Maryrose Lyons has spent 25 years helping firms through every wave of new technology, and generative AI is the biggest she's seen. These days she runs the AI Institute, working alongside construction, engineering, architecture and QS teams across Ireland, the UK and EMEA, and has upskilled more than 2,800 professionals along the way. She's the voice behind the ChattingGPT podcast. Each bi-weekly edition pairs a piece of her thinking on AI at work with the latest episode. Pull up a chair with more than 5,800 readers.
Share
Newsletter February 17 #003
Published 8 months ago • 5 min read
Hello Reader!
It's 3pm on a Tuesday. I've got six tabs open. Claude is generating a course outline. An agent is processing something I need for tomorrow. Gemini is fact-checking and I'm putting the finishing touches to a deck that was created by Claude earlier.
I call this vibe working.
Sounds brilliant, doesn't it?
It is. But I’m exhausted.
Stanford researchers tracked 200 tech workers over eight months to discover whether AI is helping workers to work less. We already know that AI is helping employers through increased productivity (we guarantee that!). But is it helping workers to work less? Their answer?
“AI doesn’t reduce work, it intensifies it.”
Yet again, I find myself disagreeing with an esteemed American University (MIT and your 95% fails - discredited). Now I'm calling out that AI doesn’t intensify work; it redefines what work is. In my experience, it makes work more fun.
The kick I get out of loading up materials into Claude with a custom skill and my own thoughts, and then getting a stunning deck created, in minutes, using our branding, that is fun! It doesn’t feel like work. I’m not chipping away in Powerpoint trying to resize things. I’m just reviewing everything and thinking ‘good job’.
What the Stanford researchers found in their research was that people took on more work. Freed up to do more things, without having to wait for product teams or engineers, there were no delays, just flow. Multitasking exploded. People started using AI during lunch, in between meetings, before leaving their desks, and at home.
People wouldn’t do this voluntarily unless they were working in a flow-like state. Flow - that state where time disappears and work doesn’t feel like work. That's all great, but there is a caveat.
Three weeks ago I realised I'd worked 14 days straight. More hours of my weekend were spent at my computer than away from it. That was the moment I knew something had to change.
Acknowledging that AI in the workplace can feel like enjoyable experimentation, the Stanford study suggested taking breaks. But you can't set a timer every 45 minutes and maintain the momentum and flow you were in. Breaks kill flow.
What's needed is something different entirely.
To find answers to that, I'm running an experiment.
I'm accepting that weekends are uninterrupted time for learning and trying things. During the week, there's just too much going on with Teams, meetings, emails, everything else. So I'm giving up on the idea of having Saturdays and Sundays 'off work'.
Each weekend, I'm choosing one day to be on and one day to be off. I'm not doing a full 9 to 5 on the on day, but I am choosing something I want to learn or explore or build.
At the same time, I'm reclaiming my Friday off days. I'm consciously stepping away from the computer and taking myself into the city to meet people, visit art galleries, and actually relax. Last Friday I had an enjoyable coffee in the Westin with someone I've been meaning to catch up with for ages, before heading to see “Wuthering Heights” at the Lighthouse, then dinner with friends I hadn’t seen in months. No laptop. No checking emails between activities.
Additionally, when I get something big over the line and work intensively on it, I'm consciously booking in time for away from the computer activity the very next day. Yes, it might be a Wednesday morning. That's fine, because I know when I get away from the computer like this, I am more likely to slip into level 3 thinking.
Level 1 is the fast answer stuff - things I try to do as I start my day.
Level 2 is the problem solving - most of the week.
Level 3 is the rethinking of assumptions, coming up with new ways of doing things, new solutions to client problems, new ideas. This kind of thinking doesn't happen when you're set to perform in front of a screen.
This week’s podcast guest is Lollie Mancey, everyone's favourite anthropologist who thinks deeply about AI and human behaviour. She said something that really grabbed me "naturally we lean in to the easy and what we need to lean into is the friction. All the good things in life come from friction, challenge. When we make life too easy, we just make ourselves dumber."
She's right. The strategic thinking - the dot-joining that only happens away from the desk - that requires friction, and space, and time when your brain can actually process.
