Completely agree with Good Culture is the Biggest Productivity Hack!
In fact, in my last post about What is an AI-Native team, I mentioned how keeping the same Amazon culture and best practices made people become faster maintaining high quality. My harness multiplied product teams at Amazon.
However, it also depends on the teams. If the director only cares about shipping and not quality, and stop caring about foundations is when surprises and numbers don’t match that promise.
The one glaring thing I find missing in this article: what's in it for your engineers if they *do* adopt AI and realize the exponential output it promises.
If you do not explicitly provide a positive incentive for people to move in that direction, the default is going to be a feeling that everyone is suddenly expected to output 4-10x more work in the same working hours or else they are suddenly going to find themselves without a job tomorrow. Worse yet, most people feel that even if they do meet this expectation, they are going to be laid off, or at best get a measly 6% cost of living raise.
That is the elephant in the room that leaders have to deal with in today's AI crazed business culture.
Thanks, and glad the article resonated! Right, great point, and that's crucial these days. If you look at companies like Nvidia, they do zero layoffs, and everyone there has a good level of psychological safety. While in a company like Meta, it's totally different, there is high competition between people which automatically influences wrong behavior.
It all starts there, a track record of doing good things for your people, emphasizing how important they are, and that they are not just a "resource", but actually extremely important for the success of the company. That's what's crucial these days imo.
Strong argument! I’d add one nuance to the bottom-up adoption point:
Direction should come from the top, while discovery should come from the bottom.
Leadership should define the outcomes, constraints, and principles. Engineers closest to the work should determine where AI is genuinely useful, share what they learn, and shape evolving practices.
Top-down mandates can create compliance theatre, while purely bottom-up experimentation can create fragmentation. Good culture connects the two.
Great to see you writing about culture! I'd add that even a good culture can crack if we're not clear on what relationship we want to have with AI as a company, because at its core this is about team identity, not tools.
The real problem isn't so much "do we use AI or not" but rather: what do we actually want to do with it as an organization? What's our culture around this? Without a clear, shared answer to that, any culture, no matter how strong its foundation, is exposed to panic or comparison slowly wearing it down...
I would push back gently on culture as the variable. Culture is real but it is hard to falsify, which makes it a comfortable explanation for results we cannot otherwise account for. The measurable version of your argument is capacity allocation. Faros AI, across 22,000 developers and 4,000 teams, found throughput per developer up 33.7% and median code review time up 441.5%. LinearB, across 8.1 million PRs, found AI-assisted work merging within 30 days 32.7% of the time against 84.5% unassisted. Teams that get value from AI are the ones that added review and verification capacity alongside it. That often looks like good culture from the outside, but the causal thing is a staffing decision, and it can be made deliberately by a team whose culture is merely fine.
I totally agree with this article where culture is the key. Great article btw
Man! Completely agree: Good Culture is the biggest productivity.
Putting AI in capable hands in the most important part. Capable hands and capable team is where the magic really happens.
Completely agree with Good Culture is the Biggest Productivity Hack!
In fact, in my last post about What is an AI-Native team, I mentioned how keeping the same Amazon culture and best practices made people become faster maintaining high quality. My harness multiplied product teams at Amazon.
However, it also depends on the teams. If the director only cares about shipping and not quality, and stop caring about foundations is when surprises and numbers don’t match that promise.
Love the message that culture comes first!
The one glaring thing I find missing in this article: what's in it for your engineers if they *do* adopt AI and realize the exponential output it promises.
If you do not explicitly provide a positive incentive for people to move in that direction, the default is going to be a feeling that everyone is suddenly expected to output 4-10x more work in the same working hours or else they are suddenly going to find themselves without a job tomorrow. Worse yet, most people feel that even if they do meet this expectation, they are going to be laid off, or at best get a measly 6% cost of living raise.
That is the elephant in the room that leaders have to deal with in today's AI crazed business culture.
Thanks, and glad the article resonated! Right, great point, and that's crucial these days. If you look at companies like Nvidia, they do zero layoffs, and everyone there has a good level of psychological safety. While in a company like Meta, it's totally different, there is high competition between people which automatically influences wrong behavior.
It all starts there, a track record of doing good things for your people, emphasizing how important they are, and that they are not just a "resource", but actually extremely important for the success of the company. That's what's crucial these days imo.
Totally agree. The people make the company. AI will just magnify the pre-existing problems within organizations if left unchecked.
Right, there's a reason why I like to say that engineering is more about people than tech. Good culture comes first, everything else just works after.
Strong argument! I’d add one nuance to the bottom-up adoption point:
Direction should come from the top, while discovery should come from the bottom.
Leadership should define the outcomes, constraints, and principles. Engineers closest to the work should determine where AI is genuinely useful, share what they learn, and shape evolving practices.
Top-down mandates can create compliance theatre, while purely bottom-up experimentation can create fragmentation. Good culture connects the two.
Insightful read. Thanks
Just wondering what's your 'incentive' here :-P
Great to see you writing about culture! I'd add that even a good culture can crack if we're not clear on what relationship we want to have with AI as a company, because at its core this is about team identity, not tools.
The real problem isn't so much "do we use AI or not" but rather: what do we actually want to do with it as an organization? What's our culture around this? Without a clear, shared answer to that, any culture, no matter how strong its foundation, is exposed to panic or comparison slowly wearing it down...
I would push back gently on culture as the variable. Culture is real but it is hard to falsify, which makes it a comfortable explanation for results we cannot otherwise account for. The measurable version of your argument is capacity allocation. Faros AI, across 22,000 developers and 4,000 teams, found throughput per developer up 33.7% and median code review time up 441.5%. LinearB, across 8.1 million PRs, found AI-assisted work merging within 30 days 32.7% of the time against 84.5% unassisted. Teams that get value from AI are the ones that added review and verification capacity alongside it. That often looks like good culture from the outside, but the causal thing is a staffing decision, and it can be made deliberately by a team whose culture is merely fine.