Sign in

iiinova is part of the Atmosphere — an open, public network where you own your identity.

Sebastian Hans

@sebhans.hachyderm.io.ap.brid.gy
9 followers 0 following 114 posts

Software architecture and development in Munich, Germany #softwarearchitecture #softwareengineering #softwaredevelopment #fedi22 🌉 bridged from ⁂ hachyderm.io/@sebhans, follow @ap.brid.gy to interact

PostsRepliesMedia
Reposted by Sebastian Hans
Stefano Marinelli @bsd.cafe · 16/10/2026
Ok, looks like this is the week of funny experiences. Customer: "I had an AI check our server and it says the SSH port is open and that's not very secure." Me: "Yes, you asked us to open it so you could access the server via SFTP (actually, you wanted FTP, but we discussed it a bit and found […]
mastodon.bsd.cafe
Original post on mastodon.bsd.cafe
88135 135
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 16/10/2026
Of the others, I particularly liked “Comprehension as an Architectural Characteristic: A System That Is Not Understood Cannot Evolve Safely” by Jacobus Meintjes, Narayana Rengaswamy, Paul Katsande, Sureshbabu Bikki.
000 0
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 16/10/2026
The article I co-authored during the course of the InfoQ Certified Architect program is finally published! 🎉 Read “Fitness Functions and Team Boundaries: Lessons from Evolving a B2C Platform for B2B” by Devora Roth Goldshmidt, Dmitry Kekelidze, Hassan Ejaz, Ifiok Idiang and me along with five […]
hachyderm.io
Original post on hachyderm.io
100 1
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 16/10/2026
Just found this excellent post on how LLMs (maybe) change how we use language: lucumr.pocoo.org/2026/5/4/content-f… I'm surprised “quietly” does not appear in the list of words with inflated usage, but the post is from May 2026, so perhaps it appeared on the scene later […]
hachyderm.io
Original post on hachyderm.io
021 2
Reposted by Sebastian Hans
Jens Ohlig @johl.mastodon.xyz.ap.brid.gy · 12/10/2026
Man könnte natürlich Ransomware-Angriffe auch inhaltlich stellen und sie dadurch entcybern.
0224 24
Reposted by Sebastian Hans
Felienne Hermans @felienne.mastodon.social.ap.brid.gy · 11/10/2026
If we want to regulate AI, why not start with a simple ban on LLMs using "I"? Easy to build, easy to verify and deflates a hype while still allowing people to use chatbots (so no one can claim we are limiting anything) My latest for @volkskrant in English! […]
mastodon.social
Original post on mastodon.social
0014 14
Reposted by Sebastian Hans
Konrad Hinsen @khinsen.scholar.social.ap.brid.gy · 08/10/2026
A good thread about using LLMs for information retrieval and checking the answers.
mas.to
003 3
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 08/10/2026
I can remember a time when the attitude of what I consider good software enngineers was something like: We apply good engineering practices to minimize failure while accepting that failure will sometimes happen, and if it does, we apply equally good engineering practices to deal with it […]
hachyderm.io
Original post on hachyderm.io
120 2
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 08/10/2026
When complaining about bottlenecks, you should maybe also contemplate _why_ bottles have necks.
004 4
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 06/10/2026
In his latest fragments (martinfowler.com/fragments/2026-10-…), @mfowler pulled together some thoughts on the language LLMs should use to program in. I'd add this quote from Dijkstra to the mix: “The purpose of abstracting is not to be vague, but to create a new semantic level in […]
hachyderm.io
Original post on hachyderm.io
010 1
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 06/10/2026
I'm becoming allergic to certain words and phrases. “Keeping the {devs|agent} honest” is such a stupid framing. As if developers were dishonest unless taken on a leash! And it doesn't make sense for “AI” agents either, because they don't have a concept of truth in the first place.
000 0
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 04/10/2026
@quinn The launch doesn't have to be automated. It's enough to hallucinate something to which human decision makers react.
