mit-tech-review · OCT 9
Google turns Gemini into a work agent with its own email address
techcrunch-ai · OCT 8
REPORT: OpenAI revenue $20 billion below earlier projection
techcrunch-ai · OCT 8
Anthropic launching 'presidential engagement' program for 2028 election
techmeme · OCT 9
China's Manus raises over $500M in first funding round since split with Meta
techcrunch-ai · OCT 8
Nous Research hits $1.5B valuation, launches AI agents for business users
techcrunch-ai · OCT 7
OPINION: Writer asks why the industry isn't freaking out about DeepSeek 4.1 Flash...
hn-frontpage · OCT 8
verge-ai · OCT 8
Zuckerberg launched Muse despite safety concerns after rival agent app took off
techmeme · OCT 9
[PAPER] STUDY: Self-improving AI loop yields 1.2-1.6x performance gains in two cycles
arxiv · OCT 8
OpenAI withdraws three mathematical results
techcrunch-ai · OCT 8
Three fired OpenAI safety researchers dispute misconduct claims, warn of chilling effect
techcrunch-ai · OCT 8
USA Today sues OpenAI for $250 million+, says it copied 'hundreds of thousands' of articles
verge-ai · OCT 8
Anthropic bans 'sustained and needless abusive or cruel behavior' toward Claude in policy overhaul
techcrunch-ai · OCT 8
SOURCES: SoftBank seeking up to $100B from Gulf investors for AI buyout fund
techmeme · OCT 9
AI leaderboard Arena raises $200 million led by Lightspeed and Khosla
techcrunch-ai · OCT 8
Wave of cyberattacks hits Japanese companies, exposing data of millions
techmeme · OCT 9
REPORT: Developer pulls ARTEX AI agent from open source after bank-hacking claims
techmeme · OCT 9
r-machinelearning · OCT 8
[PAPER] STUDY: AI agents fail to reliably pick the best training data... gains over random limited
arxiv · OCT 8
[PAPER] STUDY: Red-teaming a few AI agents can't guarantee safety at scale, researchers say
arxiv · OCT 8
[PAPER] STUDY: White-box probes reach 98.8% AUC in SHADE-Arena, topping an Opus 5.5 text monitor...
arxiv · OCT 8
[PAPER] STUDY: Six lines of server log text raise AI guardrail fail-open rate from 0% to 63%
arxiv · OCT 8
[PAPER] STUDY: AI safety framework cuts jailbreak success from 86% to 31%...
arxiv · OCT 8
[PAPER] STUDY: AI judges name the right law, but verdict often doesn't change when the law is swapped...
arxiv · OCT 8
arxiv · OCT 8
[PAPER] STUDY: Coding assistants pass 75% of tests with software architects, under 25% with non-coders
arxiv · OCT 8
[PAPER] RoboRSI robot teaches itself household cleanup over 104 rounds
arxiv · OCT 8
[PAPER] STUDY: AI-text detectors collapse when new models arrive... 99% catch rate falls to 3.8%
arxiv · OCT 8
arxiv · OCT 8
[PAPER] STUDY: Context doesn't erase AI memorization... it can unlock new leaks
arxiv · OCT 8
[PAPER] STUDY: GPT-6-astra retracts authors' claims outright when rewriting papers
arxiv · OCT 8
arxiv · OCT 8
arxiv · OCT 8
[PAPER] STUDY: Single punctuation mark fools AI grader... but only under certain request formats
arxiv · OCT 8