Toby Rogers
@tobyrogers.pm
the punk pm | talking about product, tech creativity, and innovation | head of product @ https://hedgehoglab.com | retired music journo | blue peter badge winner | adhd https://tobyrogers.pm
Customers don’t care if your product uses AI.
They care about the results they can get from using it.
AI is an enabler, not a feature.
Investors might need you to sound like an “AI company”, but your users don’t.
They just want something that works.
They care about the results they can get from using it.
AI is an enabler, not a feature.
Investors might need you to sound like an “AI company”, but your users don’t.
They just want something that works.
November 18, 2025 at 10:46 AM
Customers don’t care if your product uses AI.
They care about the results they can get from using it.
AI is an enabler, not a feature.
Investors might need you to sound like an “AI company”, but your users don’t.
They just want something that works.
They care about the results they can get from using it.
AI is an enabler, not a feature.
Investors might need you to sound like an “AI company”, but your users don’t.
They just want something that works.
What we’re seeing with AI is the principle of least effort at massive scale.
When it’s quicker to get a model to do something, people will.
The risk is becoming the horror story when AI fails at something it was never meant to do.
Powerful? Yes.
A universal shortcut? No.
When it’s quicker to get a model to do something, people will.
The risk is becoming the horror story when AI fails at something it was never meant to do.
Powerful? Yes.
A universal shortcut? No.
November 13, 2025 at 11:52 AM
What we’re seeing with AI is the principle of least effort at massive scale.
When it’s quicker to get a model to do something, people will.
The risk is becoming the horror story when AI fails at something it was never meant to do.
Powerful? Yes.
A universal shortcut? No.
When it’s quicker to get a model to do something, people will.
The risk is becoming the horror story when AI fails at something it was never meant to do.
Powerful? Yes.
A universal shortcut? No.
AI is a force multiplier for people who know what they’re doing already, not a shortcut to domain expertise.
November 12, 2025 at 10:12 AM
AI is a force multiplier for people who know what they’re doing already, not a shortcut to domain expertise.
One of the eternal truths of product management: finding a problem worth solving and building a business around it is hard.
AI doesn’t change that, it just changes how we go about it.
We’ll always need people who can join the dots.
AI doesn’t change that, it just changes how we go about it.
We’ll always need people who can join the dots.
November 11, 2025 at 11:40 AM
One of the eternal truths of product management: finding a problem worth solving and building a business around it is hard.
AI doesn’t change that, it just changes how we go about it.
We’ll always need people who can join the dots.
AI doesn’t change that, it just changes how we go about it.
We’ll always need people who can join the dots.
It's a few years old now, but I love this article on the history of iOS icons. I miss the quirky, skeuomorphic feel of the early app era.
A Short History of iOS App Icons
Writing about the big beautiful mess that is making things for the world wide web.
blog.jim-nielsen.com
November 7, 2025 at 9:09 AM
It's a few years old now, but I love this article on the history of iOS icons. I miss the quirky, skeuomorphic feel of the early app era.
PMs who focus only on users will be replaced by AI.
PMs who focus on business impact won’t.
Product management isn’t about pixels or process—it’s about purpose, positioning, and profit.
PMs who focus on business impact won’t.
Product management isn’t about pixels or process—it’s about purpose, positioning, and profit.
November 6, 2025 at 8:20 AM
PMs who focus only on users will be replaced by AI.
PMs who focus on business impact won’t.
Product management isn’t about pixels or process—it’s about purpose, positioning, and profit.
PMs who focus on business impact won’t.
Product management isn’t about pixels or process—it’s about purpose, positioning, and profit.
AI is productising the “best practice” of product management.
GPTs can now analyse your research, write your PRDs, even prioritise your roadmap.
So what’s left?
Creativity, clarity, conviction.
The human stuff.
That’s what’ll set great PMs apart from the robots.
GPTs can now analyse your research, write your PRDs, even prioritise your roadmap.
So what’s left?
Creativity, clarity, conviction.
The human stuff.
That’s what’ll set great PMs apart from the robots.
November 5, 2025 at 8:08 AM
AI is productising the “best practice” of product management.
