Joey Kudish 🇺🇦🏳️🌈
@jkudish.com
I Build Software That Works. Let’s Build Yours. TetherMobile.com founder. Laravel + AI. Newsletter: jkudish.com/newsletter
Pinned
Joey Kudish 🇺🇦🏳️🌈
@jkudish.com
· Nov 24
Human in the Loop Newsletter
Subscribe to Human in the Loop for practical insights on AI-augmented coding, productivity hacks, and how to use AI as your development partner, not replacement.
jkudish.com
Tomorrow's newsletter: The 80/20 Rule Didn't Go Anywhere
AI didn't eliminate the hard work. It just revealed where it actually lives.
The easy stuff was never the real work - we were just busy doing it. Now AI compresses it, and we hit the hard parts faster.
jkudish.com/newsletter?u...
AI didn't eliminate the hard work. It just revealed where it actually lives.
The easy stuff was never the real work - we were just busy doing it. Now AI compresses it, and we hit the hard parts faster.
jkudish.com/newsletter?u...
AI made me faster.
It also made it easier to ship output before I’ve done the thinking.
So I’m adding friction back into my workflow.
Spec first. Then code.
It also made it easier to ship output before I’ve done the thinking.
So I’m adding friction back into my workflow.
Spec first. Then code.
January 21, 2026 at 3:04 PM
AI made me faster.
It also made it easier to ship output before I’ve done the thinking.
So I’m adding friction back into my workflow.
Spec first. Then code.
It also made it easier to ship output before I’ve done the thinking.
So I’m adding friction back into my workflow.
Spec first. Then code.
I’ve had a long career of working on many projects at once.
These days it’s splitting my time between client work and product work.
The hours aren’t the hard part.
The context switching is.
So I capture decisions as I go. Notes, links, next steps. Automatically, with AI.
These days it’s splitting my time between client work and product work.
The hours aren’t the hard part.
The context switching is.
So I capture decisions as I go. Notes, links, next steps. Automatically, with AI.
January 21, 2026 at 7:50 AM
I’ve had a long career of working on many projects at once.
These days it’s splitting my time between client work and product work.
The hours aren’t the hard part.
The context switching is.
So I capture decisions as I go. Notes, links, next steps. Automatically, with AI.
These days it’s splitting my time between client work and product work.
The hours aren’t the hard part.
The context switching is.
So I capture decisions as I go. Notes, links, next steps. Automatically, with AI.
In the before times, I thought the hard part was writing code.
It isn’t.
Now more than ever with AI, the hard part is choosing what not to build.
It isn’t.
Now more than ever with AI, the hard part is choosing what not to build.
January 21, 2026 at 2:57 AM
In the before times, I thought the hard part was writing code.
It isn’t.
Now more than ever with AI, the hard part is choosing what not to build.
It isn’t.
Now more than ever with AI, the hard part is choosing what not to build.
A pattern I keep seeing:
Unclear thinking arrives faster than clear writing.
If I can’t explain something simply, I’m usually not done thinking yet.
Unclear thinking arrives faster than clear writing.
If I can’t explain something simply, I’m usually not done thinking yet.
January 21, 2026 at 12:57 AM
A pattern I keep seeing:
Unclear thinking arrives faster than clear writing.
If I can’t explain something simply, I’m usually not done thinking yet.
Unclear thinking arrives faster than clear writing.
If I can’t explain something simply, I’m usually not done thinking yet.
Best way I’ve found to use AI:
Don’t outsource judgment.
Outsource attention.
Don’t outsource judgment.
Outsource attention.
January 20, 2026 at 7:56 PM
Best way I’ve found to use AI:
Don’t outsource judgment.
Outsource attention.
Don’t outsource judgment.
Outsource attention.
Working on a quote automation feature for a client right now.
Not the model or the tooling.
It’s getting it production-ready.
Edge cases.
Data consistency.
Same input, same output.
What surprised you most when you shipped your first AI feature?
Not the model or the tooling.
It’s getting it production-ready.
Edge cases.
Data consistency.
Same input, same output.
What surprised you most when you shipped your first AI feature?
January 20, 2026 at 5:56 PM
Working on a quote automation feature for a client right now.
Not the model or the tooling.
It’s getting it production-ready.
Edge cases.
Data consistency.
Same input, same output.
What surprised you most when you shipped your first AI feature?
Not the model or the tooling.
It’s getting it production-ready.
Edge cases.
Data consistency.
Same input, same output.
What surprised you most when you shipped your first AI feature?
Codex Bar and Clawd are two of my favourite software at the moment.
