Seán Fobbe

Human Rights Lawyer ⚖️ Legal Data Scientist 📈 Open Data Advocate

My research specialties are international law, human rights law, international criminal law, natural language processing of legal text and Open Legal Data. I've also published quite a bit on cultural heritage and am the lead author of the 'Destroying the Soul of the Yazidis' report.

I love #rstats and have recently gotten into #bayesian #stats with #stan

🇺🇦 #StandWithUkraine

#HumanRights #IntLaw #OpenData #OpenSource

Academic Website: https://seanfobbe.com
Business Website: https://fobbe-consulting.com
Codeberg: https://codeberg.org/seanfobbe
ORCID: https://orcid.org/0000-0003-3808-7730

Public Key

npub1jtds8xus4y05wfhujfrzhjqmw8n6n8v85u8nf66amw23ur5m3r7sg3epp5

NIP-05 Address

[email protected]

Profile Code

nprofile1qqsf9kcrnwg2j868ym7fy33teqdhreafnkr6wre5adwah9g7p6dc3lgprpmhxue69uhhyetvv9ujumt0d4hhxarj9ecxjmntzpd08v

Publishing to

relay.momostr.pink

Author Public Key

npub1jtds8xus4y05wfhujfrzhjqmw8n6n8v85u8nf66amw23ur5m3r7sg3epp5

Show more details

Published at

2026-11-02T16:02:56Z

Kind type

0 Metadata

Event JSON

{ "id": "a002b6def15fff1fb1802b3ce1822714cd66611aab19539d48f378c6c7d32a12", "pubkey": "92db039b90a91f4726fc92462bc81b71e7a99d87a70f34eb5ddb951e0e9b88fd", "created_at": 1793635376, "kind": 0, "tags": [ [ "t", "opendata" ], [ "t", "opensource" ], [ "t", "rstats" ], [ "t", "stats" ], [ "t", "humanrights" ], [ "t", "bayesian" ], [ "t", "Stan" ], [ "t", "StandWithUkraine" ], [ "t", "intlaw" ], [ "proxy", "https://fediscience.org/users/seanfobbe", "activitypub" ], [ "L", "pink.momostr" ], [ "l", "pink.momostr.activitypub:https://fediscience.org/users/seanfobbe", "pink.momostr" ], [ "-" ] ], "content": "{\"name\":\"Seán Fobbe\",\"about\":\"Human Rights Lawyer ⚖️ Legal Data Scientist 📈 Open Data Advocate\\n\\nMy research specialties are international law, human rights law, international criminal law, natural language processing of legal text and Open Legal Data. I've also published quite a bit on cultural heritage and am the lead author of the 'Destroying the Soul of the Yazidis' report.\\n\\nI love #rstats and have recently gotten into #bayesian #stats with #stan \\n\\n🇺🇦 #StandWithUkraine\\n\\n#HumanRights #IntLaw #OpenData #OpenSource\\n\\nAcademic Website: https://seanfobbe.com\\nBusiness Website: https://fobbe-consulting.com\\nCodeberg: https://codeberg.org/seanfobbe\\nORCID: https://orcid.org/0000-0003-3808-7730\\n\",\"website\":\"https://fediscience.org/@seanfobbe\",\"picture\":\"https://fediscience.org/system/accounts/avatars/109/286/130/462/412/714/original/c263246b7551a94e.png\",\"banner\":\"https://fediscience.org/system/accounts/headers/109/286/130/462/412/714/original/e384d07493ac5c19.jpeg\",\"nip05\":\"[email protected]\"}", "sig": "c4d60ffd135f6883e307a7a9407db6445bd4ed706e75049d01595c19e1d95d6b4ab4a31950b1468a0d776d749d648b347a1dfa11a6287ee63f4b7757ea159eb6" }

Last Notes

2026-09-15 16:24:59 GMT

Zenodo is buckling under the strain of AI bot traffic, the Open Science Framework (OSF) has gone into emergency maintenance mode due to AI crawlers and other high-value sites across the web are desperately erecting barriers to ward off traffic overload.
I've been wondering what "AI" progress will look like once the open information ecosystem it has been built on is destroyed. It looks like we will find out in the not so far away future.

2026-08-09 19:25:45 GMT

"Dark Hours" is probably one of the first public smoking gun cases of a complete app (including the name) being copy/pasted by Claude Code and getting the "AI-enhanced" developer into trouble.
Godier did a good job in the aftermath by taking the app down and apologizing, but people really need to understand that "AI code" comes from somewhere and that somewhere is usually someone else's hard, copyrighted work.
https://blog.terrygodier.com/2026/08/09/mea-culpa-dark-hours.html
#DarkHours #AI #SmokingGun #Copyright

2026-05-19 08:54:28 GMT

We somehow went from "script kiddies are bad" to giving any random office worker the power to launch hundreds of programs that will hammer unknown servers across the web to make mediocre power point presentations.
Not to mention the massive usage spikes that are now hitting public software and data repositories.
What a world.