GPTs can now analyse your research, write your PRDs, even prioritise your roadmap.
So what’s left?
Creativity, clarity, conviction.
The human stuff.
That’s what’ll set great PMs apart from the robots.
GPTs can now analyse your research, write your PRDs, even prioritise your roadmap.
So what’s left?
Creativity, clarity, conviction.
The human stuff.
That’s what’ll set great PMs apart from the robots.
People don’t buy solutions to problems they don’t think they have.
Still true in the AI age.
Just because AI makes new things possible, it doesn’t mean anyone cares.
Find valuable problems first then use AI to solve them, only if it’s the right tool for the job.
Still true in the AI age.
Just because AI makes new things possible, it doesn’t mean anyone cares.
Find valuable problems first then use AI to solve them, only if it’s the right tool for the job.
November 4, 2025 at 9:23 AM
People don’t buy solutions to problems they don’t think they have.
Still true in the AI age.
Just because AI makes new things possible, it doesn’t mean anyone cares.
Find valuable problems first then use AI to solve them, only if it’s the right tool for the job.
Still true in the AI age.
Just because AI makes new things possible, it doesn’t mean anyone cares.
Find valuable problems first then use AI to solve them, only if it’s the right tool for the job.
LLMs stand on the shoulders of giants—but they don’t innovate, they regurgitate.
November 3, 2025 at 12:04 PM
LLMs stand on the shoulders of giants—but they don’t innovate, they regurgitate.
Product development will always need humans in the loop.
In the AI age, PMs are best placed to be them.
In the AI age, PMs are best placed to be them.
October 31, 2025 at 8:10 AM
Product development will always need humans in the loop.
In the AI age, PMs are best placed to be them.
In the AI age, PMs are best placed to be them.
AI will change how we build products, but not why.
People still want solutions to the same old problems. The tools evolve, the fundamentals don’t.
If you want to stay the course, stay focused on what’s timeless, not just what’s new.
People still want solutions to the same old problems. The tools evolve, the fundamentals don’t.
If you want to stay the course, stay focused on what’s timeless, not just what’s new.
October 30, 2025 at 8:52 AM
AI will change how we build products, but not why.
People still want solutions to the same old problems. The tools evolve, the fundamentals don’t.
If you want to stay the course, stay focused on what’s timeless, not just what’s new.
People still want solutions to the same old problems. The tools evolve, the fundamentals don’t.
If you want to stay the course, stay focused on what’s timeless, not just what’s new.
AI is changing what it means to be a “user.”
Soon, your user won’t be a person, it’ll be an agent acting on behalf of one.
That’s a huge shift in how we think about UX and interaction design.
When you’re building for robots instead of people, all your frames of reference need to change.
Soon, your user won’t be a person, it’ll be an agent acting on behalf of one.
That’s a huge shift in how we think about UX and interaction design.
When you’re building for robots instead of people, all your frames of reference need to change.
October 29, 2025 at 8:15 AM
AI is changing what it means to be a “user.”
Soon, your user won’t be a person, it’ll be an agent acting on behalf of one.
That’s a huge shift in how we think about UX and interaction design.
When you’re building for robots instead of people, all your frames of reference need to change.
Soon, your user won’t be a person, it’ll be an agent acting on behalf of one.
That’s a huge shift in how we think about UX and interaction design.
When you’re building for robots instead of people, all your frames of reference need to change.
Brilliant piece from @anildash.com on how the AI hype train is blinding us to its real value—and limits—as “normal” tech.
PMs: stop treating AI like magic.
It’s just another tool to help you solve problems, with the same flaws and constraints as all the others.
PMs: stop treating AI like magic.
It’s just another tool to help you solve problems, with the same flaws and constraints as all the others.
The Majority AI View - Anil Dash
A blog about making culture. Since 1999.
www.anildash.com
October 28, 2025 at 1:14 PM
Brilliant piece from @anildash.com on how the AI hype train is blinding us to its real value—and limits—as “normal” tech.
PMs: stop treating AI like magic.
It’s just another tool to help you solve problems, with the same flaws and constraints as all the others.
PMs: stop treating AI like magic.
It’s just another tool to help you solve problems, with the same flaws and constraints as all the others.