Well done @steipete 👏
Well done @steipete 👏
January 20, 2026 at 3:02 PM
Codex Bar and Clawd are two of my favourite software at the moment.
Well done @steipete 👏
Well done @steipete 👏
I build software because I like solving problems.
I build my own products because I like choosing which problems to solve.
I build my own products because I like choosing which problems to solve.
January 16, 2026 at 1:01 AM
I build software because I like solving problems.
I build my own products because I like choosing which problems to solve.
I build my own products because I like choosing which problems to solve.
don't wait until project one is "done" to start project two.
with AI, you can maintain multiple projects without burning out.
after building software for two decades, I can tell you that "done" never comes. ride the momentum while you have it.
with AI, you can maintain multiple projects without burning out.
after building software for two decades, I can tell you that "done" never comes. ride the momentum while you have it.
January 15, 2026 at 7:57 PM
don't wait until project one is "done" to start project two.
with AI, you can maintain multiple projects without burning out.
after building software for two decades, I can tell you that "done" never comes. ride the momentum while you have it.
with AI, you can maintain multiple projects without burning out.
after building software for two decades, I can tell you that "done" never comes. ride the momentum while you have it.
AI won't save you from unclear thinking.
but it'll make the consequences of unclear thinking arrive faster.
but it'll make the consequences of unclear thinking arrive faster.
January 15, 2026 at 6:58 PM
AI won't save you from unclear thinking.
but it'll make the consequences of unclear thinking arrive faster.
but it'll make the consequences of unclear thinking arrive faster.
juggling 5 projects today. different codebases, different stacks, different priorities.
a year ago this would've been chaos. now it's just Thursday.
AI ramps me back up in minutes. I just make the calls.
a year ago this would've been chaos. now it's just Thursday.
AI ramps me back up in minutes. I just make the calls.
January 15, 2026 at 6:00 PM
juggling 5 projects today. different codebases, different stacks, different priorities.
a year ago this would've been chaos. now it's just Thursday.
AI ramps me back up in minutes. I just make the calls.
a year ago this would've been chaos. now it's just Thursday.
AI ramps me back up in minutes. I just make the calls.
the developers who thrive won't be the fastest coders.
they'll be the ones who document well and work with AI to shape the architecture.
architecture > implementation speed.
they'll be the ones who document well and work with AI to shape the architecture.
architecture > implementation speed.
January 15, 2026 at 5:00 PM
the developers who thrive won't be the fastest coders.
they'll be the ones who document well and work with AI to shape the architecture.
architecture > implementation speed.
they'll be the ones who document well and work with AI to shape the architecture.
architecture > implementation speed.
my new daily routine:
mornings: review AI output.
afternoons: dedicated work with the AI - refining or thinking about what's next.
evenings: set up the next overnight AI runs.
mornings: review AI output.
afternoons: dedicated work with the AI - refining or thinking about what's next.
evenings: set up the next overnight AI runs.
January 15, 2026 at 3:56 PM
my new daily routine:
mornings: review AI output.
afternoons: dedicated work with the AI - refining or thinking about what's next.
evenings: set up the next overnight AI runs.
mornings: review AI output.
afternoons: dedicated work with the AI - refining or thinking about what's next.
evenings: set up the next overnight AI runs.
shipping fast is overrated.
shipping the right thing is underrated.
shipping the right thing is underrated.
January 15, 2026 at 3:03 PM
shipping fast is overrated.
shipping the right thing is underrated.
shipping the right thing is underrated.
Reposted by Joey Kudish 🇺🇦🏳️🌈
A muscle worth building for better prompts:
Frame your feedback around the end result, goal, or emotion you want to get to, with whatever you're building or creating.
Avoid micro-directing specific changes and how to make them.
Frame your feedback around the end result, goal, or emotion you want to get to, with whatever you're building or creating.
Avoid micro-directing specific changes and how to make them.
January 14, 2026 at 7:39 PM
A muscle worth building for better prompts:
Frame your feedback around the end result, goal, or emotion you want to get to, with whatever you're building or creating.
Avoid micro-directing specific changes and how to make them.
Frame your feedback around the end result, goal, or emotion you want to get to, with whatever you're building or creating.
Avoid micro-directing specific changes and how to make them.
Reposted by Joey Kudish 🇺🇦🏳️🌈
The more I use these agents (that now do write code that is pretty good, but ofc I need to verify and keep them in check), the more I feel we're going to see the "Microsoft Frontpage" moment in tech:
Frontpage DID make every and all web devs redundant from 2007. As we all know.
Frontpage DID make every and all web devs redundant from 2007. As we all know.