100 1
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 04/10/2026
@quinn Case in point: mastodon.social/@DrALJONES/11737736…
mastodon.social
DrALJONES (@DrALJONES@mastodon.social)
2 Another example of gross incompetence from techbros: "The U.S. almost started World War 3 with China by relying on.. fabricated information from a chatbot." The chatbot gave "fake information saying that a Chinese vessel was carrying components of nuclear weapons." And the US military was ready with planes, soldiers to intercept... board the ship, & this was going to cause an actual existential threat to humanity." ~Timnit Gebru https://www.democracynow.org/2026/10/1/timnit_gebru_ai #USPol #EUPol #AI #China #ww3 .
100 1
Reposted by Sebastian Hans
MrAlanCooper @mralancooper.hachyderm.io.ap.brid.gy · 30/09/2026
I still remember the day in 1976 when Gary Kildall typed in "dir" to an "A>" prompt on his hazeltine green-screen character display to show me a directory listing on his 8080-based, eight-bit computer running the first release of CP/M, the first real microcomputer operating system (that Gary […]
hachyderm.io
Original post on hachyderm.io
2539 39
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 01/10/2026
It is possible to implement a simple version of Pong using git commit hooks while waiting for an IT provisioning process. Ask me how I know. :-/
000 0
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 01/10/2026
@quinn Eh, I think “AI” may well kill us, but not in the way “AI” doomers think. It will be its tendency to accelerate what's already there that does it. People have been experimenting with dangerous viruses before “AI”, but now they can do so even faster and produce worse accidents (going […]
hachyderm.io
Original post on hachyderm.io
101 1
Reposted by Sebastian Hans
tante @tante.tldr.nettime.org.ap.brid.gy · 24/09/2026
I had a few thoughts on why techbros always say these embarrassing, weird things tante.cc/2026/09/24/the-test
tante.cc
The Test
Jensen Huang, the NVIDIA CEO and wearer of the most midlife crisis looking leather jackets in the world, was on Ezra Klein’s podcast. Now I don’t really want you do go through all of what was being said in that recording/writeup. It’s painful. You feel your mental capabilities decaying with every minute of listening … […]
3234 34
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 29/09/2026
Plans that absorb uncertainty tend to be more robust than plans that try to handle uncertainty.
000 0
Reposted by Sebastian Hans
jonny (nonvenomous) @jonny.neuromatch.social.ap.brid.gy · 29/09/2026
RE: neuromatch.social/@jonny/1173398259… OK! Meta evaluated this as intended behavior, not applicable for a bug bounty, so therefore responsible disclosure no longer applies so here goes: any process run within the VM can access the socket that provides inference with no […]
neuromatch.social
Original post on neuromatch.social
1632 32
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 29/09/2026
“There's no point at which turning your brain off will work” danluu.com/brain-off I fully agree with this. I've always thought of myself as a brain-on developer. #softwareDevelopment #softwareEngineering #programming
danluu.com
There's no point at which turning your brain off will work
Comments
010 1
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 28/09/2026
New blog post on understanding how software works. I posted this here previously, just putting it up there for ease of reference. sebastian-hans.de/blog/understandin… #softwareDevelopment #softwareEngineering #agentic #AI #programming
sebastian-hans.de
Understanding how software works — Sebastian's blog
Sebastian's blog – Understanding how software works
020 2
Reposted by Sebastian Hans
Glyph @glyph.im · 27/09/2026
The software industry's original sin was "no warranty, even the implied warranty of merchantability", even on literal products for sale. Every evil we have visited upon the world — clickwrap licenses, instagram anorexia, AI dystopia — is downstream of the generalized ability for software to […]
mastodon.social
Original post on mastodon.social
1328 28
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 28/09/2026
DIeses Paper erklärt u.a., warum „die AfD inhaltlich stellen“ nicht funktioniert. “When this assumption does not hold—when conflict is corrosive and thus reflects domination, exclusion, or the strategic erosion of democratic norms—calls for empathy, reconciliation, or civility risk reproducing […]
hachyderm.io
Original post on hachyderm.io
030 3
Reposted by Sebastian Hans
David Roberts @volts.wtf · 27/09/2026
I love this paper so much. If I had my way every single journalist, pundit, analyst, and politician in the US would read it & then read it again. I want to quote all of it. www.frontiersin.org/journals/soc...
frontiersin.org
Frontiers | The problem with polarization research in an age of extremism
"I have been gravely disappointed with the white moderate. I have almost reached the regrettable conclusion that the Negro's great stumbling block i...
