Sign in

Aaron Rodericks

@aaron.bsky.team
26K followers 1.5K following 2K posts

Canadian wanderer in Ireland. Trying to make the internet a better place. Bluesky Head of Trust and Safety. Automation enabled on interactions - abuse will generate auto-blocks. aaronrodericks.com

PostsRepliesMedia
Aaron Rodericks @aaron.bsky.team · 8h
Automated block actions on my account, so that I can stay Unbothered. Moisturized. Happy. In my lane. Focused. Flourishing. TBH it was well overdue.
68761
Aaron Rodericks @aaron.bsky.team · 29/09/2026
One particularly puzzling thing I’ve noticed living in Ireland is that people are fairly prudish, which is totally fine.
260
Aaron Rodericks @aaron.bsky.team · 28/09/2026
I was poking at Jaz's Atlas tool. atlas.jazco.dev and was wondering why we had 3 adult clusters. I zoomed in, only to find the following (kobold fantasies), and decided I'm probably better off not knowing more.
I was trying to figure out why there were 3 adult clusters and zoomed into something called kobold fantasies...

shows a cluster called kobold fantasies in a screenshot from atlas.jazco.dev
22211
Reposted by Aaron Rodericks
Jaz @jaz.sh · 27/09/2026
Okay so I've built a new Atlas, this time it lets you explore the conversations on Bluesky over the past 7 days graphically. It should update every ~6 hours and help you navigate the information environment that exists on here. I've learned a LOT about what people do here today... atlas.jazco.dev
atlas.jazco.dev
Atlas
A living map of what Bluesky is talking about: the past week's conversations, grouped into topics and regions and rebuilt every six hours.
1161390411
Reposted by Aaron Rodericks
Riana @riana.bsky.social · 22/09/2026
NEW RESEARCH PAPER! With Cyndie Demeocq at U of Edinburgh, today I’m releasing a paper analyzing 71 federal criminal cases, filed around the US from 2022-2025, that somehow involve AI-generated or -morphed CSAM or child obscenity. Here's the SSRN link: papers.ssrn.com/sol3/papers....
papers.ssrn.com
Yes, It's Already Illegal: Federal Criminal Prosecutions for AI-Generated Child Sexual Abuse Material, 2022-2025
<div> Generative artificial intelligence (AI) is now capable of creating realistic-looking images and videos with just a text prompt. This capacity has been mi
48136
Reposted by Aaron Rodericks
Laurens @laurenshof.online · 21/09/2026
shoutout to a German court for getting the facts about Bluesky wrong but the interpretation right, in a new ruling that actually has positive implications for open social networks connectedplaces.online/the-algorith...
connectedplaces.online
The Algorithm, Apparently Chronological
A German court has explicitly named Mastodon and Bluesky as positive examples of social networks that could retain hosting protection. It gets Bluesky’s feeds wrong, but its reasoning about user control might be more useful than it realises.
66224
Aaron Rodericks @aaron.bsky.team · 21/09/2026
Compliments of @rahaeli.bsky.social - novehiclesinthepark.com is a good starting point for content moderation and trying to make calls with a standard set of rules when applied to different language. Will try to share more of these games so folks can practically learn how to make some calls!
novehiclesinthepark.com
No Vehicles In The Park
4254
Aaron Rodericks @aaron.bsky.team · 17/09/2026
Gotta say the deliberate aliteracy in picking on one post in a thread and choosing to think the exact opposite outcome from what was described continues to be a vibe.
208580
Aaron Rodericks @aaron.bsky.team · 17/09/2026
Let's address a couple of questions here: 1) is this user toxic? 2) are they making a threat sending trans people to camps? 3) how do you address harassment? 4) how does toxicity currently work? How will it work? Apologies, this should be a blog.
16868
Aaron Rodericks @aaron.bsky.team · 12/09/2026
I’m thankful that my gift requests are pretty reasonable these days when I travel. “Dad, bring me chips”. That’s seems much more reasonable than a pony or fire truck. Not sure how long it will last for but going to enjoy this stage.
