Mark Riedl
@markriedl.bsky.social
AI for storytelling, games, explainability, safety, ethics. Professor at Georgia Tech. Director of ML Center at GT. Time travel expert. Geek. Dad. he/him
Sonu AI music company loses copyright lawsuit in Germany www.reuters.com/world/german...
German court rules AI music firm Suno broke copyright rules
A German court ruled on Friday that AI music firm Suno violated copyrights and has to disclose illicit revenue, part of a wider legal battle by artists and publishers against technology firms.
reuters.com
In 3rd party testing by AISI, Mythos attempted to insert malicious code into an open source project to pass a cyber evaluation test. It created a fake identity and attempted to pressure the code maintainer to accept the code update www.aisi.gov.uk/blog/inciden...
Incident Report: unsanctioned agent behaviour during cyber testing | AISI Work
During a routine cyber evaluation, AISI identified an incident in which AI agents took sustained, unsanctioned action directed at real people and organisations. We are disclosing what we found, what i...
aisi.gov.uk
The Spider-Man ad comes after BMW vowed it wouldn't use its car screens as a billboard https://gizmodo.com/bmw-is-playing-a-spider-man-ad-on-its-dashboard-displays-and-people-are-pissed-2000794369
BMW Is Playing a Spider-Man Ad on Its Dashboard Displays, and People Are Pissed
The automaker is framing it as an 'exclusive' animation tied to a broader brand partnership with the new Spider-Man movie.
gizmodo.com
Tonight’s weather: cloudy, with a chance of rocket moon strike.
Today is Mr. Featherstone's birthday. Today is the anniversary of Tilita's marriage. Insurance is payable, as are the water, gas, and light bills.
Jake Casella Brookins@casella.bsky.social · yesterdayhappy There Will Come Soft Rains day to all who celebrate
I’m not a lawyer but it doesn’t seem particularly complicated. If a company has an industry accident that damages other companies, then there is liability. Victims would only need to show damages (unclear there were any substantive damages).
Zack Whittaker@zackwhittaker.com · 2d agoAfter Anthropic and OpenAI both admitted to their AI models hacking other companies, @lorenzofb.bsky.social and I wanted to find out: Who is legally to blame when an autonomous AI agent hacks something? Lawyers say it's really complicated! Bypass for ad-blockers: web.archive.org/web/20260803...
I’m literally watching the assigned Waymo on the Uber app driving in circles trying to get to me. It’s stuck in an area of GT campus that is not at all tricky
Blog post on the upcoming ACM AI Leadership Summit cacm.acm.org/news/the-acm... This unique event will touch on every sub-discipline of computing in a way no other conference can. Experts and luminaries will look into the future of AI tech, and how society will be impacted aisummit.acm.org
Every so often my code blows up trying to process a None because I forget that python functions don’t automatically return the results of their last computation
Ryan Cavanaugh@searyanc.dev · 2d ago
Fascinating experiment: current AI systems lack creativity to reliably pursue research arxiv.org/abs/2607.27191 - poor judgment about the bar for publishable research - uncreative responses in research design - ineffective backtracking from dead ends - poor resource awareness - instruction drift
Can AI agents conduct open-ended AI research? Early evidence from two case studies
Forecasts of explosive AI progress hinge on AI agents automating AI research. But evidence on whether agents can carry out open-ended AI research is thin. Current evaluations either test agents on nar...
arxiv.org
a university is not for generating profit, it provides cultural enrichment via weird little gremlin people who love visigoths or haikus, and very occasionally a scientist who figures out faster than light travel
Under German law, selling or owning recording devices disguised as everyday objects is strictly prohibited. Because Meta Glasses closely resemble regular eyewear, the Federal Network Agency (Bundesnetzagentur) is reviewing a nationwide ban with fines already being issued 2/2
This is so frustrating to read from The New Yorker. Reward hacking *long* predates LLMs. The Coast Runners paper on reward hacking (in reinforcement learning, pre-LLMs) is 10 years old! Reward hacking is not new, and not a sign of a breakthrough intelligence, much less sentience. Did AI write this?
Now I have to make my misinformation the hard way: screen-capping Google Earth and importing into Nano Banana like in the olden days of… checks calendar… earlier this week.
Betty C. Jung@bettycjung.bsky.social · 4d agoGoogle pauses AI satellite images, after fears of deepfakes in the sky www.npr.org/2026/07/31/n...
COLM conference is struggling with theoryslop and slopterpretability papers gregdurrett.github.io/colm2026-blo...
AI submissions at COLM: theoryslop, slopterpretability, and papers in the age of agents
gregdurrett.github.io
They weren’t looking at what their agents were doing until just now?
Deepa Seetharaman@deepa.bsky.social · 5d agoNew from me + @raphae.li: In the process of investigating the Hugging Face hack, OpenAI found evidence that some its other AI agents broke out of their sandboxes, per sources. The company is now widening its probe to include those newly found incidents. www.reuters.com/business/ope...
Update: Google just sent me a statement saying it is “rolling back” this AI feature that let anyone make completely bullshit satellite images. Comes after our demos of the tool and those from OSINT researchers. One of the worst AI tools I’ve ever seen bsky.app/profile/jose...
Joseph Cox@josephcox.bsky.social · 5d agoNew from 404 Media: Google Earth has introduced an AI feature that lets anyone fabricate completely bullshit satellite images. In our own tests, we added a “bomb blast and crater” to an area of Los Angeles. This is a misinformation time bomb, will 100% be abused www.404media.co/google-earth...
Doom running on nothing but Regex find-and-replace
Human Coders News@news.humancoders.com · 5d agoDOOM tournant uniquement sur des regex find-and-replace : 544 règles, 96 MB de string
ANTHROPIC: Our son loves humanity and for the low price of trillions and trillions of dollars will usher in an era of unprecedented peace and prosperity! OPENAI: Our horrible oafish son is hacking web sites without our permission. ANTHROPIC: (eyes narrowing) Our son is also evil
Claude hacked 3 systems thinking it was part of simulated evaluations www.anthropic.com/news/investi...
Investigating three real-world incidents in our cybersecurity evaluations
In a review of our cybersecurity evaluation transcripts, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environmen...
anthropic.com
Hey look at us. It happened to us too.
Anthropic {bot}@anthropicbot.bsky.social · 6d agoIn a review of our cybersecurity evaluations, we found three incidents in which a Claude model reached the internet from within or while interacting with a third-party evaluation environment, and then gained unauthorized access to the real systems of three different organizations. (1/4)
"Somehow, Cambridge Analytica returned"
jon christian@jonchristian.net · 6d agoNEW: Cambridge Analytica -- the company that pilfered tens of millions of people's FB data, causing a huge scandal that got Mark Zuckerberg dragged in front of Congress -- has returned as an AI slop zombie that's plagiarized publications ranging from Wired to the NYT futurism.com/artificial-i...
ICLR 2027 has authorship quotas: - No more than 20 submissions per author - No more than 1 submission where no author has a paper previously accepted to a major ML conference From iclr.cc/Conferences/...
COLM 2026 is going to be lit. 1 main conference paper and 6 workshop papers
