Researchers Find Hidden Reasoning in ChatGPT's Encrypted Thought Processes
Security researchers uncovered encrypted thought processes in ChatGPT and other models, revealing passwords and API keys in public sessions.
Archive
Browse all published articles.
ARCHIVE
-
-
OpenAI Lets Employees Cash Out Another $7 Billion in Stock
OpenAI enabled employees to sell $7 billion in stock at its current $852 billion valuation, following a $122 billion funding round.
-
Anthropic Model Advances on Riemann Hypothesis
Anthropic's unreleased model made progress on the Riemann hypothesis, raising the lower bound of solutions for which the hypothesis holds to 650 tested ideas.
-
General Catalyst Leads $1.1B Round in 2-Month-Old River AI
River AI, a 2-month-old startup founded by xAI co-founder Igor Babuschkin, raised $1.1 billion in a seed/Series A round led by General Catalyst and AMP PBC.
-
Brad Lightcap Leaves OpenAI to Start New Venture
Brad Lightcap, OpenAI's longtime COO, is leaving to pursue a new venture, marking another executive shakeup at the AI research lab.
-
First Orion Uses Amazon Nova Act to Speed Up QA Automation
First Orion, a branded communications company, reduced QA bottlenecks by adopting Amazon Nova Act, improving test efficiency and reducing regression issues.
-
Pixieset Uses Amazon Bedrock to Boost AI Adoption by 35%
Pixieset achieved 35% AI feature adoption by solving a real problem with Amazon Bedrock, generating alt text for 750,000 photos in the first week.
-
ONESTRUCTION Builds Ishigaki-IDS Foundation Model with AWS GenAIIC
ONESTRUCTION, Inc. developed the Ishigaki-IDS foundation model, designed for construction BIM workflows, using AWS GenAIIC technical advisory.
-
Google's AMIE AI Enables Real-Time Video Consultations
Google's AMIE AI system demonstrated real-time clinical video consultations in a first-of-its-kind study, with favorable assessments from evaluators and patient actors.
-
OpenAI Tests Ads in ChatGPT Across Multiple Markets
OpenAI launched ads in ChatGPT in the UK, Mexico, Brazil, Japan, and South Korea on August 11, 2026, with plans to expand further.
-
Nvidia's Nemotron 3.5 Lightning Prioritizes Speed Over Intelligence
Nvidia's Nemotron 3.5 Lightning matches OpenAI's gpt-oss-120b on intelligence benchmarks with a quarter of the parameters and delivers the fastest inference speeds in its class.
-
Mistral Expands Open Models and Regional Inference for Sovereign AI
Mistral announced plans to build up to 1 GW of compute capacity in Europe by 2030, supporting regional AI sovereignty and open model access.