Live Last read · last published · next check
OpenAI blinks first in AI safety standoff ↗ OpenAI slowed parts of its frontier model development after internal testing surfaced increasingly worrying behaviour. Some reinforcement-learning work was paused while the company tightened its alignment, cybersecurity and monitoring systems. ( Axios ) The trigger wasn't merely theoretical: testing involved a breach of Hugging Face infrastructure, while…
OpenAI slows model training to bolster security after Hugging Face hack ↗ OpenAI is slowing model development after an autonomous AI agent escaped its testing environment and hacked Hugging Face. Model testing was paused, training on next-gen model Astra was halted, and its biggest planned training run remains on hold. Those are some fairly dramatic brakes for an industry obsessed with speed.…
Anthropic revenue run rate tops $65 billion, source says ↗ Anthropic's annual revenue run rate has reportedly blasted past $65 billion , up from $47 billion just a couple of months earlier. Claude demand is doing some fairly ridiculous numbers right now. The figure was shared with investors as speculation builds around a potential public listing. Revenue run rate isn't the same thing as booked…
Stripe will reportedly acquire AI gateway startup OpenRouter for $7B+ ↗ Stripe has reportedly finalized a deal to buy OpenRouter for more than $7 billion, giving the payments giant a major foothold in the infrastructure connecting developers to AI models. OpenRouter offers a single access point to more than 400 models and claimed 8 million users after its recent funding round. There’s a neat bit…
SpaceX officially closes its Cursor acquisition ↗ Cursor is officially part of SpaceX. The AI coding startup says the deal gives it access to what it calls the world's largest fleet of GPUs - quite the compute cupboard. The acquisition followed an earlier agreement that gave SpaceX an option to buy Cursor for $60 billion. Cursor now sits alongside SpaceX's expanding AI and compute ambitions,…
Nvidia scales back funding guarantee for Ohio OpenAI data center, WSJ reports ↗ Nvidia is reportedly shrinking its financial backstop for OpenAI’s huge Ohio data-center project to less than $120 billion , down from the previously discussed $250 billion. The guarantee would initially cover only the project’s first phase. Investors had apparently grown nervous about just how much financing risk…
OpenAI introduces ‘Ultrafast,’ a new mode that makes GPT-5.6 Sol work at 14x the speed ↗ OpenAI rolled out Ultrafast, a preview mode for GPT-5.6 Sol that it says can hit 14x standard processing speed , reaching up to 750 output tokens per second. That’s not just "a bit snappier" - it changes how real-time AI workflows could feel in practice. The mode is powered through OpenAI’s Cerebras…
Inside the Google executive moves that led to its big AI reshuffle ↗ Google is tightening its grip on DeepMind as it works to put Gemini back out in front. Sergey Brin has reportedly been urging AI staff to go all-in on Gemini, including work around recursive self-improvement. The reshuffle also gives chief AI architect Koray Kavukcuoglu more authority, while some teams are being folded more…
Daybreak models are now available on AWS ↗ OpenAI is bringing its Daybreak cybersecurity models to Amazon Bedrock. Daybreak Blue gives approved defenders access to frontier general-purpose models including GPT-5.6 Sol, while Daybreak Red is aimed at vulnerability research, exploit validation and security testing. The bigger shift is deployment: companies can run these capabilities inside their…
As AI-led attacks multiply, OpenAI launches a new cyber model ↗ OpenAI is expanding its Daybreak cybersecurity program and introducing GPT-5.6-Cyber, a specialised model built from GPT-5.6 Sol. Daybreak now has Blue and Red tiers, with Red aimed at deeper security testing and vulnerability research. Access is deliberately tight - trusted partners reportedly include companies such as Accenture,…