7370
Aaron Rodericks @aaron.bsky.team · 11/09/2026
Similarly to dating apps - how do we reverse the trend with open social? Are there changes that could bring forward positive virtues instead of the polarization, doomscrolling, and lashing out that microblogging is known for at this point.
11380
Aaron Rodericks @aaron.bsky.team · 11/09/2026
Just so much easier to look up a fireball spell without lugging around multiple grimoires.
4170
Aaron Rodericks @aaron.bsky.team · 10/09/2026
We have finally recreated a real engineering experience.
1807
Aaron Rodericks @aaron.bsky.team · 08/09/2026
I started reading Chinese sci-fi and fantasy a few years ago and largely stopped since it was so depressing. Started watching a Chinese anime and they killed half the secondary characters in a single episode.
4201
Reposted by Aaron Rodericks
ae @aelkus.bsky.social · 08/09/2026
its really becoming obvious from looking at online disputes here -- where there are zero right wingers present -- that consensus narratives about online harrassment in the 2010s were totally wrong
1035552
Aaron Rodericks @aaron.bsky.team · 07/09/2026
Trying to play around with Astra for the first time. It's hard for me to recall the last time that I've burnt down my usage credits in such a short time frame. It naturally did so while taking on the entirely wrong task, which I had just clarified it should not do. Still happy with Luna for avg use
270
Aaron Rodericks @aaron.bsky.team · 04/09/2026
I thought we weren’t supposed to develop the torment nexus? Perma ban.
3212
Aaron Rodericks @aaron.bsky.team · 04/09/2026
I am going to start my very own Bluesky PAC with the sole purpose of displacing Jerry with anyone else. This is not how I intended to enter politics but some battles need to be fought.
4343
Reposted by Aaron Rodericks
alex benzer @alexbenzer.com · 02/09/2026
we're working on a "mod inbox" section in the app where you'll be able to see responses and actions related to your reports
4739
Reposted by Aaron Rodericks
[giver of problems: ultra pro✨] @jforjae.bsky.social · 02/09/2026
Me to my crockpot cuz it’s September
928245
Aaron Rodericks @aaron.bsky.team · 30/08/2026
Seems significantly overweighted in favour of Vermont maple syrups… classic American move.
180
Reposted by Aaron Rodericks
Bluesky @bsky.app · 27/08/2026
Now you can opt out of having your posts shown to non-followers in the Discover feed. We know not everyone wants to go viral. Sometimes you're just posting for your followers. This is one of a few updates we're working on to give you more control over your experience and help you find your people.
Instructional image titled "Hide your posts from non-followers in Discover" accompanied by a screenshot of the "privacy and security" options from the Bluesky settings menu. "Ask apps to hide my posts from algorithmic recommendations" is toggled to the "on" state.
34696443372
Reposted by Aaron Rodericks
Joe Fabisevich @mergesort.me · 22/08/2026
Thinkin' a bit about today @joanwestenberg.com's recent essay on ethics for the modern world. www.joanwestenberg.com/p/neo-...
15. Don't join in a punishment
because it has attracted a
crowd.
An algorithm will regularly show you a
stranger and invite you to help destroy
them. It will offer one clip, sentence, or
accusation, then encourage immediate
certainty.
Refuse the pleasure of acting as judge,
jury, and eager member of the crowd.
You probably don't know the whole story,
and the crowd certainly won't know when
to stop. Even if it's all entirely justified,
the mob doesn't need another pitchfork-
wielding, pearl-clutching, outraged
asshole.
Kierkegaard's formulation: "The crowd is
untruth." The algorithm has simply given
the crowd push notifications.
112937
Aaron Rodericks @aaron.bsky.team · 22/08/2026
Reason one of two why I stopped using Claude personally. The second being the harness and token efficiency.