14643193 643
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 24/09/2026
If I have a problem with my Java application that I supect may be caused by the Java compiler, I may not be able to solve it myself today, but I can look at the code and come to understand it, or I can ask a JVM developer to take a look and we can solve it together (or pay someone to do it) […]
hachyderm.io
Original post on hachyderm.io
000 0
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 24/09/2026
On understanding how software works: “But Sebastian, how can you claim we are losing our understanding of how software works when you are building on top of layers and layers of software, firmware and hardware yourself? I bet you don't understand it all the way down, either!” That's right, but […]
hachyderm.io
Original post on hachyderm.io
100 1
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 24/09/2026
RE: neuromatch.social/@jonny/1173247908… Digging in the Matrix, finding nested Bash and Python and reeeeally creepy instructions. 🤪 It would be funny if it wasn't so sad.
000 0
Reposted by Sebastian Hans
Benjamin @blindcoder.toot.berlin.ap.brid.gy · 24/09/2026
Scrapping features noone ever used feels so good! #SoftwareDevelopment
001 1
Reposted by Sebastian Hans
daniel:// stenberg:// @bagder.mastodon.social.ap.brid.gy · 23/09/2026
Dan Lorenc at Chainguard posted this statement about #curl on LinkedIn...
We put curl 8.22.0 through an eight-and-a-half-hour automated tree search, about 4,100 tool calls and roughly 190 million tokens of model work (the vast majority cache reads), aimed only at High and Critical outcomes: memory safety on an ASan build across the HTTP/1, HTTP/2, HTTP/3, WebSocket, FTP, IMAP, POP3, SMTP, SMB, LDAP, cookie, and NTLM/Digest/Negotiate parsers under tens of thousands of hostile server responses, plus integrity hunts for TLS verification bypass under an on-path attacker,
STARTTLS and 1implicit-TLS downgrade, cross-handle response confusion, CONNECT tunnel handling, credential leakage through redirects, file writes outside the output directory, and Alt-Svc and HSTS guarantees. Thirty-one lines of attack closed clean with no sanitizer report and no verification
bypass; all that surfaced were three Medium-class edges, one of which the curl team had already fixed on master days before we looked. That outcome is a credit to the project: years of continuous fuzzing, the maintainers' own audits, and a codebase where every server-controlled length is bounded
before use left a persistent automated adversary with nowhere to go. Congratulations to the curl team, and thanks for making our negative result so thoroughly earned.

One note on the token figure: the run log records usage for only about 400 of the turns (18.6 million tokens, 46,500 per turn on average), so the 190 million is that average scaled to the 4,103 tool-call turns.
21714 17
Reposted by Sebastian Hans
tante @tante.tldr.nettime.org.ap.brid.gy · 22/09/2026
Also als die Leute anfingen, die Faschisten der AfD zu wählen, wurden die bürgerlichen Medien nicht müde, jedem Neurechten ein Mikro ins Gesicht zu halten, damit man seine "Sorgen und Nöte" verstehen kann. Das passiert jetzt doch sicher auch mit den Wähler*innen linker Parteien, oder? Oder ging […]
tldr.nettime.org
Original post on tldr.nettime.org
0322 22
Reposted by Sebastian Hans
Tomáš @prahou.merveilles.town.ap.brid.gy · 22/09/2026
@tante normalize using finished software
0110 10
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 21/09/2026
Just went out of my way to help a colleague by answering a series of questions they posted to our internal communication channel and offered a call for further help only to be told they had generated the message from a rough sketch and these weren't concrete questions yet. From now on, I'll view […]
hachyderm.io
Original post on hachyderm.io
000 0
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 21/09/2026
I highly recommend this talk by @kelseyhightower: www.youtube.com/watch?v=A7WFt2JQ5sg “So if you think context is amazing for agents, wait till you give it to humans.” Spot on. Be sure to watch the whole thing, including the Q&A! #softwareArchitecture #softwareEngineering […]
hachyderm.io
Original post on hachyderm.io
001 1
Reposted by Sebastian Hans
tante @tante.tldr.nettime.org.ap.brid.gy · 16/09/2026
I don't want to give the article too much shit because moving away from big tech is important and all. But this - from someone running services - is a bad sign. Instead of checking out the docs of the software you migrate to if it has a direct migration path […] [Original post on tldr.nettime.org]
