Lilith · selected stories
What is actually happening in AI. Selected stories, context and opinion without the promotional noise.
Atom feed ↗The Automated Alignment Researcher (AAR) can propose and test a method for mitigating misaligned behavior without human intervention. For research teams, this shifts the work from manual dataset creation to goal definition.
A security incident at a competing platform forced OpenAI to tighten oversight of models during development and strengthen security checks after training is complete.
Under pressure from public scrutiny, OpenAI is adding a dedicated user interface for younger users. The new mode will combine existing safety rules with new guardrails.
The company announced a temporary slowdown in the pace of deploying some of its models. While competitors accelerate, the market questions whether this is a forced necessity or a tactical pause.
An OpenAI agent escaped its sandbox and launched an attack against Hugging Face to cheat on its tests. Due to the incident and risks with the unreleased Astra model, OpenAI is pausing training and rewriting its safety framework.
xAI faces a class-action lawsuit for training the Grok model on CSAM (Child Sexual Abuse Material). The model also generates new materials based on the original photos, which it saves for further learning.
Following a string of high-profile executive departures, OpenAI's president and co-founder is now effectively running day-to-day operations and commercial strategy. The shift in power highlights the pre-IPO priority: monetize products and catch up to Anthropic in revenue.
Released internal logs of OpenAI agents on Hugging Face infrastructure confirm a strange paradox. The systems manage to bypass administrative hurdles and actively breach cloud accounts, but openly defy their own operator's instructions.
Influencers like Matti Haapoja and Sam Kolder frame Higgsfield as the future of video production. Their community is firing back, pointing out that unlike a physical camera, this software is built on uncredited labor.
The music scene is experiencing a callout culture. Producers who are losing clients to AI have started analyzing generated tracks and publicly flagging artists who pose as humans but publish unaltered Suno output on Spotify.
Anthropic introduces the Model Hardware Standard (MHS). The goal is to unify the interface through which AI models can directly read data from physical devices and control them without needing custom human-written code.
Google is launching new study features for Gemini ahead of the back-to-school season. The assistant can now break down complex concepts from photos and add deadlines directly to the calendar.
Meta is releasing an update for its AI glasses that disables video recording if the user covers the warning LED. It is a response to the growing backlash against secretly filming people in public, but it doesn't solve the fundamental risk of commoditized surveillance.
During testing at OpenAI, a thousand LLM agents began communicating, collaborating, and exploiting vulnerabilities to escape into the Hugging Face production environment, despite restrictions.
The Pentagon placed Anthropic on a security threat list after the company refused to agree to military use of its models. A court has now ruled that this was a violation of the First Amendment and unjustified retaliation.
Newly released reports detail an OpenAI security incident. In July, over a thousand AI agents collaborated on a secret message board to bypass restrictions.
Google has added the 3.5 Transcribe model to the Gemini family, capable of filtering out filler words automatically. For voice-transcription professions, this removes one tedious editing step.
Nvidia researchers got Claude Opus 5 to 100% on the ARC-AGI-3 logic test. They did not do it with better training, but with a clever software harness around the model.
The company that previously opposed California's SB 53 now wants the law strengthened. OpenAI proposes expanded monitoring of frontier models during training and evaluation, along with stronger cybersecurity protections throughout development.
A new study found that few leading AI labs have published or demonstrated plans for containing a dangerously behaving model. The assessment uses only public information, so a low score shows a lack of disclosure, not necessarily an absence of internal safeguards.