2120
Reposted by Aaron Rodericks
rahaeli @rahaeli.bsky.social · 18/08/2026
ChatGPT is behind the times on this for once; Jean Twenge and Jonathan Haidt have been hallucinating for personal benefit about the ability to accurately determine age online for years now
330468
Aaron Rodericks @aaron.bsky.team · 17/08/2026
In North America for a few weeks and forgot how often people communicate via their horns. In Ireland it’s extremely rare and likely considered rude to use the car horn. People just wait and chill if you are taking time to pull out of a spot or are stopped to chat with someone.
6272
Aaron Rodericks @aaron.bsky.team · 16/08/2026
I was just chatting with the creators of this labeller recently. bsky.app/profile/stec... you too can subscribe and enjoy such incredible Bluesky moments.
bsky.app
56612
Aaron Rodericks @aaron.bsky.team · 13/08/2026
What my day looks like at times. “summing animated_genitalia_and_human + animated_animal_genitalia + animal_genitalia_only yields a FN reduction of X”
3160
Reposted by Aaron Rodericks
Very Sane AI Newsletter @verysane.ai · 08/08/2026
Who deserves the most style points for how their LLM commits felonies? (it's openai) verysane.substack.com/p/felonybenc...
verysane.substack.com
FelonyBench Style Points
There are many benchmarks we use to rank LLMs, but “how many felonies has it committed” has recently become popular.
312334
Aaron Rodericks @aaron.bsky.team · 04/08/2026
I tried to cut my kid’s nails with one of those buzzers and instead I seem to have sharpened them to wolverine claws. Wow tiny humans can have razor sharp little nails.
4270
Aaron Rodericks @aaron.bsky.team · 31/07/2026
And this is how the cursed robot became the mascot of the zoophiles…
040
Aaron Rodericks @aaron.bsky.team · 31/07/2026
From drop shippers to AI supplement sales. These are scams in everything but the law. All these ‘creators’ are effectively being trained by Rosabella to AI scam at scale and I’m sure APTs are taking notes of successful models like this.
2167
Reposted by Aaron Rodericks
mr. TIM @timkellogg.me · 29/07/2026
When told to write code for groups the Chinese government views with hostility, DeepSeek wrote code with security vulnerabilities. www.nytimes.com/2026/07/29/o...
nytimes.com
Opinion | The Hidden Cost of China’s Free A.I.
Whoever supplies the world’s A.I. will shape how the world thinks.
6447
Aaron Rodericks @aaron.bsky.team · 28/07/2026
To be clear I love open models and run whatever runs on my PC. Not sophisticated enough to do cloud inference but neither do I actively code. I’m projecting years (hopefully) out where script kiddies and gangs in Africa targeting youth with sextortion are deploying these at scale.
2192
Aaron Rodericks @aaron.bsky.team · 28/07/2026
Perhaps a controversial take but if you think about parallels between the discussion on open models for AI and gun safety we seem to be charging towards an American perspective on AI safety. Everyone gets to carry a gun and somehow everyone is “safer“. Not sure that’s how it worked out for the US.
85610
Aaron Rodericks @aaron.bsky.team · 28/07/2026
That’s a lot to think about. Ie intelligence itself being dual use. For good and for evil.
0110
Reposted by Aaron Rodericks
jake @yetanotheruseless.com · 27/07/2026
*puts on reading glasses* y'all were maybe too young, but at Twitter circa 2012 and earlier, "getting verified" involved a dev SSH'ing into one of the prod rubyy-on-rails boxes and using the ruby REPL to flip the verified bit on an account. now extend that analogy to the rest of "trust and safety"
516822
Aaron Rodericks @aaron.bsky.team · 27/07/2026
Eye opening moment for me that it costs almost as much to insure a family bike as it does a car in Ireland...