2026-04-24 13:04:55 GMT

AI economics seem less and less favorable: https://www.theverge.com/ai-artificial-intelligence/917380/ai-monetization-anthropic-openai-token-economics-revenue
Crash coming closer, I guess. Perhaps RAM will be affordable again next year?

2026-02-16 12:29:56 GMT

AI companies copy all written works they can get their hands on and call it fair use, if someone does it to their models it suddenly becomes "unauthorized distillation" and should be actionable in court.
The double-standard is ridiculous.
https://www.theregister.com/2026/02/14/ai_risk_distillation_attacks/

2026-02-04 12:55:08 GMT

The problem with "AI" in legal --- all ethical problems aside --- is that legal is heavily reliant on language in addition to logic. Determining the "correct" use of legal language is a difficult and ill-posed problem with multiple viable solutions and no means to safely choose between them without authoritative case law coming to the rescue.
Verifying legal solutions is a problem with a high human verification burden, so producing mass "AI" solutions mostly increases human workload, instead of reducing it.
#Legal #AI #Drafting #CaseLaw

2025-12-29 21:52:56 GMT

Wow, Linux gaming is really close to being ready for prime time.
I just managed to run Final Fantasy VII Remake on Debian 13 (Trixie) via Steam out of the box, no special adjustments necessary. Skyrim and Tales of Arise worked just as well.
I've got to test this more, but in some ways they seemed to run even better (i.e. faster) than on Windows.
#LinuxGaming

2025-10-01 15:29:16 GMT

I'm assuming Linkedin usage is in free fall. Impressions have been poor for a while, but the switch from daily to cumulative impression counting in the default Dashboard is serious evidence that Linkedin is seeing this across the board and is trying to actively hide it.
Why is cumulative impression counting a dishonest use of statistics in this particular case? "Cumulative" means that impressions counts for every day are added to the previous day. This implies:
1) The absolute numbers always go up
2) The slope of the line is always positive (i.e. points upwards)
3) The y-axis labels show larger numbers to accommodate the summation (i.e. 100-200 daily impressions are plotted on 100/200/300 y-axis labels, the cumulative numbers suddenly need a 500/1000/1500 y-axis)
Cumulative sums can be appropriate in certain contexts, but not this one.
https://fediscience.org/system/media_attachments/files/115/299/615/929/803/238/original/e8b033358eec0ec0.png

2025-04-11 08:25:33 GMT

🔔 New Pre-Print 🔔
We introduce the "Corpus of Resolutions: UN Security Council (CR-UNSC)", a novel international legal-political dataset containing texts and metadata for all resolutions of the United Nations Security Council (UNSC) from resolution 1 (1946) through resolution 2722 (2024), as published by the UN Digital Library (UNDL).
Pre-Print: https://zenodo.org/records/15191558/files/Fobbe-Gasbarri-Ridi_2025_Words-of-Power-Introducing-UN-Security-Council-Resolutions-Corpus.pdf?download=1
The dataset provides full resolution texts in all six official UN languages, draft texts and full meeting records in English, as well as dozens of metadata variables, for a grand total of 82 variables in structured tabular CSV format, accompanied by extensive narrative documentation. In addition to the main tabular dataset the CR-UNSC is accompanied by specialized variants for network analysis and bibliography management.
The computational workflow is engineered as a fully automated end-to-end extract-transform-load (ETL) data pipeline with citation analysis and NLP components, unit tests and extensive reporting. The declarative workflow is fault-tolerant, resumable and stores intermediate results in over one hundred individual checkpoints. The code is published open source under the GNU General Public License Version 3 (GPLv3).
We intend to update the corpus at least once per year.
The most recent version of the CR-UNSC corpus will always be available open access via Zenodo through its Concept DOI located at https://doi.org/10.5281/zenodo.7319783
#Law #UnitedNations #DigitalHumanities @npub1uc0…cpgf @npub1m7c…g454
https://fediscience.org/system/media_attachments/files/114/318/369/738/070/473/original/baf797db23399611.png