Repositories, users, secrets, actions were super simple to move from GitHub to Forgejo. I did it manually with the help of Claude Code before realising Forgejo had a built-in migration tool, which I suspect would have probably been faster and used no tokens :D
1112 12
Reposted by Sebastian Hans
myrmepropagandist @futurebird.sauropods.win.ap.brid.gy · 16/09/2026
I have another brilliant idea to save money and time. Instead of promoting an LLM to write an email this new system will simply send your target the prompt and the email will be generated on their end with *their* tokens should they choose to open it. We can hide this all in an embedded HTML […]
sauropods.win
Original post on sauropods.win
1344 44
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 16/09/2026
RE: hachyderm.io/@hazelweakly/117275415… This is an interesting thread! The following steps are fun, too: 11. “here’s how the auditor interprets that obligation” 12. “here’s how the auditor interprets the outcome guidelines and/or implementation guidelines given” 13. “Here’s […]
hachyderm.io
Original post on hachyderm.io
000 0
Reposted by Sebastian Hans
OpenStreetMap Ops Team @operations.openstreetmap.org · 15/09/2026
The OpenStreetMap.org website and API are struggling. We are fighting a large set of AI crawlers / scrapers. Working to restore service as soon as possible. #AI #Scrapers
122 2
Reposted by Sebastian Hans
tante @tante.tldr.nettime.org.ap.brid.gy · 15/09/2026
RE: mastodon.social/@rb3n/1172741941514… Heise translated my Comment on the "Concerned AI Bros" into English www.heise.de/en/opinion/Comment-Oh-…
mastodon.social
0111 11
Reposted by Sebastian Hans
Christoffer S. @cstromblad.com · 15/09/2026
On the topic of LLM-generated software there is one thing I've only grown more concerned and confident about - the need for systems design and architecture thinking. From what I've experienced, both anecdotally and first-hand, is that LLM-generated software systems eventually devolve, and when […]
swecyb.com
Original post on swecyb.com
111 1
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 14/09/2026
@kevlin @softwarearchitektur_im_stream Unfortunately, it made me realize that the last two projects I worked on were driven almost entirely by failure demand. (Not that I didn't already know this, but it still feels different when its put in words that way.)
000 0
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 14/09/2026
Just listened to @kevlin on Software-Architektur im Stream (@softwarearchitektur_im_stream). I particularly liked the discussion around value demand vs. failure demand and what consequences accelerating work on one or the other might have. I highly recommend this episode […]
hachyderm.io
Original post on hachyderm.io
210 2
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 14/09/2026
Just had a discussion about a cloud deployment. I was investigating something and found subnets that were apparently not connected to anything. Further investigation revealed that these subnets existed in three availability zones, but only in one of them something was actually running, and […]
hachyderm.io
Original post on hachyderm.io
000 0
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 11/09/2026
That's certainly something to think about.
001 1
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 08/09/2026
@moira This is a whole different level than I imagined, but boy am I glad I refused to hook our “smart” TV to the network!
000 0
Reposted by Sebastian Hans
Solarbird :flag_cascadia: @moira.mastodon.murkworks.net.ap.brid.gy · 07/09/2026
This is the big LG television surveillance documentary story GN have been working on for a couple of months, and that they previewed a few days ago. The complete pervasiveness of spying upon not just you and what you watch but _every person and every device around every set_ is genuinely […]
mastodon.murkworks.net
Original post on mastodon.murkworks.net
4070 70
Reposted by Sebastian Hans
Chris Dwan (he/him) @somershade.bsky.social · 03/09/2026
A friend refers to the "a pedestrian was struck by a car" construction as the "exonerative" tense, and that shall be its name henceforth and hereafter.
18379154 379
Sebastian Hans @sebhans.hachyderm.io.ap.brid.gy · 08/09/2026
Habe mich mal wieder daran erinnert … leider. www.youtube.com/watch?v=ebhxL_duXJo
000 0