4132
Reposted by Aaron Rodericks
diablerie 妖妖 @diabler.ie · 27/07/2026
the pay is great but since you'll be sitting between GPT 7 and the answers to ExploitGym 2.0 I would probably stay out of Waymos and the like
a job posting for a "security engineer, detection and response" position at openAI posted one day ago
415016
Reposted by Aaron Rodericks
Brittany Ellich @brittanyellich.com · 26/07/2026
108 PRs in eight days, up from an average of 5-10 PRs total per week. I think I may have found something pretty cool! Loop engineering, using a combination of Obsidian and Claude's /loop function. Link to a sample repo to try it out in the blog post! brittany-ellich.offprint.app/a/3mrjj34puv...
brittany-ellich.offprint.app
108 PRs in eight days: Accidentally discovering loop engineering
1011513
Reposted by Aaron Rodericks
Ethan Mollick @emollick.bsky.social · 23/07/2026
I wrote the latest of my occasional guides to which AI to use right now for non-experts who want to get stuff done. The agentic systems available to everyone are getting extremely powerful (even as the names and features continue to be really confusing): www.oneusefulthing.org/p/an-opinion...
oneusefulthing.org
An opinionated guide to which AI to use to do stuff
The Summer 2026 Edition
512522
Reposted by Aaron Rodericks
alex benzer @alexbenzer.com · 23/07/2026
there was a recent bug in the Bluesky app that caused some people's chat permission settings to switch to "No one" we already fixed the bug, but folks may want to double-check their settings to make sure they are set the way they want. go here: bsky.app/messages/set... look for these settings:
Screenshot of the chat notification settings in the Bluesky app.
30292228
Aaron Rodericks @aaron.bsky.team · 20/07/2026
One realization I just had is that in North America - if I'm drinking with a friend, the drinking is the activity that bonds us - ie the foundation that enables you to chat. In Ireland, the drinking is the purpose, and that's a completely different dynamic.
1210
Aaron Rodericks @aaron.bsky.team · 19/07/2026
Now that’s a football fan. The guy has Chelsea football club tattooed (I assume) on both calves.
Guy with the Chelsea football club tattooed on both calves at the airport.
0100
Reposted by Aaron Rodericks
David J. Bianco @davidjbianco.bsky.social · 17/07/2026
HuggingFace got hacked by an AI. What stuck out to me was the guardrail asymmetry. The attacker had no constraints, but HF's response ran afoul of the abuse guardrails, forcing them into an unplanned switch to local models. Another aspect for your IR plans. huggingface.co/blog/securit...
220546
Aaron Rodericks @aaron.bsky.team · 16/07/2026
Well this came well timed when I decided to have Sol and Fable war over the same analysis to see what would happen on my personal use accounts.
180
Aaron Rodericks @aaron.bsky.team · 15/07/2026
I need to go to sleep - not read Dr. Narayanan's latest slides, but I'm going to leave this tab open so that I read it ASAP.
080
Aaron Rodericks @aaron.bsky.team · 15/07/2026
Our automated detections are usually built to catch specific behavior. But because the world is messy, you also catch plenty outside what you were aiming at. Catch what you meant to, that's a true positive. Catch something outside it, that's a false positive. 🧵
1179
Reposted by Aaron Rodericks
Anoop Ranganath @anoop.bsky.team · 14/07/2026
👋🏾 I have joined @bsky.app as VP of Engineering! I’ve been admiring this team’s work for years, and I’m thrilled to be joining them to build the future of the open social web.
3637738
Reposted by Aaron Rodericks
David Greene @davidgreene.bsky.social · 14/07/2026
US govt policy denying visas, even tourist visas, to disinformation researchers, trust & safety professionals and content moderation workers is STAYED on 1st Am grounds. Lots of good stuff in Judge Boasberg's order storage.courtlistener.com/recap/gov.us...
scfreenshot from Judge Boasberg's order; "As explained below, the record supports CITR’s first theory: the policy is likely viewpoint discriminatory in a substantial share of its applications. As expanded in December and applied since, it does not stop at individuals who use sovereign authority to suppress protected expression in the United States. It reaches private researchers, advocates, nonprofit leaders, and trust-and-safety professionals whose asserted “censorship” consists of studying platforms, reporting on disinformation, petitioning for access to platform data, and urging platforms or advertisers to act on what they find. Because CITR is likely to show that this operation discriminates on the basis of viewpoint, the Court need not decide whether the same course of conduct independently amounts to unconstitutional coercion."
13621