2025-04-05 10:52:54 GMT

I'm currently migrating my #OpenSource projects from #GitHub to #Codeberg and I have to say I'm impressed.
+ Codeberg web interface is much (!) faster and more responsive compared to GitHub
+ Git clone and push feel about twice as fast as GitHub (not that it matters much)
+ Much cleaner interface than GitHub, less bloat and less featuritis
+ Codeberg activity overview is very helpful and well organized
+ Codeberg migration tool works well
Note: the OSS git software that Codeberg runs is called Forgejo.
Codeberg: https://codeberg.org/
Forgejo: https://forgejo.org/

2025-04-02 15:34:54 GMT

I'll be honest, I don't intend to submit academic papers to US-based journals and conferences anymore.
1. Self-censorship of science at US universities and scientific organizations is in full swing and I see virtually no institutional resistance (although some individual scholars are protesting). I don't expect to get a fair hearing with human rights-oriented empirical legal research in peer review any longer.
2. The way this is going I would expect even currently neutral journals to fold within the usual timeframe of a paper review process. I don't feel like burning a paper in the hopes of getting it published before the door closes.
3. I prefer to submit my research to journals backed by institutions that support human dignity, diversity and the freedom of scientific research.
We've all seen where this went in Europe during the 1930s. I hope the US will learn from our history, but I am not optimistic right now.
#AcademicMastodon #ScienceMastodon

2025-03-25 14:26:46 GMT

This is neat: the United Nations have adopted a new set of 8 #OpenSource prinicples:
🔶 The UN Open Source Principles 🔶
1. Open by default: Making Open Source the standard approach for projects
2. Contribute back: Encouraging active participation in the Open Source ecosystem
3. Secure by design: Making security a priority in all software projects
4. Foster inclusive participation and community building: Enabling and facilitating diverse and inclusive contributions
5. Design for reusability: Designing projects to be interoperable across various platforms and ecosystems
6. Provide documentation: Providing thorough documentation for end-users, integrators and developers
7. RISE (recognize, incentivize, support and empower): Empowering individuals and communities to actively participate
8. Sustain and scale: Supporting the development of solutions that meet the evolving needs of the UN system and beyond.
Source: https://unite.un.org/news/osi-first-endorse-united-nations-open-source-principles
https://fediscience.org/system/media_attachments/files/114/223/532/160/352/885/original/221b643eddf269fe.png

2025-02-20 15:46:31 GMT

I hope you're all doing risk assessments and checking the exposure of your business/NGO/public institution to the US weaponizing its tech infrastructure (Google, Microsoft, Apple) and shutting down your operations if they disagree with your mission?
#Security #RiskAssessment #Tech

2025-02-20 14:21:43 GMT

Can anyone tell me what the availability/uptime and usability for Codeberg is these days?
Edit: here's a link to https://codeberg.org/
I am considering migrating all my open source projects there but have heard there were issues in the past.
#Codeberg #OpenSource #OpenScience #Git #AcademicMastodon

2025-01-23 18:08:57 GMT

- reply

Going forward will probably have to add a line to every post asking people to refresh then 😂

2025-01-23 18:07:10 GMT

- reply

Weird. My server (fediscience.org) shows the preview. I did have to refresh it once, though.

2025-01-23 18:05:38 GMT

Now a test for social preview images!
https://seanfobbe.com/posts/2024-05-14_complete-citation-network-un-security-council-published/

2024-10-02 21:51:57 GMT

- reply

Link to my #OpenAccess data repository: https://seanfobbe.com/data/
More Open Access goodness is available on my academic website: https://www.seanfobbe.com

2024-10-02 21:50:11 GMT

Thank you all for 150,000+ legal dataset downloads!
Get your ethically sourced, high-quality, open access and public domain-distributed legal data here: https://seanfobbe.com/data/
Currently available are 18 open legal datasets that cover the following subjects in international law and German domestic law: (cont.)
#OpenData #OpenScience #OpenSource #Law
https://fediscience.org/system/media_attachments/files/113/240/026/710/718/843/original/95f68563277641dd.png

2024-07-31 15:32:39 GMT

🔔 New Teaching Section 🔔
This #OpenAccess section on my website collects all the materials I’ve created to help people learn more about Legal Data Science.
🔸 Introduction to Legal Data Science
🔸 Tutorials
🔸 Essays
🔸 More materials!
Legal Data Science is the application of computational statistical methods to the legal domain. In other words, it combines programming, statistics and domain expertise (Conway 2013).
Link: https://seanfobbe.com/teaching/
#Teaching #Tutorials #OpenSource #DataScience
https://fediscience.org/system/media_attachments/files/112/881/808/516/763/760/original/38ab4364dab84d92.png

2024-05-15 15:55:03 GMT

- reply

The alternative "platform decay" (also by Doctorow) is my preferred term in polite settings.
Source: https://www.eff.org/deeplinks/2023/04/platforms-decay-lets-put-users-first