Today I saw this at a YouTube video of young team working by slack with cursor and MCPs… for changes and updates of their site without even the engineer being there (good practice or not, they are moving and improving). While in my company we are stuck because of whatever bureaucracy and product definition. Don’t take me wrong I see the value of certain level of process but is a shift, I see how many are unlocking their creativity while others continue denying the transformation.
great one! Although I don’t think that the culture stems only from being AI native, it’s an amplifier but not a direct root cause. In order to achieve this kind of dynamics you need to create a team of empowered hackers with a strong bias towards action and total autonomy. The companies like that existed before but rarely achieved such hyper growth while still at the hacker stage - this is where ai walks in as the key catalyst IMO.
The holy grail is to preserve it at scale - would love to see more reads on that every few company milestones :)
Fascinating perspective. From what I’m seeing in Peru (especially in financial services), corporations are still far behind on enabling innovation through AI. Bureaucracy and rigid processes make it incredibly hard to move fast or adopt an AI-native mindset. It makes me wonder how big the gap will be when these AI-native employees really start coming into the workforce. Thanks for sparking this reflection — will definitely be sharing!
couldn't agree more! this was a really fun read. and SPOT on. it's night and day difference between an ai-native company (employees) and a saas-era company (employees move MUCH slower).
Totally aligned on this. We're living the AI-native reality right now - our PM/UX team defaults to Lovable for prototyping, and it's completely changed our velocity. What you describe about ownership and autonomy really hits home. Our team builds instead of writing docs nobody reads or sitting through endless alignment meetings.
Having said that, do you have any runbooks or best practices for your team to follow when taking Lovable prototypes to production with minimal friction?
The things that you have described works flawlessly in a small organization. But when you need to scale and do not surprise your customers with releases here and there and you have to synchronize your GTM, AI-native will still need to be managed.
Hi, nice article! I found it via the "Mary Decodes Design" Substack.
I don’t fully agree with the thesis: “The hidden superpower: cheap failures. When you can move this fast, the cost of failure plummets.”
Because it also means:
* Lower quality: covering edge cases usually takes much more time.
* Rough user flows that can break at some point or leave users stuck in unclear situations.
* Less sense of professional experience: an individual contributor using AI may generate outcomes based on their personal perspective, highlighting what they consider important and skipping what they don’t, but this personal view isn’t always accurate. In contrast, a team’s collective experience tends to balance features more effectively.
* Learning loops are important, but equally important is sharing those learnings and knowledge which isn’t always the case for individual AI contributors.
* “Move fast and break things” works well for prototyping, but it requires additional effort to identify what works well and what needs improvement or removal adding extra complexity.
1. How do product decisions get made? While making small tweaks to the product might not require a lot of oversight / scrutiny, if you were to make a brand new product within Lovable (as an example, if you were to launch a new PowerPoint-like product under the Lovable umbrella), what is the process like at Lovable?
2. The traditional 3-way specialization model (product, design, engineer) has another hidden benefit of a health tension between the functions. E.g., Design would always push for customer delight and ease of use, Engineering will always push for elegant and maintainable technical solution, while Product will always push for an business objective that aligns to the broader product strategy. This tension keeps the shift from leaning too hard to any direction. Do you see that in the new model you observed at Lovable?
Great perspective! Living this from the inside of a large company trying to make employees AI native, I can see the friction between that and the org charts. It does feel like now there are too many cooks in the kitchen when we can move that fast and take on more.
What's your take on customer centricity in a world where we move fast? The only problem I have with the unlock of speed is that it can prompt some people to build for the sake of ideas, without a solid customer problem or vision in mind. While failing fast helps with this by invalidating/validating initial product direction, my experience right now is that people throw things at the wall without deeper thinking, just because an idea looks cool and innovative.
Id love your take on how to balance speed, AI native, and deep customer empathy!
Wow, this hits too close to real, especially the "typical tech company" section and that cartoon. AI-native isn’t just a skillset, it’s a mindset. I’ve worked at a traditional org trying to bolt AI onto old workflows like duct tape on a rocket. Doesn’t fly. The unlock I’ve found at Braintrust is to keep it small, and think even smaller, people-wise. Hire the "I don't know but I'll figure it out" mindset folks willing to experiment, learn, fail, share their learnings, and then try again.
48 here and AI first - not sure about build fast :) but I am breaking stuff...is that still ok?
Today I saw this at a YouTube video of young team working by slack with cursor and MCPs… for changes and updates of their site without even the engineer being there (good practice or not, they are moving and improving). While in my company we are stuck because of whatever bureaucracy and product definition. Don’t take me wrong I see the value of certain level of process but is a shift, I see how many are unlocking their creativity while others continue denying the transformation.
i'm not saying process is not needed. but process needs to be built up from the ground up, not forced from top down onto this new way of working...
Yes 100%
great one! Although I don’t think that the culture stems only from being AI native, it’s an amplifier but not a direct root cause. In order to achieve this kind of dynamics you need to create a team of empowered hackers with a strong bias towards action and total autonomy. The companies like that existed before but rarely achieved such hyper growth while still at the hacker stage - this is where ai walks in as the key catalyst IMO.
