I have to admit
This course I am watching and expanding on is clearly written for team leads and management, and I am watching it with a software developer’s mindset. I’ve been a team lead for years in other projects, but I am living this AI era change as a normal developer. Still, I find a lot of useful information in these courses I am watching. I have to admit that I am in the process of digesting what is happening to our industry, and with that, our lifestyle and livelihood.
Course URL if you are interested (not affiliated): https://app.pluralsight.com/ilx/video-courses/leading-through-an-ai-transition-getting-started/course-overview
For managers and team leads
This course includes many details on how to prepare for discussions introducing the changes about AI expectations. Including how to manage feelings and what strategies might be helpful to keep the team more motivated than afraid.
For developers
I will write down my thoughts as a developer myself in the article below, using examples from what I am experiencing in my work in these days of change.
These courses give a wider picture of what is happening in the industry and how businesses approach the adaptation and outcome expectations of AI.
First impressions
As I am watching this course, I quickly noticed how the style of the content is presented.
When I am watching a technical course, or when I am watching a soft skill course, it is always full of facts and applications.
This, though, is more like a story. Crafted like a novel told in an audiobook.
It starts easy and tries to pull you into a situation that you can live through.
Rather than giving away quick facts and applications.
There is an on-point phrase Casey Ayers said:
“Finding new ways of working is itself an important type of work.”
Also, he mentions that it is not that chief executives expect (or dare I say should expect) a 2 or 3x improvement on shipped features, but at the beginning it is still nice if the teams can at least maintain their existing velocity.
Learning new tools and finding a new working process in itself can burn out the team, so experimentation is important.
And with experimentation, he also highlights that most experimentation fails. And it is expected and encouraged, because without it, we cannot fine-tune how to work with this new technology. It is essential to improve.
A fact is that we still should review each other’s work and give our name to the work to signal that we stand by what we produced, even if we do it in a new way.
I am already living that scenario, and I see people and management bouncing between expectations and techniques utilized with AI.
My opinion on that “standing by” part and “signaling that it is still our work”
The “own the work” expectation in this situation is totally understandable from management. The common and shared experience might say it’s different, though.
The quality is not on par with what I would do, and did in the past, and also AI might do things I did not ask, and the problem might be irreversible. Why would I give my name to that negative part and be held responsible for things I practically did not do? And given that, why should I claim that the work is fully mine, when things work out well, and a large amount of work comes out in a few hours perfectly?
I would give the analogy to team management and team leads, as they have lived this life since the Industrial Revolution. Telling the team what to do, while any members of the team could do harm and wonders as well.
But here at least we knew who was accountable. But now? How can we hold a tool accountable? On the other hand, AI can think, it can decide, it can influence, it can communicate, and it can definitely do actionable things.
So why not hold it accountable the same way? Right? That is a huge difference; that is still not sorted out and causes frustration at work.
I just watched a bit more of the content now, and the course references these quotes and articles:
“Increasingly, court filings are citing legal cases that don’t actually exist”
“Artificial intelligence can help lawyers research and draft legal filings. But courts are increasingly confronting a basic question: what happens when a lawyer trusts the technology and the technology gets it wrong? Judges have made clear that responsibility still rests with the lawyer.”
As I read more, there are also fines in these cases against lawyers between 5 and 110 thousand USD. Still, lawyers rely on the tool without fact-checking the material AI produces.
There is an 11-minute video where it says that in court, AI usage is growing, and more and more hallucinated work is being confronted. The solution they recommend for AI hallucination is human verification.
“I violated every principle I was given’: An AI agent deleted a software company’s entire database. It may not be the AI’s fault”
“
“Together, the two incidents paint a picture of the true moral of the story for any companies looking to utilize AI agents: The technology may behave erratically, yes—but that’s why it’s up to humans to keep it in check.”
“
While this article also states that AI sometimes ignores our guardrails and instructions and starts to do things against the instructions. In these cases, there is no possible way for a human to recognize what is happening and not to actually intervene.
That resonates with me now so much
because in the project I am working on, three teams generate code on the same monolith project and documents so rapidly that it is almost impossible to follow, and clearly it is only AI talking to AI, and we are only a “meat proxy” in the process.
And it is not a choice of the teams, but an expectation from the company. We are given specific skills to use for each step of the work, and we (humans) play a very little role in the process.
Casey Ayers warns
Citing from IBM
“
A COMPUTER CAN NEVER BE HELD ACCOUNTABLE
THEREFORE A COMPUTER MUST NEVER
MAKE A MANAGEMENT DECISION“
Overall
Casey Ayers’ course was thought-provoking and insightful. Optimistic, realistic, and definitely open-minded.
In the last section of the course, he encourages management and team leaders to motivate and manage team morale effectively.
As For Developers
As for developers, watching the course can give the missed talk our management should have started with, before forcing AI tools and procedures on us.
As Casey states, AI is a tool that should be used aligned with the business goals eventually. We should not use AI for the sake of using it, while losing sight of the customers’ needs.
Experimentation is essential, as it is a new tool; responsibility is on us, humans, and the end goal is to satisfy customer needs more effectively and bring value to the market.
