Thomas Letan
@lthms.soap.coffee
📤 152
📥 108
📝 390
“Decent solutions provider.” He/him.
https://soap.coffee/~lthms
pinned post!
I’ve written down my thoughts about what I want to do in 2026 wrt. LLMs
loading . . .
How I Want to Use LLMs in 2026
Agentic tools are here, and they are here to stay. As a consequence, I will integrate LLMs in my workflow, but I want to do it on my own terms.
https://soap.coffee/~lthms/posts/how-i-want-to-use-llms-in-2026.html
4 months ago
0
3
1
reposted by
Thomas Letan
🦌≈
about 1 month ago
the first rule of compilers is to have fun
3
149
25
Yesterday I wrote a MR like if it was still 2025
about 1 month ago
1
0
0
You really cannot trust Claude is my personal first quarter conclusion. Doesn’t mean you can’t get it to produce meaningful stuff. Just, if it doesn’t want to do it your way, it won’t 😂
about 2 months ago
0
0
0
that's the time of the month where i am supposed to write a retrospective. The thing is, i just worked this month 😆
about 2 months ago
0
0
0
reposted by
Thomas Letan
Aleix Morgadas
about 2 months ago
🧠 Rethinking Pair Programming in the Age of AI Why "pairing with AI" is harmful to your team and what to do instead
learnings.aleixmorgadas.dev/p/rethinking...
loading . . .
Rethinking Pair Programming in the Age of AI
Why "pairing with AI" is harmful to your team and what to do instead
https://learnings.aleixmorgadas.dev/p/rethinking-pair-programming-in-the
2
7
4
reposted by
Thomas Letan
Obsidian
about 2 months ago
The Obsidian team is growing from three engineers to four engineers. Competitive SF salary. Fully remote, live anywhere. Apply below.
4
317
58
reposted by
Thomas Letan
srrrse
about 2 months ago
This resonates
mariozechner.at/posts/2026-0...
loading . . .
Thoughts on slowing the fuck down
Thoughts on slowing the fuck down
https://mariozechner.at/posts/2026-03-25-thoughts-on-slowing-the-fuck-down/
1
18
4
reposted by
Thomas Letan
Gergely Orosz
about 2 months ago
If you use GitHub (especially if you pay for it!!) consider doing this *immediately* Settings -> Privacy -> Disallow GitHub to train their models on your code. GitHub opted *everyone* into training. No matter if you pay for the service (like I do). WTH
github.com/settings/cop...
90
2054
1633
The zen of using the `latest` tag of a docker image with auto upgrade enabled
2 months ago
0
0
0
I haven’t read about SRAM since my PhD 😅
add a skeleton here at some point
3 months ago
0
1
0
My little cluster is getting into shape. - miniflux for RSS - continuwuity + n8n for automation (not used yet) - soju for IRC - garage for easy hosting of static files
3 months ago
1
1
0
So XMPP is a little weird. I wanted to use it has a n8n control plane of sorts but I think Matrix will be a better fit
3 months ago
0
0
0
Claude Code does not deal with frustration very well.
3 months ago
0
0
0
reposted by
Thomas Letan
Chris Short
3 months ago
Wrapping Code Comments
#devopsish
loading . . .
Wrapping Code Comments
I was today years old when I realized that:
https://matklad.github.io/2026/02/21/wrapping-code-comments.html?&aid=recXR8yvjkAlgpUhe&_bhlid=eba530744267af66ea0840d3b38742fa20034b5a
0
0
1
I really don’t like the pets vs cattle DevOps Analogy It’s more like pottery vs 3D printing
add a skeleton here at some point
3 months ago
0
1
0
The bliss of watching all the VMs running my k3s cluster being destroyed and recreated, and everything comes back to life.
3 months ago
0
0
1
reposted by
Thomas Letan
Xe
3 months ago
1
283
61
reposted by
Thomas Letan
Rami Ismail (رامي)
3 months ago
oh no they scraped your stuff without permission oh no
add a skeleton here at some point
13
616
106
What can go wrong?
3 months ago
0
0
0
reposted by
Thomas Letan
Richard Kadrey
3 months ago
2026 basically
304
40966
11248
i'm back on IRC 😎
3 months ago
0
0
0
reposted by
Thomas Letan
Tim Kellogg
3 months ago
skill issue
4
165
11
reposted by
Thomas Letan
amos @ home
3 months ago
just fixed tracey hover — so you can see the rule you're implementing just by hovering it
github.com/bearcove/tra...
5
49
3
My favorite time of a hobby project.
3 months ago
1
0
0
I have deployed a miniflux instance tonight. If you are looking for a RSS reader and don’t mind it’s not professional grade SLA feel free to reach out
3 months ago
0
0
0
Claude being like “your idea is nice but for your *scale* that’s no use, let’s go back to *my* idea.” I mean, excuse me? you’ll spend the next 10 minutes exploring my idea is what will happen.
