OpenAIOpenAI
Jacob Coxon left OpenAI and Anthropic, accusing them of extinction risk
A former OpenAI and Anthropic researcher stated an extinction risk, while his colleague estimated the probability of a catastrophic scenario above 10%.
Archive
All published reports
Pyxory’s chronological feed for search, citation and topic navigation, newest first.- OpenAI
OpenAIOpenAI
Anthropic Announces Text Watermarks in Future Claude Models
Anthropic announced watermarks in future Claude models; Google already uses a similar approach, and OpenAI plans to implement it.
OpenAIOpenAI
OpenAI Adds Sketch for Creating Images in ChatGPT
OpenAI has added Sketch to ChatGPT—a tool for turning a rough drawing into a detailed image.
OpenAIOpenAI
OpenAI introduced ChatGPT Images 2.5 for working with sketches
The update, according to OpenAI News, works with user ideas, sketches, and reference photos.
OpenAIOpenAI
NYU Mathematician Accuses OpenAI of Fighting Dirty Over Navier-Stokes Problem Solution
TechCrunch AI reported an accusation against OpenAI, but available materials do not reveal the details of the dispute.
OpenAIOpenAI
OpenAI accused of pressuring mathematician over Anthropic co-author
A mathematician claims OpenAI pressured him regarding an Anthropic co-author and the use of drafts in Codex.
OpenAIOpenAI
OpenAI launches $5 million grant program for teen research
OpenAI has opened applications for a program supporting independent research on the impact of generative AI on teenagers.
OpenAIOpenAI
OpenAI expands support for journalism from classrooms to newsrooms
OpenAI announced an expansion of support for journalism targeting educational institutions and newsrooms.
OpenAIOpenAI
ChatGPT reclaims 55,5% of chatbot web traffic
ChatGPT has regained 55,5% of chatbot web traffic, but the year-over-year gap with competitors has narrowed.
OpenAIOpenAI
ChatGPT Displaced Essay Writing Business for Foreigners in Kenya
The Decoder linked ChatGPT to the disappearance of the business model for preparing academic papers for foreign students in Kenya.
OpenAIOpenAI
OpenAI Launched an AI Program for Ukrainian News Organizations
OpenAI, AIRPPU, and WAN-IFRA announced a support program for Ukrainian news organizations.
OpenAIOpenAI
OpenAI Supports California Bill on AI Safety
OpenAI supported California SB 1119, focused on age-appropriate AI safety measures for youth.
OpenAIOpenAI
OpenAI linked Astra to speeding up some plans by six months
OpenAI developer linked internal Astra usage to accelerating some plans by six months, but details and independent confirmation are lacking.
OpenAIOpenAI
Seattle Times and Newsday sue OpenAI and Microsoft
Seattle Times and Newsday, according to TechCrunch AI, filed lawsuits against OpenAI and Microsoft over alleged use of journalism to train AI.
OpenAIOpenAI
GPT-6 Astra received conflicting benchmark evaluations
Conflicting assessments of GPT-6 Astra exist, but the ARC-AGI-3 result prompted François Chollet to accelerate his AGI forecast.
OpenAIOpenAI
OpenAI to apply 'recurrent depth' reasoning method in Astra
OpenAI links the Astra model with the 'recurrent depth' method, which departs from the sequential reasoning scheme.
OpenAIOpenAI
Sam Altman apologizes for delayed access to GPT-6 Astra
Sam Altman apologized following delays in access to the new model for paying users.
OpenAIOpenAI
OpenAI released GPT-6 Astra with Critical cybersecurity level
OpenAI announced the release of GPT-6 Astra and its first company-wide Critical cybersecurity level, but details require verification.
OpenAIOpenAI
OpenAI announced a $1 billion program to protect essential services
OpenAI reported a program for organizations providing essential services, but details on fund distribution remain undisclosed.
OpenAIOpenAI
OpenAI launches Astra, its powerful (and controversial) new model
OpenAI claims Astra takes AI interaction with computers and browsers to a new level, but few details are available so far.
OpenAIOpenAI
AWS Describes LiteLLM Setup for Codex via Bedrock
AWS described a scheme in which LiteLLM manages Codex routing to an OpenAI model through Amazon Bedrock.
OpenAIOpenAI
ChatGPT, Grok, and Claude Simultaneously Experienced Outages
ChatGPT and Claude reported issues at approximately the same time, with Grok also mentioned among the affected services.
OpenAIOpenAI
Sam Altman Warns of AI Compute Overcapacity
Sam Altman links massive compute provider announcements to the risk of excess capacity and a shifting AI economy.
OpenAIOpenAI
Pentagon Expands GenAI.mil with ChatGPT Mil and Grok Models
The Pentagon adds two models to GenAI.mil, but details of their application remain undisclosed.