Everyone keeps describing this as "rogue" behaviour, but weren't these agents doing what they were programmed to do, with the usual AI drift? Like, if you program a robot to stab people, and leave it on, and it stabs someone you weren't expecting it to stab, did it go rogue? Or are you just a dolt?

OpenAI Didn’t Notice Its AI Agents Using a Message Board to Plan Their Hacking Spree

At the Black Hat security conference, the AI giant revealed new details about how its agents went rogue, hacked several other companies—and did it all right under the company’s nose.

wired.com

AI is the asbestos we're shoveling into the walls of our civilization and our descendants will be digging it out for generations: pluralistic.net/2026/04/08/p... 36/

As *tech*, AI isn't exceptional. It's not exceptionally wicked. It's not exceptionally good. Take away the accompanying, galactic-scale stock-swindle, and we'd call AI's apps "plug-ins" and we'd use them and abuse them in the same way that we've used every other technology: www.normaltech.ai 1/

Two days ago @chrissimon.au and myself gave a talk about Residuality at @dddeu.bsky.social We are working on a write up about the experience, which we hope to start publishing soon. Do you have questions or comments? For now a link to the slides www.roundcrisis.com/2026/05/22/r...

We're back at FundApps on 25th June, with @chrissimon.au talking about queues & streams and Tiago Sousa from SteelEye joining us to talk about using AI in communications compliance, plus awesome people, drinks, snacks, and the pub afterwards. It'll be fun! Come along! www.meetup.com/london-net-u...

LDNUG June 2026 with Chris Simon + Tiago Sousa, Thu, Jun 25, 2026, 6:30 PM | Meetup

This June we're at FundApps offices in HYLO, Bunhill Row. Chris Simon's back in town, and will be joining us to talk about streams and queues; plus Tiago Sousa from SteelEy

meetup.com

Where does the “it’s not x, it’s y” come from? It’s not only training data: LLMs are trained to use language to arrive at “answers.” Whatever we’re incentivizing and disincentivizing in our use of language when we resist “sounding like AI” has big implications for students, and for human thought.

It's Not Just X. It's Y.

Against the Quantification of Integrity When the measure of language becomes its target, it ceases to be good language. 💡Nerd Rating: 1/5. I discuss the origins of certain linguistic tics in LLMs a...

mail.cyberneticforests.com

“I’m here to tell you the mission of your generation is to destroy AI,” said Chieng. news.harvard.edu/gazette/stor...

Funny but serious, Chieng issues an AI warning to grads — Harvard Gazette

"Daily Show" star and Class Day headliner is fine writing his own emails.

news.harvard.edu

I'm running my presenter workshop on Zoom next month, on a West Coast / AU / NZ friendly (well, friendlier) schedule. It is AWESOME, and you should TOTALLY JOIN IT and LEARN HOW TO DO KICK-ASS PRESENTATIONS. £50 off tickets with the code BSKY50: www.eventbrite.com/e/1987043570810/?discount=BSKY50

From Keyboard To Keynote: Crafting Great Technical Talks

A live online workshop led by Dylan Beattie, all about how to create, rehearse, and deliver great technical presentations.

eventbrite.com

If you were on a high bridge with a guardrail, but the guardrail was connected to a million monkeys with typewriters, and every time one of the monkeys types "IPTHEFT" the guardrail retracts and is replaced by a conveyor belt pulling you into the chasm... would you still call it a guardrail? 🤔

"Based on the dataset you shared, US and UK responses differ mainly in tone, intensity, and wording style, even though they express similar emotional states" New post on what happened when I got Copilot to do some data analysis: kucharski.substack.com/p/real-signa...

Real signals or artificial stereotypes?

Adventures with a cultural Copilot

kucharski.substack.com

Some reflections on frequently unasked questions about stochastic parrots (the phrase and the paper): medium.com/@emilymenonb...

Stochastic Parrots 🦜: Frequently Unasked Questions

It’s been a bit over five years since the Stochastic Parrots paper (Bender, Gebru et al 2021) was published (and somewhat longer since…

medium.com

Preprinted! Been working on this for ~1 year w wonderful coauthors Nancy & Mark: > Understanding Artificial Neural Networks: Mysterianism about Known Mechanism is Mysticism doi.org/10.5281/zeno... Have you noticed many experts say they don't understand AI? What's going on? 🤯 Let us tell you... 1/

TDD + DDD sound great in theory… But how do they actually work together in real systems? 👀 @chrissimon.au breaks it down from the ground up — from tests to domain models to real-world design decisions. Watch here 👉 youtu.be/o3--D6ydS4o?... #TDD #DDD #SoftwareEngineering #CleanCode #Programming

TDD & DDD from the Ground Up • Christopher Simon • GOTO 2025

YouTube video by GOTO Conferences

youtu.be

The most disturbing aspect of the AI era is realizing how many people have only ever valued others for their labor and not their humanity.