3 months ago
1
0
0
I can now redeploy my cluster control plane without my website to go down 🎉.
3 months ago
0
0
0
reposted by
Thomas Letan
Leo 🏴☠️
3 months ago
hi folks! long time no see. i’m officially open to work again, primarily interested in the dx + devtools + ai space if anyone needs a principal engineer or technical product manager with 16+ years in the industry, my DMs are open!
linkedin.com/in/leostera
github.com/leostera
loading . . .
https://linkedin.com/in/leostera
1
20
9
One post to go and no more 2024 entries on my homepage. Three and the same is true for 2025 😁. I have two planned already. We’ll see when the third lands!
3 months ago
1
0
0
Blogged something out of frustration today 😅
loading . . .
Teaching Firefox About slack:// URLs
There is nothing like finally taking the time to fix an annoyance that’s been puzzling you for 5 years.
https://soap.coffee/~lthms/posts/teaching-firefox-about-slack.html
3 months ago
0
1
1
So, for real: when does the rain stop in France??
3 months ago
0
0
0
This week-end I learned BetterStack has a Terraform provider and I think it is beautiful.
loading . . .
lthms’ cloud lab status
Welcome to lthms’ cloud lab status page for real-time and historical data on system performance.
https://lthms-cloud-lab.betteruptime.com/
3 months ago
0
0
0
So debugging with Claude reminds me a lot of that feelings of having to watch someone else doing all the keystrokes via a Meet room. In my case we were debugging a k3s cluster issue and it understands that a lot better than me so it makes sense that I take the copilot seat.
3 months ago
1
1
0
“Good question.” Claude said. And the user knew: all the reassuring comments the agent made in the past few minutes were probably worthless.
3 months ago
0
2
0
Having Claude debugging a kubernetes issue on its own is mesmerizing
3 months ago
0
2
0
My little k3s setup is coming along nicely. Two VMs, journal logs and hardware metrics pushed to BetterStack. Next step is actually deploy something there 😄.
3 months ago
1
0
0
Unsurprisingly, I’m not the first one noticing the similarities:
tante.cc/2026/01/15/s...
add a skeleton here at some point
3 months ago
0
0
0
After the fast food and fast fashion industry, are we about to see the emergence of the fast software industry?
3 months ago
0
0
1
holy moly my cluster boots now.
3 months ago
0
1
0
gotta say, deploying a k3s cluster on Vultr + CoreOS + Ignition is harder than I thought.
3 months ago
0
0
0
I’m all in with the FedoraOS + Ignition way of deploy things™ but oh my god how annoying it is to get it right at first. I need to stop creating and destroying VMs that are costing me 2cts each time and try on local VMs 😂
4 months ago
0
0
0
reposted by
Thomas Letan
maxrn
4 months ago
the world if "ps | grep xyz" didn't show "grep xyz" in the output
1
1
1
Week-end challenge: setup a simple k3s cluster.
4 months ago
0
0
1
I’m resuming trying to use Claude Code in a way that would fit an engineering process. I’m using an issue to test the command that I’m implementing:
github.com/lthms/vee/is...
I like it because it’s not a trivial one. Like, I really wasn’t sure about the best way to implement this.
loading . . .
Docker Compose support for ephemeral sessions · Issue #8 · lthms/vee
Problem Ephemeral sessions only support single-container environments, preventing use with projects that require supporting services (databases, queues, mock APIs). Context Some projects need multi...
https://github.com/lthms/vee/issues/8
4 months ago
1
0
0
People will tell you being polite to a code assistant is just a waste of tokens. Meanwhile, Claude sometimes starts its prompts to subagents with “Please.”
4 months ago
0
0
0
So upgrading once’s firmware does help restoring sound on Linux
4 months ago
0
0
0
reposted by
Thomas Letan
Hoshino Lina (星乃リナ) 🩵 3D Yuri Wedding 2026!!!
4 months ago
Specifically, the kind of person who, despite being largely on "the same side" as you on the broader issues, will come out knives swinging at the smallest slight, real or perceived, and then deflect and blame-shift and put words in your mouth to justify it. I don't need people like that in my life.
1
22
2
I can now start new Claude instances inside Docker containers with a few keystrokes
4 months ago
0
0
0
I’m getting bold with my “I don’t like Claude Code vanilla TUI’s UX, let's vibe code something better.” I’m still looking for the
#$WORK
formula, but for personal projects, I don’t see a way back to coding myself, and that makes me nostalgic.
4 months ago
1
0
0
reposted by
Thomas Letan
Gergely Orosz
4 months ago
Even though AI agents make me more productive: when I use them, I end up working more, not less! It’s such a contradiction: a thing that should be saving time ends up taking away more overall time… A reason I find it hard to believe AI will lead to fewer hours/days worked…
16
104
7
Load more
feeds!
log in