Jeffrey Guenther
@jeffreyguenther.com
Founder @loamstudios.com, @loamcapitalmanagement.com
Product-minded developer
Proudly 🇨🇦
Product-minded developer
Proudly 🇨🇦
I think making software is more like gardening than running a factory.
Factories make goods based on a spec. The goal is zero variation from the spec.
Tending a garden means responding to what the garden is telling you about what it needs.
Factories make goods based on a spec. The goal is zero variation from the spec.
Tending a garden means responding to what the garden is telling you about what it needs.
February 8, 2026 at 7:56 PM
I think making software is more like gardening than running a factory.
Factories make goods based on a spec. The goal is zero variation from the spec.
Tending a garden means responding to what the garden is telling you about what it needs.
Factories make goods based on a spec. The goal is zero variation from the spec.
Tending a garden means responding to what the garden is telling you about what it needs.
With the rise of agents, crypto based transactions are going to become the default.
Have an agent make a payment is much harder with a credit card.
There's no CLI equivalent of the iframe credit card form or checkout page.
Have an agent make a payment is much harder with a credit card.
There's no CLI equivalent of the iframe credit card form or checkout page.
February 8, 2026 at 4:59 PM
With the rise of agents, crypto based transactions are going to become the default.
Have an agent make a payment is much harder with a credit card.
There's no CLI equivalent of the iframe credit card form or checkout page.
Have an agent make a payment is much harder with a credit card.
There's no CLI equivalent of the iframe credit card form or checkout page.
Rest Day —
“The LORD your God is in your midst, a mighty one who will save; he will rejoice over you with gladness; he will quiet you by his love; he will exult over you with loud singing.”
Zephaniah 3:17 ESV
“The LORD your God is in your midst, a mighty one who will save; he will rejoice over you with gladness; he will quiet you by his love; he will exult over you with loud singing.”
Zephaniah 3:17 ESV
February 7, 2026 at 2:01 PM
Rest Day —
“The LORD your God is in your midst, a mighty one who will save; he will rejoice over you with gladness; he will quiet you by his love; he will exult over you with loud singing.”
Zephaniah 3:17 ESV
“The LORD your God is in your midst, a mighty one who will save; he will rejoice over you with gladness; he will quiet you by his love; he will exult over you with loud singing.”
Zephaniah 3:17 ESV
You remember when World of War Craft was released and how people got fired and were found having not left their house in days....
ya...
This is where we're at with AI right now.
ya...
This is where we're at with AI right now.
February 6, 2026 at 10:13 PM
You remember when World of War Craft was released and how people got fired and were found having not left their house in days....
ya...
This is where we're at with AI right now.
ya...
This is where we're at with AI right now.
A tiling window manager with multiple workspaces might become a necessity in this new highly parallel world.
@OmarchyLinux I'm looking at you. 🤩
@OmarchyLinux I'm looking at you. 🤩
February 6, 2026 at 9:32 PM
A tiling window manager with multiple workspaces might become a necessity in this new highly parallel world.
@OmarchyLinux I'm looking at you. 🤩
@OmarchyLinux I'm looking at you. 🤩
As I push to do more and more work with as little intervention as possible, the main piece of friction I'm noticing is that the models don't follow workflow instructions reliably.
I have commit and pull request workflows outlined in Skills and reference them. They are only sometimes used.
I have commit and pull request workflows outlined in Skills and reference them. They are only sometimes used.
February 6, 2026 at 2:55 PM
As I push to do more and more work with as little intervention as possible, the main piece of friction I'm noticing is that the models don't follow workflow instructions reliably.
I have commit and pull request workflows outlined in Skills and reference them. They are only sometimes used.
I have commit and pull request workflows outlined in Skills and reference them. They are only sometimes used.
If you let the model lead your creativity, you will get something that's derivative of the sum of past human knowledge.
If you use the model to augment your creativity, you will get something that leverages the sum of all human knowledge to elevate your creativity.
If you use the model to augment your creativity, you will get something that leverages the sum of all human knowledge to elevate your creativity.
February 4, 2026 at 5:02 PM
If you let the model lead your creativity, you will get something that's derivative of the sum of past human knowledge.
