HN user

atmosx

6,593 karma

SRE / infra at <startup>.

blog: https://www.convalesco.org/

NOTE: The opinions expressed on HN are my own and do not represent my employer's views in any way. Nothing posted here should be considered official or sanctioned by my employers or any other organisation I’m affiliated with.

[ my public key: https://keybase.io/atmosx; my proof: https://keybase.io/atmosx/sigs/NpgSmczuTq4qXkimbkMPxQXEYfPFV64joiMwMjJDw_w ]

meet.hn/city/gr-Drama

Socials:

- github.com/atmosx

- libretooth.gr/@atmosx

Interests:

DevOps, Philosophy, Geopolitics, Remote Work, Startups, Art, Snowboarding

---

Posts130
Comments3,646
View on HN
www.theguardian.com 2mo ago

Sensor tampering to win weather bets

atmosx
1pts0
lindseyhallwrites.substack.com 3mo ago

I Stumbled Across My Boyfriend's ChatGPT and It Ended Our Relationship

atmosx
13pts7
www.psychologytoday.com 4mo ago

When Writing Becomes Detached from Thought

atmosx
4pts0
www.nature.com 7mo ago

Unexpected events and prosocial behavior: the Batman effect

atmosx
4pts1
www.theatlantic.com 9mo ago

If the University of Chicago won't defend the humanities, who will?

atmosx
166pts179
www.brunellocucinelli.com 10mo ago

Technology, Humanism and Artificial Intelligence

atmosx
1pts0
pubmed.ncbi.nlm.nih.gov 1y ago

Treating wounds in pediatrics with medical grade honey: A case series

atmosx
7pts2
www.climatiq.io 2y ago

Measuring gas emissions in data centres (2022)

atmosx
1pts0
bearmccreary.com 2y ago

Scoring "The Rings of Power" P1 – Journey to Middle-Earth

atmosx
2pts0
dailyphilosophy.substack.com 2y ago

Who Controls Technology? (Daily Philosophy Newsletter #291)

atmosx
2pts0
www.wired.com 2y ago

Greece's New Political Star Is a TikTok Creation

atmosx
1pts1
www.the-express.com 2y ago

iPhone users warned using wrong USB-C cable could destroy device for good

atmosx
2pts2
www.spinellis.gr 3y ago

The hypocritical call to pause giant AI

atmosx
2pts0
www.keyboardsettlement.com 3y ago

MacBook Butterfly Keyboard Litigation Settlement Website

atmosx
16pts6
www.ft.com 3y ago

How social media platforms put profits before people

atmosx
2pts0
twitter.com 4y ago

Homeland Security sends warning letter due bad tweet language

atmosx
49pts44
kr-asia.com 5y ago

China’s tech workers pushed to limits by surveillance software

atmosx
3pts0
en.wikipedia.org 5y ago

FBI multinational sting operation via compromised network (16 countries)

atmosx
3pts0
www.androidpolice.com 5y ago

The EU may be looking into breaking end-to-end encrypted chats

atmosx
3pts0
www.catharsismagazine.com 5y ago

What It Means to Be a Hero: American Superhero Comics vs. Japanese Shonen Manga

atmosx
2pts0
twitter.com 6y ago

DHH about startups “changing the world” and workaholics

atmosx
17pts1
www.youtube.com 7y ago

Linguistics, Style and Writing in the 21st Century (2015)

atmosx
3pts0
www.nytimes.com 7y ago

Will Sports Betting Transform How Games Are Watched, and Even Played?

atmosx
4pts0
www.forbes.com 7y ago

Can Uber Be Beaten 1 City at a Time? Daimler's Bet

atmosx
1pts0
www.newscientist.com 7y ago

Famed mathematician claims proof of 160-year-old Riemann hypothesis

atmosx
2pts1
www.youtube.com 8y ago

Steven Pinker on Good Writing, with Ian McEwan (2014)

atmosx
2pts0
list.zfsonlinux.org 8y ago

SQLite Corruption when running on ZFS (Linux and FreeBSD)

atmosx
3pts0
github.com 8y ago

Awesome List of resources on leading people and being a manage

atmosx
2pts0
blog.acolyer.org 8y ago

CLKSCREW: Exposing the perils of security-oblivious energy management

atmosx
2pts1
www.youtube.com 9y ago

The Philosophy of Stoicism – Massimo Pigliucci

atmosx
2pts0

No one should be afraid of anything. Fear is a terrible advisor. Keep your eyes open, try to read the context as careful as you can and adapt as best as you can. Don’t spent too much time trying to be an oracle, never works out…

If you study the last 5k years, then the real question how on earth did the Europeans (primarily) and then the Americans managed to stay ahead of China for as long as they did (the "Needham Question").