LLMs can help you be creative, but they can’t create for you. You’ve still got to join the dots.
October 28, 2025 at 8:18 AM
LLMs can help you be creative, but they can’t create for you. You’ve still got to join the dots.
AI isn’t all hype, it’s a tectonic shift in how products get built.
With a billion people already using LLMs regularly, you can’t just wait for the bubble to pop.
If you’re a PM and you’re not figuring out how to work with AI now, you’ll be figuring out how to survive because of it later.
With a billion people already using LLMs regularly, you can’t just wait for the bubble to pop.
If you’re a PM and you’re not figuring out how to work with AI now, you’ll be figuring out how to survive because of it later.
October 27, 2025 at 11:46 AM
AI isn’t all hype, it’s a tectonic shift in how products get built.
With a billion people already using LLMs regularly, you can’t just wait for the bubble to pop.
If you’re a PM and you’re not figuring out how to work with AI now, you’ll be figuring out how to survive because of it later.
With a billion people already using LLMs regularly, you can’t just wait for the bubble to pop.
If you’re a PM and you’re not figuring out how to work with AI now, you’ll be figuring out how to survive because of it later.
When the web is drowning in LLM-generated content, what we need isn’t more curators—it’s more sense-makers.
The humans (not algorithms) who help us find the signal in the noise.
The humans (not algorithms) who help us find the signal in the noise.
October 26, 2025 at 8:42 PM
When the web is drowning in LLM-generated content, what we need isn’t more curators—it’s more sense-makers.
The humans (not algorithms) who help us find the signal in the noise.
The humans (not algorithms) who help us find the signal in the noise.
Three things I’ve been thinking about this week:
- AI browsers turning the web into a workspace.
- Taste as a product skill.
- Fun as strategy.
Everything’s speeding up, but the pattern’s simple: execution is easy, good judgement is hard.
- AI browsers turning the web into a workspace.
- Taste as a product skill.
- Fun as strategy.
Everything’s speeding up, but the pattern’s simple: execution is easy, good judgement is hard.
October 26, 2025 at 8:35 AM
Three things I’ve been thinking about this week:
- AI browsers turning the web into a workspace.
- Taste as a product skill.
- Fun as strategy.
Everything’s speeding up, but the pattern’s simple: execution is easy, good judgement is hard.
- AI browsers turning the web into a workspace.
- Taste as a product skill.
- Fun as strategy.
Everything’s speeding up, but the pattern’s simple: execution is easy, good judgement is hard.
When AI makes building software almost effortless, how many people will build their own tools instead of buying someone else’s?
And what happens to the companies selling software when they’re no longer just competing with rivals, but with what their customers can spin up themselves?
And what happens to the companies selling software when they’re no longer just competing with rivals, but with what their customers can spin up themselves?
October 25, 2025 at 8:38 AM
When AI makes building software almost effortless, how many people will build their own tools instead of buying someone else’s?
And what happens to the companies selling software when they’re no longer just competing with rivals, but with what their customers can spin up themselves?
And what happens to the companies selling software when they’re no longer just competing with rivals, but with what their customers can spin up themselves?
AI browsers are a good reminder that product’s core job hasn’t changed: solve real problems.
The “interface” will just be whatever form best fits the moment—voice, chat, app, ambient.
The job for PMs and designers now is to think beyond screens. Where else can we bring value into people’s lives?
The “interface” will just be whatever form best fits the moment—voice, chat, app, ambient.
The job for PMs and designers now is to think beyond screens. Where else can we bring value into people’s lives?
October 24, 2025 at 5:37 AM
AI browsers are a good reminder that product’s core job hasn’t changed: solve real problems.
The “interface” will just be whatever form best fits the moment—voice, chat, app, ambient.
The job for PMs and designers now is to think beyond screens. Where else can we bring value into people’s lives?
The “interface” will just be whatever form best fits the moment—voice, chat, app, ambient.
The job for PMs and designers now is to think beyond screens. Where else can we bring value into people’s lives?
I built a custom GPT to generate wacky but believable product ideas.
Why?
Not to outsource innovation, but to give myself a starting point for vibe-coding experiments.