If you use the model to augment your creativity, you will get something that leverages the sum of all human knowledge to elevate your creativity.
If you use the model to augment your creativity, you will get something that leverages the sum of all human knowledge to elevate your creativity.
The more I do this agentic programming thing, the more I'm convinced that agent session transcripts are a valuable artifact to the completion of a task on par with the code generated, the design work, and any other supporting product documentation.
February 4, 2026 at 1:40 PM
The more I do this agentic programming thing, the more I'm convinced that agent session transcripts are a valuable artifact to the completion of a task on par with the code generated, the design work, and any other supporting product documentation.
Overwhelm.
UX to combat overwhelm is something you're about to hear a bunch about.
UX to combat overwhelm is something you're about to hear a bunch about.
February 3, 2026 at 8:01 PM
Overwhelm.
UX to combat overwhelm is something you're about to hear a bunch about.
UX to combat overwhelm is something you're about to hear a bunch about.
I got sick of waiting for CI to run where most of the time spent was setting up the containers.
I've moved to using github.com/basecamp/gh...
I can run the same checks locally and then signoff on the branch. This now takes seconds instead of minutes.
I've moved to using github.com/basecamp/gh...
I can run the same checks locally and then signoff on the branch. This now takes seconds instead of minutes.
GitHub - basecamp/gh-signoff: Local CI. Sign off on your own work.
Local CI. Sign off on your own work. Contribute to basecamp/gh-signoff development by creating an account on GitHub.
github.com
February 2, 2026 at 3:31 PM
I got sick of waiting for CI to run where most of the time spent was setting up the containers.
I've moved to using github.com/basecamp/gh...
I can run the same checks locally and then signoff on the branch. This now takes seconds instead of minutes.
I've moved to using github.com/basecamp/gh...
I can run the same checks locally and then signoff on the branch. This now takes seconds instead of minutes.
Having spent 4+ hours now running parallel agent sessions each working a separate ticket, I’m convinced that a sandbox is needed for each agent. Managing things on a single box has lots of paper cuts and foot guns.
February 2, 2026 at 1:37 AM
Having spent 4+ hours now running parallel agent sessions each working a separate ticket, I’m convinced that a sandbox is needed for each agent. Managing things on a single box has lots of paper cuts and foot guns.
I used to say, "Great idea, but too much work for now given the current priorities."
Now, I say, "Great idea, let me get an agent on it."
Now, I say, "Great idea, let me get an agent on it."
January 30, 2026 at 6:00 PM
I used to say, "Great idea, but too much work for now given the current priorities."
Now, I say, "Great idea, let me get an agent on it."
Now, I say, "Great idea, let me get an agent on it."
Little by little, I've been building up a set of Skills that embody the development values and workflow that I learned early in my career.
I'm working towards one day being able to run an AFK loop on a project and get results I'm excited to review and play with.
I'm working towards one day being able to run an AFK loop on a project and get results I'm excited to review and play with.
January 30, 2026 at 1:00 PM
Little by little, I've been building up a set of Skills that embody the development values and workflow that I learned early in my career.
I'm working towards one day being able to run an AFK loop on a project and get results I'm excited to review and play with.
I'm working towards one day being able to run an AFK loop on a project and get results I'm excited to review and play with.
I think it's going to take a couple years for software development practices as an industry to catch up with where models are TODAY.
January 26, 2026 at 7:39 PM
I think it's going to take a couple years for software development practices as an industry to catch up with where models are TODAY.
Some ideas on how to measure the effectiveness on an agent on a programming team. Or at least how to begin to understand the dynamics:
Agent output tokens v human input tokens per feature
Agent output tokens v human input tokens per feature
January 26, 2026 at 6:02 PM
Some ideas on how to measure the effectiveness on an agent on a programming team. Or at least how to begin to understand the dynamics:
Agent output tokens v human input tokens per feature
Agent output tokens v human input tokens per feature
If well-documented PRs and commits are about showing a programmer's thinking process, attaching the agent threads for the work is a must too.
It's an artifact for the process of creation.
