Pawel Brodzinski
@pawelbrodzinski.bsky.social
Leader of an org where anyone can make any decision (Lunar Logic).
Doing anything that no one else wants to do.
A mouthful on product development, startups, org design, lean/agile, IT in general. Also AI. AI too.
Doing anything that no one else wants to do.
A mouthful on product development, startups, org design, lean/agile, IT in general. Also AI. AI too.
If your cycle time sucks, automating coding won't do anything to your delivery pipeline.
You've got 99 problems, but the pace of coding ain't one.
by @paulisthrivving.xyz
medium.com/thrivve-part...
You've got 99 problems, but the pace of coding ain't one.
by @paulisthrivving.xyz
medium.com/thrivve-part...
AI Won’t Fix Your Broken Delivery System. Here’s What Will.
You’re solving the wrong problem. And now you’re solving it faster.
medium.com
February 9, 2026 at 6:16 PM
If your cycle time sucks, automating coding won't do anything to your delivery pipeline.
You've got 99 problems, but the pace of coding ain't one.
by @paulisthrivving.xyz
medium.com/thrivve-part...
You've got 99 problems, but the pace of coding ain't one.
by @paulisthrivving.xyz
medium.com/thrivve-part...
Do you want to improve the engagement of your teams by 20%+ with literally no investment whatsoever?
Easy.
Distribute more autonomy across the board.
Seriously. It's enough to move engagement from "meh" to "engaged."
brodzinski.com/2026/01/lego...
Easy.
Distribute more autonomy across the board.
Seriously. It's enough to move engagement from "meh" to "engaged."
brodzinski.com/2026/01/lego...
What LEGO Can Teach Us about Autonomy and Engagement - Pawel Brodzinski on Leadership in Technology
A simple experiment with building LEGO models in groups demonstrates how important autonomy is for high engagement at work.
brodzinski.com
January 28, 2026 at 5:18 PM
Do you want to improve the engagement of your teams by 20%+ with literally no investment whatsoever?
Easy.
Distribute more autonomy across the board.
Seriously. It's enough to move engagement from "meh" to "engaged."
brodzinski.com/2026/01/lego...
Easy.
Distribute more autonomy across the board.
Seriously. It's enough to move engagement from "meh" to "engaged."
brodzinski.com/2026/01/lego...
"In time, he won’t just lose comprehension of what has been built; he’ll lose the ability to reason about it at all. Even if he can code, those skills will atrophy, along with the learning that sustains them."
@swardley.bsky.social on swarm AI vibe coding
www.linkedin.com/posts/simonw...
@swardley.bsky.social on swarm AI vibe coding
www.linkedin.com/posts/simonw...
The Architect and His Army of Twenty-Two | Simon Wardley
I love this. A swarm of agents and the human reduced to a very narrow form of defined decision making (which in itself can be automated) whilst the swarm makes nearly all the decisions.
In time, he w...
www.linkedin.com
January 28, 2026 at 2:42 PM
"In time, he won’t just lose comprehension of what has been built; he’ll lose the ability to reason about it at all. Even if he can code, those skills will atrophy, along with the learning that sustains them."
@swardley.bsky.social on swarm AI vibe coding
www.linkedin.com/posts/simonw...
@swardley.bsky.social on swarm AI vibe coding
www.linkedin.com/posts/simonw...
"A historic moment in solo founding and startup history."
36.3% of startups are founded solo.
A good question to ask: What is their success rate as compared to team-led startups?
36.3% of startups are founded solo.
A good question to ask: What is their success rate as compared to team-led startups?
January 28, 2026 at 8:26 AM
"A historic moment in solo founding and startup history."
36.3% of startups are founded solo.
A good question to ask: What is their success rate as compared to team-led startups?
36.3% of startups are founded solo.
A good question to ask: What is their success rate as compared to team-led startups?