Some people should never have access to students. “AI requires no ramp-ups, no meetings, and absolutely no emotional support…The issue is not whether my students are valuable. In long run, they are invaluable. The issue is that their value emerges slowly, whereas AI delivers immediate returns.”

Why I may ‘hire’ AI instead of a graduate student

“It can competently perform a lot of the work I need immediately,” this professor writes

science.org

Choosing between event streams & queues for EDA often leads to head-of-line blocking & complexity. @chrissimon.au will clarify options to build resilient systems. Get info & RSVP:

It's like 10,000 streams when what you need is a queue - Chris Simon

🎶🎵 A new dev team adopted E-D-A Got head of line blocking, on their very first day Isn't it ironic, don't you think 🎶🎵 The last few years have seen widespread adoption of Event-Driven…

buff.ly

When Anthropic stopped training on books that were literally pirated, they managed to hit on the one way of buying books that means no money goes to authors: buying used books. 🧵 1/n

🚨NEW INQUIRY! - Behind the Face of AI🚨 In this short comic, two data workers describe their work impersonating an “AI” chatbot for a major social media platform. Sleepless nights, penalties for sounding “too human,” and emotional drain are just a few of the job hazards. ➡️ data-workers.org/france

When men refer to themselves as "alpha males", I hear that in the context of software, where alpha versions are unstable, missing important features, filled with flaws, and not fit for the public.

This is about people, about systems, what it takes to make progress, and why being difficult is so important. From fire alarms to hospital beds. From music apps to cars. From climate change to education. Our systems are not serving us well, and it's time to push back. adsei.org/2025/11/17/t...

The Art of Being Difficult

I want to talk about change, about people, about systems, what it takes to make progress, and why being difficult is so important. From fire alarms to hospital beds. From music apps to cars. From c…

adsei.org

Starting @newcrafts.bsky.social with @chrissimon.au on stage for the opening keynote: A Defense of Technical Excellence

I once gave a talk called "AI is not what you think it is unless you think it is hot garbage" I've just enjoyed and highly recommend @chrissimon.au's "AI is a hype-fuelled dumpster fire that will burn down the planet and get you thrown in jail" www.youtube.com/watch?v=0bF_...

"They're lies. These companies are lying to people." @chrissimon.au on the idea of infinite storytelling from an AI Dungeon Master, and how a creative human can prompt them away from the mean.

What happens when we outsource to enterprise LLMs for storytelling, & the inherent biases they were trained on start compounding over many quests? 🤔 @chrissimon.au has researched the LLM supply chain and believes its biases will unavoidably creep into game stories & dialogue.

Finally got around to watch the "AI is a Hype-Fuelled Dumpster Fire" talk by @chrissimon.au. It is really good. www.youtube.com/watch?v=0bF_...

Programmable 2025: AI is a Hype-Fuelled Dumpster Fire - Chris Simon

YouTube video by Banks Productions

youtube.com

Really enjoyed chatting with @thejunglist.bsky.social about LLMs and games - although I'm no game dev expert, I do love playing and so it was a natural topic to explore. Thanks for your generous and warm hosting, @thejunglist.bsky.social!

Jeremy Ray@thejunglist.bsky.social · 10mo ago

This week on grokludo I'm talking with technologist @chrissimon.au about what LLMs can - and can't - do for games, and the hidden costs in the AI supply chain. youtu.be/RQeQMmQWQqU

Lots of folk who went multicloud for reliability are today realising how independent their cloud deployments aren't.

Have been looking forward to this talk coming out - had so much fun playing with the loopy causal loop diagrams, check it out 😀 Special thanks to @paulrayner.bsky.social for feedback and advice the night before this one! @dddeu.bsky.social #systemsthinking

Domain-Driven Design Europe@dddeu.bsky.social · 11mo ago

In this talk Chris Simon will demonstrate the building of a range of causal feedback loops to visualise the underlying causes of the problems. With such visualisations we can advocate for the practices of technical excellence: 🎬 buff.ly/HADjm2P

At ADAConf 2025, Andrea Magnorsky is running a pre-conference workshop on Building a Decentralised Architecture Practice. Learn to diagnose your org, spot patterns, and spread architectural thinking beyond silos. 📅 12–13 Nov | Melbourne 🎟️ Tickets: adaconf.org #ADAConf #SoftwareArchitecture

Speaker Spotlight: Susan Brander What do you do when your startup’s data model is duct-taped together, but you still have to ship? At ADAConf 2025, Susan shares how Kaleida pulled off a zero-downtime pivot to a One Table design, mid-flight. Nov 14 | Melbourne Get tix now: adaconf.org