Beyond the Hype. GenAI's Impact on Development

Generative AI is going to change software development in 2026. That part is obvious. The useful question is what changes, what gets cheaper, what gets faster, and what gets worse. This webinar is a practical walkthrough of what GenAI is commodifying, what it speeds up, what it makes possible, and where it harms teams. The goal is not hype. The goal is to help you make better decisions about how you use these tools without creating a false sense of velocity and a bigger maintenance burden.

1) What GenAI commodifies

GenAI is rapidly commodifying work that used to feel like real development, especially work that is repetitive or has a clear pattern. That does not mean the output is reliable, secure, or fit for purpose. It means the first draft is cheap.

We will cover how GenAI is commodifying:

• Creating CRUD code
• First drafts of documentation, both technical and user-facing
• UI mockups and basic design work.
• Basic accessibility improvements, because many tools generate UIs that meet accessibility standards by default.
• Initial impact analysis for new features
• Finding edge cases faster, while still missing many along the way.
• TDD/BDD is becoming easier because the tool can generate the test cases and the underlying code.

A less experienced team can produce an application faster with a little help from AI. The missing ingredient is taste. Taste is what tells you whether you built a usable user experience, handled edge cases, addressed security, kept the system testable, and avoided technical debt. The bill for all of this will come due later.

GenAI also creates a large volume of output, much of which is not relevant to the task at hand. That shifts effort from producing artifacts to sifting through artifacts. Edge case analysis is a good example. The tool will find many edge cases, but the volume creates duplicates and noise. You spend time eliminating duplicates, and you risk missing cases that were not found.

2) What GenAI speeds up

GenAI can speed up strategic thinking when you provide enough context and use it as a sparring partner. The speed-up is not magic. It depends on your ability to frame the problem, provide constraints, and evaluate the quality of what comes back.

3) What GenAI makes possible

GenAI makes some things possible that were previously expensive or slow. The cost of trying things drops. The cost of cleaning up mistakes can rise.

We will cover how GenAI makes possible:

• Smaller teams - risk fewer perspectives to catch problems, loss of redundancy, and loss of mentorship.
• Running more experiments in the sense of Lean Startup and Lean UX.
• Testing more variations, risks: if you do not understand statistical power, you can create a lot of variation testing that is irrelevant.
• Technical spikes becoming cheap, to the point where you can test across different programming languages.
• Product Owner prototyping without needing outside help.

As with commodification and speed-up, the new possibilities come with changing risks.

4) Where GenAI harms teams

GenAI does harm, and the harm is predictable. The pattern is simple. Output gets cheaper. Judgment gets more expensive.

We will cover the major harm patterns:

• Quality risk, because LLMs are trained on the average of the code on the internet and the average quality is not great.
• Vibe coded security and related risks
• As the volume of code increases, all risks increase.
• The QA bottleneck becomes worse.
• Code review becomes a new bottleneck.
• Skills atrophy, especially for junior developers who do not learn the skills they need.
• As skills atrophy, false confidence in the output increases.
• Taste for good UX design and code quality comes from experience and practice, and that practice is getting lost.
• Collaboration is reduced as we spend more time talking to the LLM and less time talking to each other.
• More time is spent using critical thinking and less time doing rote work, but there is a limit to how much critical thinking we can sustain in a day.
• Sustainable pace becomes harder because it is perceived that LLMs help us go faster, while the cost of working with the LLM is higher. The cognitive work is heavier.
• Increased cognitive load is the key issue, because the work shifts toward constant evaluation, risk detection, and judgment.
• Increased stress in teams, because we spend more time looking for risks and mistakes and less time on the creative side of the work.

5) What the software development community misunderstands

There are several misunderstandings that create bad decisions and bad expectations.

We will address:

• Why the oft cited benchmarks that are cited as evidence of AI effectiveness are meaningless for real work.
• Why generating User Stories and Product Requirements Documents is a waste of time
• Why software development is not just writing code. (And never has been)
• Why debugging code you did not write is hard. You did not pay the price when it was written, so you pay a higher price when there is a production issue.

6) Surviving and thriving in the age of AI

The risks are not abstract. They show up as maintenance burden, rework, and pressure.

We will cover:

• It is cheaper to generate artifacts, but the more artifacts you have the greater the maintenance burden.
• The false velocity trap. More user stories faster, doesn’t mean more value. Increased velocity feels good and can mean increased technical debt, which means more rework, which increases complexity, which increases technical debt.
• Speed theatre. Typing was never the bottleneck. GenAI is making things that were already fast even faster.

Why listen to me

I am not coming at this as a cheerleader or a doomsayer. I am coming at it as an explorer. I don’t want to sell you tokens, nor do I need to justify hundreds of billions of dollars in GPU investments.

I have survived every tech hype cycle that has happened, since before Gartner started tracking them (4GLs, Apple Newton, CORBA, Clippy, Flash, Java Applets, 4GL, Segway, No Code, the Metaverse, …) . In many of those cases, I even had to survive with the technology. We were promised the moon, and we spent years in some cases trying to make it work. (Hello CORBA, Newton and Java Applets.) GenAI is the loudest hype cycle, I’ve ever witnessed. Trillions of dollars in investments, yes. Trillions of dollars of value? No evidence of that.

My job in this webinar is to strip away the hype and focus on what delivers value. That includes naming the risks clearly, because pretending the risks do not exist does not make them go away.

I do not want GenAI to exist, for many reasons (theft of intellectual property, destruction of value, irresponsible rollouts, energy use). Yet it is here. So I am choosing to deal with reality instead of wishing it away.

In practice, I am using GenAI as a sparring partner, to challenge me to improve my writing, to automate small tasks, and to help me build several applications. That means I am not speculating from the sidelines. I am learning where it helps, where it hurts, and what it costs.

Who this is for

This webinar is for developers, tech leads, Scrum Masters, Product Owners, and managers who want a grounded view of AI in software delivery and who are responsible for outcomes, not hype.