Reposted by Pawel Brodzinski
experienced engineers: one change, test, one change, test
junior engineers: batch everything because they're in a hurry
this is exactly backwards
the person least capable of batching is the one most likely to batch
junior engineers: batch everything because they're in a hurry
this is exactly backwards
the person least capable of batching is the one most likely to batch
January 28, 2026 at 1:33 AM
experienced engineers: one change, test, one change, test
junior engineers: batch everything because they're in a hurry
this is exactly backwards
the person least capable of batching is the one most likely to batch
junior engineers: batch everything because they're in a hurry
this is exactly backwards
the person least capable of batching is the one most likely to batch
Realities of AI unicorns these days (which become unicorns before they even announce what they're working on).
January 27, 2026 at 2:14 PM
Realities of AI unicorns these days (which become unicorns before they even announce what they're working on).
Shipping something may be 10x cheaper.
Shipping meaningful stuff is about as difficult as it was.
Shipping meaningful stuff is about as difficult as it was.
January 26, 2026 at 12:59 PM
Shipping something may be 10x cheaper.
Shipping meaningful stuff is about as difficult as it was.
Shipping meaningful stuff is about as difficult as it was.
The new business plan:
1. Get a bunch of people from major AI companies.
2. Start a startup that will "rethink something-something AI."
3. Get half a billion in seed money round before you accomplish anything.
4. And they lived happily ever after.
techcrunch.com/2026/01/20/h...
1. Get a bunch of people from major AI companies.
2. Start a startup that will "rethink something-something AI."
3. Get half a billion in seed money round before you accomplish anything.
4. And they lived happily ever after.
techcrunch.com/2026/01/20/h...
Humans&, a 'human-centric' AI startup founded by Anthropic, xAI, Google alums, raised $480M seed round | TechCrunch
Humans&, a startup that believes AI should empower people, not replace them, has reportedly raised a $480 million seed round at a $4.48 billion valuation.
techcrunch.com
January 23, 2026 at 8:31 PM
The new business plan:
1. Get a bunch of people from major AI companies.
2. Start a startup that will "rethink something-something AI."
3. Get half a billion in seed money round before you accomplish anything.
4. And they lived happily ever after.
techcrunch.com/2026/01/20/h...
1. Get a bunch of people from major AI companies.
2. Start a startup that will "rethink something-something AI."
3. Get half a billion in seed money round before you accomplish anything.
4. And they lived happily ever after.
techcrunch.com/2026/01/20/h...
"The “wisdom of crowds” works only if the crowd is made of distinct individuals. When one operator can speak through thousands of masks, that independence collapses."
by @garymarcus.bsky.social
substack.com/home/post/p-...
by @garymarcus.bsky.social
substack.com/home/post/p-...
AI bot swarms threaten to undermine democracy
When AI Can Fake Majorities, Democracy Slips Away
substack.com
January 23, 2026 at 12:56 PM
"The “wisdom of crowds” works only if the crowd is made of distinct individuals. When one operator can speak through thousands of masks, that independence collapses."
by @garymarcus.bsky.social
substack.com/home/post/p-...
by @garymarcus.bsky.social
substack.com/home/post/p-...
"Many managers think they empower their teams to solve problems autonomously. Unfortunately, that autonomy only lasts when managers feel relaxed enough to “allow” that autonomy. When managers feel too much pressure, that autonomy vanishes."
by @jrothman.com
www.jrothman.com/newsletter/2...
by @jrothman.com
www.jrothman.com/newsletter/2...
Leadership Tip 25: How to Choose Autonomy to Avoid Malicious Compliance - Johanna Rothman
This is Johanna Rothman's January 2026 Pragmatic Manager newsletter. The unsubscribe link is at the bottom of this email. Many managers think they empower their teams to solve problems autonomously. U...
www.jrothman.com
January 23, 2026 at 10:13 AM
"Many managers think they empower their teams to solve problems autonomously. Unfortunately, that autonomy only lasts when managers feel relaxed enough to “allow” that autonomy. When managers feel too much pressure, that autonomy vanishes."
by @jrothman.com
www.jrothman.com/newsletter/2...
by @jrothman.com
www.jrothman.com/newsletter/2...
