone, wrap it with a new interface, or actually rewrite it still takes people who understand the business, not just the code. AI shortens the assessment. It doesn’t make the call.

Open Source

🔀 Why AI is quietly pushing IBM i toward Git

ARCAD Software’s leadership sat down with ITJungle to explain a decade of Git integration lessons: dual repos for security, SQL as the bridge to off-platform developers, and progressive rollouts that don’t strand veteran RPG programmers. Their sharpest point: AI tools need Git’s branching and pull-request isolation before anyone should let AI touch production code.

Breaking It Down: What Git Actually Is

Git is a version control system. That’s it.

It tracks every change to your code. Who changed what. When.

Linus Torvalds built it in ten days back in 2005, to manage patches on the Linux kernel. It worked. Then everyone else showed up. Today Git isn’t optional. It’s the default.

Before Git, RPG shops leaned on source physical files and library-based promotion: check out a member, lock it so nobody else could touch it, make your change, promote it from DEV to TEST to PROD one library at a time. Some shops layered commercial change management tools like Rational Team Concert, Aldon, or Turnover on top of that. All of it worked. None of it was built for two developers editing the same program at once.

Git gives you branches instead. Multiple developers, multiple releases, all in flight at once, isolated until they’re ready to merge.

For IBM i, that’s a real shift. Most shops have run one-at-a-time change management for decades. It worked. It just doesn’t scale to parallel development, and it doesn’t speak the same language as the rest of the enterprise.

Then there’s AI.

AI tools work best inside a Git repository. A branch is a sandbox, a place for AI to make changes without touching production. A pull request is the human checkpoint before anything moves forward.

Give an AI tool free rein in a live library, and it’s working across everything at once. Give it a branch, and it’s contained, reviewed, safe.

Git is how you get that isolation. Not a policy. A structure.

That’s why it matters on IBM i. Not because it’s trendy. Because it’s the container AI needs to be useful without being reckless.

Trends

🤝 AI strips busywork. What fills the void?

Zoom’s founder is betting AI will free people from coordination drag, sparking more human interaction, not less.

But where does that happen?

Some return to office mandates have devs showing up to empty offices, while manufacturing, warehousing, and other hands-on industries still need boots on the ground.

The real question isn’t remote versus office, it’s human versus chatbot, and how we choose to connect once AI clears the busywork away.

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!