Confucius came before Plato and most important discoveries came from China. The two wrote a the "good citizen" ought to be and how a city/country/empire ought to be run. Even meritocracy has been traditionally stronger in China, over centuries.

In the west we're talking about it since Homer, but in practice has been applied sporadically, while China has been much more consistent.

I would put all the “nuke is safe” ppl in an island alongside the reactor and go back every 100 years to check if they went Atlantis or not yet…

I get that it’s a a form of clean energy and we kind of need that in the EU but a remote, “protected”, isolated location would be my first choice.

Not familiar with blink but I use tmux heavily. Tried herdr but did t stuck. The main selling point is the status report: tmux requires a plugin and even that doesn’t work as good as herdr. By reporting I mean that while a tmux plugin can report a very basic “status” in the powerline, herdr does so natively in greater detail (reports if the agent waits for use input etc). All solutions I came across with status report might support one agent / cli app and not others… so there is that.

Coupled with tmux and tmuxinator. Vim in the editor pane1, shell in pane2, agent(s) in panes3+. I rename the pane’s name to map try work done there. I rarely work with more than three agents, most of the it’s just one because I spent time reviewing the changes, performing tests in the shell pane to make sure everything is as it should be.

Ps. I use an the vim-ai plugin (dot files available in GitHub alongside vim config) but it’s not ideal for agetic workflows.

The Damon should default to local host, however mariaDB is not some obscure daemon that runs for the kicks… so maybe, engineers should take the time to setup a firewall instead of assuming that all daemons listen to local host? Especially when the server is online? Just saying…

The text is far from perfect. It’s shallow and most of the times average. LLM is okay-ish for a first draft, maybe, but any professional writer will tell that it’s a rather poor author.

None that should matter to you. Maybe you should start with something cheaper instead of SOTA e.g. Deepseek, Mistral, etc. They're not as good but not as bad as some say either. These cheaper LLMs will allow you to experiment much more until you get some sort of workflow on using LLMs and the switch to the expensive ones.

Well they're not rounding people because of their religion or sexuality and putting them in "retraining" camps yet.

Right. They pay Turkey to do that: https://www.rescue.org/eu/article/what-eu-turkey-deal

I don’t think the EU is a classic dictatorship, but it’s a colossal failure nonetheless, has a severe lack of democracy and acceptance. And their personnel is mediocre, not like the US administration but it’s closer than ppl in this forum realize.

That couldn't be further from the truth.

All the referendums I can recall (French and Dutch in 00s, Italian recently, Greek and British) were all resounding “No”’s… plus the rise of the far right is kinda hard to miss…

Gallup, March 2025: No country approves of their own leadership, or that of U.S., over the EU [2]

That’s literally insane. I don’t believe it. I have yet to find someone who is fond of the EU leadership. lol. …

You have to remember that the EU is fundamentally the world's most successful peace project, founded after the horrors of WW2 to make war in Europe not just unthinkable, but materially impossible.

You're confusing the EU with NATO. Germany is already planning on re-arming[^1] because NATO is dead. The EU is by all means an unfinished project and if I had to bet, I would bet that it won't survive the next two decades. There's no pro-EU majority in any European country right now. Not one. It's a great idea, very poorly implement that gets (some times unfairly, some times fairly) all the blame for everything wrong with Europe right now. Migration, energy, housing, etc.

[1^]: https://www.defensenews.com/global/europe/2026/04/22/germany...

This one, on the other hand, seems to be genuinely boring. I couldn't get past the intro.

The professionals…

On a related note reading HN comments is a prime example of sleepy text. Gets me every time.

They’ll train on your uploaded docs and photos, to be sure, but it won’t be correlated to your chats and projects, unless Anthropic is doing things that would make their counsel have heart attacks.

And if they do, they'll apologize with a blog post.

Are you self-hosting code in Forjego by any chance? If yes, there's an integrated system called "actions"[1].

Jenkins is popular but complex. I would look elsewhere, woodpeckerCI looks like a good solution but I don't have direct experience with it.

I've worked with Tekton in the past, but it's tailored around kubernetes, e.g. a good for k3s (self-hosted k8s cluster).

If you're not self-hosting and you use something like github, you need to either leave an HTTP server open to the internet (insecure) or use something like ngrok, smee.io, cloudflare tunnel, hookdeck, etc. These tools will manage to forward the webhook in a secure way to the internal system.

[1] https://forgejo.org/docs/latest/admin/actions/