Reposted by Pawel Brodzinski
If you already know what's in your backlog for the next N iterations of your product, you're not doing iterative development. You're working through a plan.
January 23, 2026 at 9:35 AM
If you already know what's in your backlog for the next N iterations of your product, you're not doing iterative development. You're working through a plan.
Gallup: "After dropping in 2020 during the pandemic, employee engagement is on the rise again, reaching a record-high 23%."
That was 2 years ago. Now engagement dropped again.
Actually, whether it's 23% or 21%, it's still appalling.
Reason? Lack of autonomy: brodzinski.com/2026/01/auto...
That was 2 years ago. Now engagement dropped again.
Actually, whether it's 23% or 21%, it's still appalling.
Reason? Lack of autonomy: brodzinski.com/2026/01/auto...
Limited Autonomy Is the Main Reason for Low Engagement Levels - Pawel Brodzinski on Leadership in Technology
Hierarchy discourages leaders from distributing autonomy, which is the key reason for appalling engagment levels in the modern workplace.
brodzinski.com
January 22, 2026 at 12:12 PM
Gallup: "After dropping in 2020 during the pandemic, employee engagement is on the rise again, reaching a record-high 23%."
That was 2 years ago. Now engagement dropped again.
Actually, whether it's 23% or 21%, it's still appalling.
Reason? Lack of autonomy: brodzinski.com/2026/01/auto...
That was 2 years ago. Now engagement dropped again.
Actually, whether it's 23% or 21%, it's still appalling.
Reason? Lack of autonomy: brodzinski.com/2026/01/auto...
"We optimized for shipping trash, because the metrics for trash are easier to track than the metrics for quality."
www.linkedin.com/pulse/from-p...
www.linkedin.com/pulse/from-p...
From Products to Obstacles and Subscriptions: A Story of Decline
We used to build products. Now all we build is obstacles wrapped in subscriptions.
www.linkedin.com
January 22, 2026 at 9:44 AM
"We optimized for shipping trash, because the metrics for trash are easier to track than the metrics for quality."
www.linkedin.com/pulse/from-p...
www.linkedin.com/pulse/from-p...
Reposted by Pawel Brodzinski
Just 6 days remaining for my Effective Public Speaking Kickstarter. See the speaking myths here: www.jrothman.com/tag/speaking... and the Kickstarter here: www.kickstarter.com/projects/joh...
#Speaking #Kickstarter #ProjectWeLove #Fiction
#Speaking #Kickstarter #ProjectWeLove #Fiction
speaking myth Archives - Johanna Rothman
www.jrothman.com
January 21, 2026 at 4:52 PM
Just 6 days remaining for my Effective Public Speaking Kickstarter. See the speaking myths here: www.jrothman.com/tag/speaking... and the Kickstarter here: www.kickstarter.com/projects/joh...
#Speaking #Kickstarter #ProjectWeLove #Fiction
#Speaking #Kickstarter #ProjectWeLove #Fiction
"This is Continuous Integration 101. While a build is in progress, the software’s like Schrödinger’s Cat – simultaneously working and not working."
by @jasongorman.bsky.social
codemanship.wordpress.com/2026/01/21/f...
by @jasongorman.bsky.social
codemanship.wordpress.com/2026/01/21/f...
Finally! Proof That Agentic AI Scales (For Creating Broken Software)
Some of the marketing choices made by the “AI” industry over the last few years have seemed a little… odd. The latest is a “breakthrough” in “agentic AI” c…
codemanship.wordpress.com
January 21, 2026 at 2:52 PM
"This is Continuous Integration 101. While a build is in progress, the software’s like Schrödinger’s Cat – simultaneously working and not working."
by @jasongorman.bsky.social
codemanship.wordpress.com/2026/01/21/f...
by @jasongorman.bsky.social
codemanship.wordpress.com/2026/01/21/f...