January 14, 2026 at 7:48 PM
The more I use these agents (that now do write code that is pretty good, but ofc I need to verify and keep them in check), the more I feel we're going to see the "Microsoft Frontpage" moment in tech:
Frontpage DID make every and all web devs redundant from 2007. As we all know.
Frontpage DID make every and all web devs redundant from 2007. As we all know.
anyone else attending Nomad Summit in Chiang Mai this weekend?
would love to meet other builders in person.
would love to meet other builders in person.
January 15, 2026 at 9:03 AM
anyone else attending Nomad Summit in Chiang Mai this weekend?
would love to meet other builders in person.
would love to meet other builders in person.
the hardest part of building solo isn't the work.
it's deciding which work matters most today.
it's deciding which work matters most today.
January 15, 2026 at 3:01 AM
the hardest part of building solo isn't the work.
it's deciding which work matters most today.
it's deciding which work matters most today.
what feature did you almost cut from your product that ended up being the one users love most?
January 15, 2026 at 1:03 AM
what feature did you almost cut from your product that ended up being the one users love most?
the developers who'll struggle most aren't the ones who can't code.
they're the ones who can't describe what they want built.
AI raised the floor for implementation. clarity became the ceiling.
they're the ones who can't describe what they want built.
AI raised the floor for implementation. clarity became the ceiling.
January 14, 2026 at 7:59 PM
the developers who'll struggle most aren't the ones who can't code.
they're the ones who can't describe what they want built.
AI raised the floor for implementation. clarity became the ceiling.
they're the ones who can't describe what they want built.
AI raised the floor for implementation. clarity became the ceiling.
shipped code in 4 languages this week. PHP, JavaScript, Python, Rust.
not because I planned it. because each project needed the right tool.
AI makes switching contexts painless. I focus on what to build, not remembering syntax.
not because I planned it. because each project needed the right tool.
AI makes switching contexts painless. I focus on what to build, not remembering syntax.
January 14, 2026 at 7:02 PM
shipped code in 4 languages this week. PHP, JavaScript, Python, Rust.
not because I planned it. because each project needed the right tool.
AI makes switching contexts painless. I focus on what to build, not remembering syntax.
not because I planned it. because each project needed the right tool.
AI makes switching contexts painless. I focus on what to build, not remembering syntax.
what's your setup for running AI overnight?
I'm using Claude Code with Ralph loops - queue up tasks before bed, review in the morning.
curious what others are doing.
I'm using Claude Code with Ralph loops - queue up tasks before bed, review in the morning.
curious what others are doing.
January 14, 2026 at 6:02 PM
what's your setup for running AI overnight?
I'm using Claude Code with Ralph loops - queue up tasks before bed, review in the morning.
curious what others are doing.
I'm using Claude Code with Ralph loops - queue up tasks before bed, review in the morning.
curious what others are doing.
the last 10% before launch is the hardest 10%.
not because it's complex. because it's all the small things you kept saying "I'll fix later."
later is now. I'm putting finishing touches on several projects that will launch soon!
not because it's complex. because it's all the small things you kept saying "I'll fix later."
later is now. I'm putting finishing touches on several projects that will launch soon!
January 14, 2026 at 5:04 PM
the last 10% before launch is the hardest 10%.
not because it's complex. because it's all the small things you kept saying "I'll fix later."
later is now. I'm putting finishing touches on several projects that will launch soon!
not because it's complex. because it's all the small things you kept saying "I'll fix later."
later is now. I'm putting finishing touches on several projects that will launch soon!
my 18+ years of experience are in PHP + Javascript, but over the last months I've confidently shipped Rust, Python, and even some Lua.
once you know the fundamentals of software engineering, you can become good enough (with the help of AI) in any language and solve real problems.
once you know the fundamentals of software engineering, you can become good enough (with the help of AI) in any language and solve real problems.
January 14, 2026 at 4:03 PM
my 18+ years of experience are in PHP + Javascript, but over the last months I've confidently shipped Rust, Python, and even some Lua.
once you know the fundamentals of software engineering, you can become good enough (with the help of AI) in any language and solve real problems.
once you know the fundamentals of software engineering, you can become good enough (with the help of AI) in any language and solve real problems.
I don't separate "client work" from "my work."
same standards. same tools. same care.
the only difference is who's paying.
same standards. same tools. same care.
the only difference is who's paying.
January 14, 2026 at 3:01 PM
I don't separate "client work" from "my work."
same standards. same tools. same care.
the only difference is who's paying.
same standards. same tools. same care.
the only difference is who's paying.