The holy grail is to preserve it at scale - would love to see more reads on that every few company milestones :)
Fascinating perspective. From what I’m seeing in Peru (especially in financial services), corporations are still far behind on enabling innovation through AI. Bureaucracy and rigid processes make it incredibly hard to move fast or adopt an AI-native mindset. It makes me wonder how big the gap will be when these AI-native employees really start coming into the workforce. Thanks for sparking this reflection — will definitely be sharing!
couldn't agree more! this was a really fun read. and SPOT on. it's night and day difference between an ai-native company (employees) and a saas-era company (employees move MUCH slower).
ps - I wrote some similar thoughts recently, where I talk about, what I call "AI-Native Workers": https://www.thesignal.club/p/why-building-ai-native-is-the-biggest
Totally aligned on this. We're living the AI-native reality right now - our PM/UX team defaults to Lovable for prototyping, and it's completely changed our velocity. What you describe about ownership and autonomy really hits home. Our team builds instead of writing docs nobody reads or sitting through endless alignment meetings.
Having said that, do you have any runbooks or best practices for your team to follow when taking Lovable prototypes to production with minimal friction?
The future is about to be evenly distributed 🚀🚀🚀🚀
You speak truth and pull no punches. Looking forward to moving at AI Native speed in the Enterprise.
I’m living through the same thing and can definitely relate.
This was terrific.
The things that you have described works flawlessly in a small organization. But when you need to scale and do not surprise your customers with releases here and there and you have to synchronize your GTM, AI-native will still need to be managed.
Was just thinking that!
If it sounds too good to be true …
AI native development will run like cheetah’s compared to standard corporate bureaucracies.
But eventually you’ll have to heard cats, and I see l a lot of risk there.
I hope I’m wrong because AI-in-everything is the new normal.
Hi, nice article! I found it via the "Mary Decodes Design" Substack.
I don’t fully agree with the thesis: “The hidden superpower: cheap failures. When you can move this fast, the cost of failure plummets.”
Because it also means:
* Lower quality: covering edge cases usually takes much more time.
* Rough user flows that can break at some point or leave users stuck in unclear situations.
* Less sense of professional experience: an individual contributor using AI may generate outcomes based on their personal perspective, highlighting what they consider important and skipping what they don’t, but this personal view isn’t always accurate. In contrast, a team’s collective experience tends to balance features more effectively.
* Learning loops are important, but equally important is sharing those learnings and knowledge which isn’t always the case for individual AI contributors.
* “Move fast and break things” works well for prototyping, but it requires additional effort to identify what works well and what needs improvement or removal adding extra complexity.
Really fascinating read! Thank you! I'm curious tho on 2 things:
1. How do product decisions get made? While making small tweaks to the product might not require a lot of oversight / scrutiny, if you were to make a brand new product within Lovable (as an example, if you were to launch a new PowerPoint-like product under the Lovable umbrella), what is the process like at Lovable?
2. The traditional 3-way specialization model (product, design, engineer) has another hidden benefit of a health tension between the functions. E.g., Design would always push for customer delight and ease of use, Engineering will always push for elegant and maintainable technical solution, while Product will always push for an business objective that aligns to the broader product strategy. This tension keeps the shift from leaning too hard to any direction. Do you see that in the new model you observed at Lovable?
Design also pushes for cohesion, system-level thinking that aligns with user’s mental models.
Narrowing beyond that introduces biases that tip the product towards failure.
Met an old Apple mate over the weekend. I showed her my Lovable projects — AI-assisted, vibe-coded, pure fun.
She hadn’t heard of Lovable. Wasn’t using any AI at all.
“Not allowed,” she said. “And I’m too swamped at work to explore.”
No surprise there — old tech is still clocking in while the future’s building side projects on a Friday night.
I’ve used my leave to play, test, build, break, and learn. And I’ve loved it.
Going back soon… to the land of “that tool’s not approved” and “that vendor’s not onboarded.” Honestly? Dreading it a bit.
Great perspective! Living this from the inside of a large company trying to make employees AI native, I can see the friction between that and the org charts. It does feel like now there are too many cooks in the kitchen when we can move that fast and take on more.
What's your take on customer centricity in a world where we move fast? The only problem I have with the unlock of speed is that it can prompt some people to build for the sake of ideas, without a solid customer problem or vision in mind. While failing fast helps with this by invalidating/validating initial product direction, my experience right now is that people throw things at the wall without deeper thinking, just because an idea looks cool and innovative.
Id love your take on how to balance speed, AI native, and deep customer empathy!
Wow, this hits too close to real, especially the "typical tech company" section and that cartoon. AI-native isn’t just a skillset, it’s a mindset. I’ve worked at a traditional org trying to bolt AI onto old workflows like duct tape on a rocket. Doesn’t fly. The unlock I’ve found at Braintrust is to keep it small, and think even smaller, people-wise. Hire the "I don't know but I'll figure it out" mindset folks willing to experiment, learn, fail, share their learnings, and then try again.