1) Software engineering is not about building features
2) Refactoring is a business problem
3) Software engineering is not an engineering practice
4) LLMs will not replace software engineers
5) Architects don't make decisions
by @swardley.bsky.social & @tudorgirba.com
medium.com/feenk/rewild...
2) Refactoring is a business problem
3) Software engineering is not an engineering practice
4) LLMs will not replace software engineers
5) Architects don't make decisions
by @swardley.bsky.social & @tudorgirba.com
medium.com/feenk/rewild...
Rewilding Software Engineering
Chapter 6: Myths we tell ourselves
medium.com
January 21, 2026 at 10:28 AM
1) Software engineering is not about building features
2) Refactoring is a business problem
3) Software engineering is not an engineering practice
4) LLMs will not replace software engineers
5) Architects don't make decisions
by @swardley.bsky.social & @tudorgirba.com
medium.com/feenk/rewild...
2) Refactoring is a business problem
3) Software engineering is not an engineering practice
4) LLMs will not replace software engineers
5) Architects don't make decisions
by @swardley.bsky.social & @tudorgirba.com
medium.com/feenk/rewild...
A discovery phase for a new product development should fail more often than not.
Where I understand "failure" as not continuing the effort.
Failure rate is, in fact, a success rate.
pawelbrodzinski.substack.com/p/50-of-disc...
Where I understand "failure" as not continuing the effort.
Failure rate is, in fact, a success rate.
pawelbrodzinski.substack.com/p/50-of-disc...
50% of Discovery Phases Should Fail
We often consider product discovery a success when it's followed by development. Against such a measuring stick, we should expect most discovery phases to fail.
pawelbrodzinski.substack.com
January 19, 2026 at 3:11 PM
A discovery phase for a new product development should fail more often than not.
Where I understand "failure" as not continuing the effort.
Failure rate is, in fact, a success rate.
pawelbrodzinski.substack.com/p/50-of-disc...
Where I understand "failure" as not continuing the effort.
Failure rate is, in fact, a success rate.
pawelbrodzinski.substack.com/p/50-of-disc...
Reporter: “What makes you tick?”
Tolkien: “I don’t tick. I am not a machine. If I did tick, I should have no views on it, and you had better ask the winder.”
—J.R.R. Tolkien to Harvey Breit, New York Times Book Review, June 5, 1955, quoted in Letters, 217-18
Tolkien: “I don’t tick. I am not a machine. If I did tick, I should have no views on it, and you had better ask the winder.”
—J.R.R. Tolkien to Harvey Breit, New York Times Book Review, June 5, 1955, quoted in Letters, 217-18
January 13, 2026 at 1:28 PM
Reporter: “What makes you tick?”
Tolkien: “I don’t tick. I am not a machine. If I did tick, I should have no views on it, and you had better ask the winder.”
—J.R.R. Tolkien to Harvey Breit, New York Times Book Review, June 5, 1955, quoted in Letters, 217-18
Tolkien: “I don’t tick. I am not a machine. If I did tick, I should have no views on it, and you had better ask the winder.”
—J.R.R. Tolkien to Harvey Breit, New York Times Book Review, June 5, 1955, quoted in Letters, 217-18
The next multibillion-dollar funding series for OpenAI will starve tens of thousands of startups from their seed rounds.
These are the tradeoffs the startup ecosystem is looking at this year.
pawelbrodzinski.substack.com/p/2026-the-y...
These are the tradeoffs the startup ecosystem is looking at this year.
pawelbrodzinski.substack.com/p/2026-the-y...
2026: The Year of Scarce Funding for Early-Stage Startups
Growing gap between VC fundraising and spending, along with AI companies appetite for capital will get early-stage startups to change their tactics.
pawelbrodzinski.substack.com
January 13, 2026 at 10:23 AM
The next multibillion-dollar funding series for OpenAI will starve tens of thousands of startups from their seed rounds.
These are the tradeoffs the startup ecosystem is looking at this year.
pawelbrodzinski.substack.com/p/2026-the-y...