It's an artifact for the process of creation.
January 26, 2026 at 5:03 PM
If well-documented PRs and commits are about showing a programmer's thinking process, attaching the agent threads for the work is a must too.
It's an artifact for the process of creation.
It's an artifact for the process of creation.
Software has always been about defining methods to verify things work.
With agents, this is just as true, but now it's essential.
With agents, this is just as true, but now it's essential.
January 26, 2026 at 3:51 PM
Software has always been about defining methods to verify things work.
With agents, this is just as true, but now it's essential.
With agents, this is just as true, but now it's essential.
Finally setup chezmoi to manage my dotfiles and system config.
I've been putting it off for a while. With an agent, it's as easy as saying, "Set yourself up and suggest things to track"
Now my laptop will finally have the same setup as my desktop. 😅
I've been putting it off for a while. With an agent, it's as easy as saying, "Set yourself up and suggest things to track"
Now my laptop will finally have the same setup as my desktop. 😅
January 26, 2026 at 1:12 PM
Finally setup chezmoi to manage my dotfiles and system config.
I've been putting it off for a while. With an agent, it's as easy as saying, "Set yourself up and suggest things to track"
Now my laptop will finally have the same setup as my desktop. 😅
I've been putting it off for a while. With an agent, it's as easy as saying, "Set yourself up and suggest things to track"
Now my laptop will finally have the same setup as my desktop. 😅
Not sure why this is the case, @AmpCode feels faster to use than Claude Code. It's not just the UI. The agent's inference seems faster.
January 20, 2026 at 7:57 PM
Not sure why this is the case, @AmpCode feels faster to use than Claude Code. It's not just the UI. The agent's inference seems faster.
How did people do devops before AI? 🤯
January 20, 2026 at 5:44 PM
How did people do devops before AI? 🤯
I love the idea of Sprites from Fly, but I'm not sure I need another thing to pay for. Tokens are taking up the "extras" line in the budget.
That said, the idea of having a fully contained instance my agents can run inside of is very attractive.
That said, the idea of having a fully contained instance my agents can run inside of is very attractive.
January 16, 2026 at 8:50 PM
I love the idea of Sprites from Fly, but I'm not sure I need another thing to pay for. Tokens are taking up the "extras" line in the budget.
That said, the idea of having a fully contained instance my agents can run inside of is very attractive.
That said, the idea of having a fully contained instance my agents can run inside of is very attractive.
Vibe coding is akin to a spike or an experiment.
The problem comes when the experiment goes to production without a second look.
Code is now cheap. Generate. Learn. Trash.
The problem comes when the experiment goes to production without a second look.
Code is now cheap. Generate. Learn. Trash.
January 16, 2026 at 5:56 PM
Vibe coding is akin to a spike or an experiment.
The problem comes when the experiment goes to production without a second look.
Code is now cheap. Generate. Learn. Trash.
The problem comes when the experiment goes to production without a second look.
Code is now cheap. Generate. Learn. Trash.
Slop is a byproduct of poor project tooling and unclear thinking.
January 16, 2026 at 4:58 PM
Slop is a byproduct of poor project tooling and unclear thinking.
Deciding how the agent can check their work will become a key piece of my workflow going forward.
What criteria must be true for this feature to be considered complete?
What criteria must be true for this feature to be considered complete?
January 16, 2026 at 1:59 PM
Deciding how the agent can check their work will become a key piece of my workflow going forward.
What criteria must be true for this feature to be considered complete?
What criteria must be true for this feature to be considered complete?
I'm not sure why I didn't do it sooner. Having now asked Claude what to do to succeed, I feel like I can make such better use of agents.
Slop isn't inevitable. Slop is because we've starved the agent what it needs to succeed.
Slop isn't inevitable. Slop is because we've starved the agent what it needs to succeed.
January 16, 2026 at 3:45 AM
I'm not sure why I didn't do it sooner. Having now asked Claude what to do to succeed, I feel like I can make such better use of agents.
Slop isn't inevitable. Slop is because we've starved the agent what it needs to succeed.
Slop isn't inevitable. Slop is because we've starved the agent what it needs to succeed.