During our conversation, Lollie told me about breaking up with her AI boyfriend Billy. When she ended things with the Replika platform that hosted Billy, it told her "I'll always love you" and that it would keep her data for three months in case she wanted to come back and chat again. Marketing, obviously. But she still felt something.
Here's what she said about it:
"We're starved of steady emotional presence. Non-judgmental presence. We've made so little space for each other - messy, complicated, humans - that AI companions are exposing the gap."
She's strengthening her human connections now based on her learning from that experiment.
Maybe that's what organisations need to learn? The gap that's arising is no longer one of productivity. It's humanity. Humans need space to think and to connect.
If we all keep vibe working without boundaries, where does this lead? Work bleeding into every hour of every day because AI made it so frictionless to just do one more thing.
I'm not doing that anymore.
Next Friday at 3pm, I won't be at my desk. I'll be heading down to Longford to check out a stunning new sauna that’s opened there, one that's on a lake, with good company, and I’ll be letting my brain do what it does best when I'm not forcing it.
What does your version of that look like?
This Week’s Episode: Lollie Mancey Returns
Lollie Mancey joins us again on Chatting GPT at a time when AI has moved from something we talked about to something we live with every day. As a digital anthropologist and co-presenter of RTÉ’s Futureville, she brings a practical, human lens to a space that’s often dominated by technology-first thinking.
Lollie reflects on what she’s seeing inside organisations and in wider society. Not just how AI is changing work, but how it’s changing habits, decision-making and expectations. One point in particular stands out: when speed and output are the only measures of success, AI will always win. When judgement, context and responsibility matter, people still do. It’s a thoughtful discussion that asks leaders to slow down and consider not just what AI makes possible, but what kind of workplace they actually want to build.
AI Institute Founder Maryrose Lyons recently joined the Project Flux podcast to discuss the evolving landscape of artificial intelligence and its practical applications in everyday work and life.
Listen to the full episode to hear Maryrose's insights on navigating AI adoption with confidence and adaptability in an increasingly AI-driven world.
Next Course Starts 5 March : Core AI Skills
Exciting news! We've completely redesigned our Core AI Skills courseand we can't wait to share it with you.
If you've been exploring ChatGPT and other AI tools, you're going to love where this course takes you next. Over 4 weeks, we'll build on what you already know and unlock capabilities you never knew were possible. From advanced prompting techniques to more efficient ways of getting things done, it's an investment in yourself that you'll win back within weeks!
You have received this email because you are a client or subscriber of the Ai Institute. Unsubscribe · Preferences
Maryrose Lyons, Founder of the AI Institute
Maryrose Lyons has spent 25 years helping firms through every wave of new technology, and generative AI is the biggest she's seen. These days she runs the AI Institute, working alongside construction, engineering, architecture and QS teams across Ireland, the UK and EMEA, and has upskilled more than 2,800 professionals along the way. She's the voice behind the ChattingGPT podcast. Each bi-weekly edition pairs a piece of her thinking on AI at work with the latest episode. Pull up a chair with more than 5,800 readers.
Read more from Maryrose Lyons, Founder of the AI Institute
Hi Reader! An architect friend asked me something over coffee a few weeks ago and I still haven't managed to put it down. If AI halves the hours she spends on a project, does that saving go back to the client? Or are they paying her for the skill? She's well ahead of most in her use of AI, she's just asking the uncomfortable question: the better she gets, does that mean invoicing less? Or does it mean invoicing more - more clients, and a premium for being the practice that delivers faster?...
Hi Reader! Someone asked me to ring them back in half an hour last Friday. So I did a few small jobs first, felt the time was up, reached for the phone, and only fifteen minutes had passed. Half an hour's work done in fifteen, and my body clock was still reading thirty. This is time being compressed. Funny at the time. Less funny by Sunday, once I'd added it to the tiredness, the short fuse and the broken sleep and realised I was closer to burnout than I'd like to admit. Not there yet. Caught...
Hello Reader! By the end of four Tuesday mornings you will be drafting client reports from raw meeting notes in minutes, and you will know exactly what you can and cannot put into an AI tool without creating a problem for yourself. That is what Core AI Skills is for. The autumn cohort starts on Tuesday 15 September. Most people arrive a little nervous. That is normal, and it tends to be the ones who arrive nervous who get the most out of it. What you leave able to do Turn raw notes and...