When you know the output’s disposable, you stop worrying about being “right.”
That’s when the fun (and real creativity) starts.
Why?
Not to outsource innovation, but to give myself a starting point for vibe-coding experiments.
When you know the output’s disposable, you stop worrying about being “right.”
That’s when the fun (and real creativity) starts.
October 23, 2025 at 2:15 PM
I built a custom GPT to generate wacky but believable product ideas.
Why?
Not to outsource innovation, but to give myself a starting point for vibe-coding experiments.
When you know the output’s disposable, you stop worrying about being “right.”
That’s when the fun (and real creativity) starts.
Why?
Not to outsource innovation, but to give myself a starting point for vibe-coding experiments.
When you know the output’s disposable, you stop worrying about being “right.”
That’s when the fun (and real creativity) starts.
AI isn’t a feature, it’s a lever.
If it doesn’t make your customer’s life better, don’t use it.
If it doesn’t make your customer’s life better, don’t use it.
October 23, 2025 at 10:27 AM
AI isn’t a feature, it’s a lever.
If it doesn’t make your customer’s life better, don’t use it.
If it doesn’t make your customer’s life better, don’t use it.
If you think about it, apps are just wrappers for digital experiences.
As AI becomes a wrapper for apps, it raises a bigger question:
What happens when the wrapper eats the product?
Do we actually need apps anymore?
As AI becomes a wrapper for apps, it raises a bigger question:
What happens when the wrapper eats the product?
Do we actually need apps anymore?
October 22, 2025 at 7:58 PM
If you think about it, apps are just wrappers for digital experiences.
As AI becomes a wrapper for apps, it raises a bigger question:
What happens when the wrapper eats the product?
Do we actually need apps anymore?
As AI becomes a wrapper for apps, it raises a bigger question:
What happens when the wrapper eats the product?
Do we actually need apps anymore?
First impressions of ChatGPT Atlas:
The UX feels so close to Dia I had to check I’d actually switched browser.
Where it really shines (obviously) is the deep ChatGPT integration.
Starting an ad-hoc chat around a web page, then pulling it into a project with full context is wildly powerful.
The UX feels so close to Dia I had to check I’d actually switched browser.
Where it really shines (obviously) is the deep ChatGPT integration.
Starting an ad-hoc chat around a web page, then pulling it into a project with full context is wildly powerful.
October 22, 2025 at 12:52 PM
First impressions of ChatGPT Atlas:
The UX feels so close to Dia I had to check I’d actually switched browser.
Where it really shines (obviously) is the deep ChatGPT integration.
Starting an ad-hoc chat around a web page, then pulling it into a project with full context is wildly powerful.
The UX feels so close to Dia I had to check I’d actually switched browser.
Where it really shines (obviously) is the deep ChatGPT integration.
Starting an ad-hoc chat around a web page, then pulling it into a project with full context is wildly powerful.
In a world of vibe-coded slop products, taste will be a product management superpower.
October 22, 2025 at 9:04 AM
In a world of vibe-coded slop products, taste will be a product management superpower.
AI browsers will change everything.
Not just how we search, but how we use digital tools.
As a PM, you need to start asking: how do I deliver value in an AI-first world?
With LLM-powered browsers like Dia, Comet and Atlas, plus OpenAI’s Apps SDK, the internet’s about to feel very different.
Not just how we search, but how we use digital tools.
As a PM, you need to start asking: how do I deliver value in an AI-first world?
With LLM-powered browsers like Dia, Comet and Atlas, plus OpenAI’s Apps SDK, the internet’s about to feel very different.
October 21, 2025 at 9:26 PM
AI browsers will change everything.
Not just how we search, but how we use digital tools.
As a PM, you need to start asking: how do I deliver value in an AI-first world?
With LLM-powered browsers like Dia, Comet and Atlas, plus OpenAI’s Apps SDK, the internet’s about to feel very different.
Not just how we search, but how we use digital tools.
As a PM, you need to start asking: how do I deliver value in an AI-first world?
With LLM-powered browsers like Dia, Comet and Atlas, plus OpenAI’s Apps SDK, the internet’s about to feel very different.