What does faster code cost when the hard part is deciding whether it belongs in the system?
An AI assistant can draft a function in seconds. It cannot, by itself, know why your team rejected a similar design three years ago, which customer depends on an undocumented exception, or how a convincing answer will behave under your production constraints. The speed is real. So is the cost of discovering the missing context late.
That is the decision behind AI4DEV.
On Thursday, October 15, Howard Pearlmutter will lead a live, online, two-hour seminar for developers and the people who make decisions with them. We will look across the software-development lifecycle: where AI can extend human ability, where it can create a false sense of completion, and how experienced judgment changes the value of the tools.
THE QUESTION IS BIGGER THAN “CAN IT WRITE CODE?”
Code generation is only one part of development. Teams also have to understand existing systems, explore requirements, shape architecture, investigate defects, design interfaces, write and examine tests, document decisions, integrate changes, and decide what is ready to release.
AI can assist with many of those activities. But a quick draft can move the work rather than finish it. A plausible test may test the wrong behavior. A tidy refactor may remove the odd branch that protects a real customer. An attractive prototype may leave security, operations, or maintainability as somebody else's surprise.
The practical question is not whether to be enthusiastic or skeptical. It is where assistance earns its place, what evidence you need before you trust the result, and which work still requires a person who understands the system. Those answers change with the codebase, the team, and the stakes.
WHY SPEND TWO HOURS ON THIS?
Because a tool subscription is easy to buy; a shared way to judge the work is harder to build. If developers, managers, and product leaders use different meanings for “done,” faster production can multiply downstream disagreement. If they can distinguish a useful first pass from a dependable result, the same tools become more valuable.
AI4DEV is for the developer who has inherited confident mistakes; the engineering leader weighing speed against reliability; the product leader who needs to know what a prototype proves; and the executive who wants more than a vendor's before-and-after demo. One admission is $250. Consider bringing a developer and a decision maker from the same organization so they can continue the conversation in a common language.
WHY HOWARD?
Howard Pearlmutter has spent decades doing the connecting work between new technology, working software, and the people accountable for decisions. He has built commercial systems, taught advanced technical audiences, and briefed leaders when the technologies that later defined an era were still emerging.
Valtech trusted him to teach advanced Java architecture and explain technical choices to audiences that included professionals from JPMorgan, Goldman Sachs, EDS, Barclays, and Orange. Its leaders praised both his technical expertise and his ability to teach. In another era, GM, Ford, and other major automakers needed sophisticated factory designs translated into something executives could see and examine. Howard built the web-enabled product that turned those designs into animated, interactive management reports.
His AI experience did not begin with chatbots. He was applying extensible-language techniques to shipped software decades before today's models. More recently, a custom AI-for-development seminar drew 47 people across development and management. Client president John Schur reported that Howard engaged skeptics, addressed their concerns, and described a careful path forward. That is the sort of conversation AI4DEV is built to support: technically literate, open to opportunity, and alert to what the work actually demands.
THE VALUE OF THE LONG VIEW
This is not a tour of whichever AI product has the loudest launch this week. Interfaces, model capabilities, and vendors will change. Your need to understand context, verify behavior, make tradeoffs, and decide who is responsible will remain. Howard's unusual range—from early AI and software construction to executive communication—makes this a chance to think across those boundaries in one sitting.
If AI is already entering your team's work, the decisions are already being made. Some are explicit; others are hidden in defaults, habits, and unexamined output. Give the subject two focused hours before the next shortcut becomes your architecture.
AI4DEV with Howard Pearlmutter
Thursday, October 15, 2026
2:00–4:00 p.m. Mountain Time
Live online • $250 per admission