These are the tradeoffs the startup ecosystem is looking at this year.
pawelbrodzinski.substack.com/p/2026-the-y...
Reposted by Pawel Brodzinski
"estimation reveals that the increase in static analysis warnings and code complexity acts as a major factor causing long-term velocity slowdown"
Executive Summary: marathons are easy to start
arxiv.org/abs/2511.04427
Executive Summary: marathons are easy to start
arxiv.org/abs/2511.04427
Does AI-Assisted Coding Deliver? A Difference-in-Differences Study of Cursor's Impact on Software Projects
Large language models (LLMs) have demonstrated the promise to revolutionize the field of software engineering. Among other things, LLM agents are rapidly gaining momentum in their application to softw...
arxiv.org
January 13, 2026 at 5:12 AM
"estimation reveals that the increase in static analysis warnings and code complexity acts as a major factor causing long-term velocity slowdown"
Executive Summary: marathons are easy to start
arxiv.org/abs/2511.04427
Executive Summary: marathons are easy to start
arxiv.org/abs/2511.04427
As AI's role grows, digital startups are becoming more like brick-and-mortar businesses in one critical respect: The cost of operations is substantial and has to be factored in.
An average customer has to be net-positive for the business.
An average customer has to be net-positive for the business.
January 12, 2026 at 12:01 PM
As AI's role grows, digital startups are becoming more like brick-and-mortar businesses in one critical respect: The cost of operations is substantial and has to be factored in.
An average customer has to be net-positive for the business.
An average customer has to be net-positive for the business.
"When we talk about AI risk, most organisations think about hallucinations, bias, or compliance gaps. Few think about what happens when the training data itself becomes the attack vector."
www.linkedin.com/posts/openno...
www.linkedin.com/posts/openno...
From Celestial Cats to "<SUDO>": What New Backdoor Research Reinforces About Internal LLM Risk | OpenNovations
When we talk about AI risk, most organisations think about hallucinations, bias, or compliance gaps. Few think about what happens when the training data itself becomes the attack vector.
Two years ag...
www.linkedin.com
January 12, 2026 at 10:44 AM
"When we talk about AI risk, most organisations think about hallucinations, bias, or compliance gaps. Few think about what happens when the training data itself becomes the attack vector."
www.linkedin.com/posts/openno...
www.linkedin.com/posts/openno...
One measuring stick I used extensively last year (and will use it this year):
➤ Look at the existing trends.
➤ Extrapolate their dynamics.
➤ Ask: What does the endgame look like if the trends remain stable?
The answer, almost universally, is something entirely unsustainable.
Do the math.
➤ Look at the existing trends.
➤ Extrapolate their dynamics.
➤ Ask: What does the endgame look like if the trends remain stable?
The answer, almost universally, is something entirely unsustainable.
Do the math.
January 12, 2026 at 8:57 AM
One measuring stick I used extensively last year (and will use it this year):
➤ Look at the existing trends.
➤ Extrapolate their dynamics.
➤ Ask: What does the endgame look like if the trends remain stable?
The answer, almost universally, is something entirely unsustainable.
Do the math.
➤ Look at the existing trends.
➤ Extrapolate their dynamics.
➤ Ask: What does the endgame look like if the trends remain stable?
The answer, almost universally, is something entirely unsustainable.
Do the math.
"Any time saved by (their) AI prompting gets consumed by verification overhead, the work just gets passed along to someone else – in this case me."
ammil.industries/i-know-you-d...
ammil.industries/i-know-you-d...
I know you didn't write this · ammil industries
Exploring the ethics and productivity impacts of undisclosed AI-generated content in professional work, and why verification overhead is now the hidden cost of AI adoption.
ammil.industries
December 23, 2025 at 11:41 AM
"Any time saved by (their) AI prompting gets consumed by verification overhead, the work just gets passed along to someone else – in this case me."
ammil.industries/i-know-you-d...
ammil.industries/i-know-you-d...