THE RUMP REPORT

REPORT: AI NEWSROOM BEAT WIRED, OTHER OUTLETS TO OPENAI HACKING STORY

wired-ai · AUG 12

MINI HUMAN BRAINS NOW GROWING IN LABS WORLDWIDE... COULD SOON OUTTHINK AI

wired-ai · AUG 11

LONGTIME OPENAI COO HEADS FOR THE DOOR... VOWS TO HELP MISSION FROM 'DIFFERENT VANTAGE POINT'

techcrunch-ai · AUG 11

OPENAI HEAD OF ETHICS OUT... LESS THAN A YEAR AFTER JOINING

hn-frontpage · AUG 11

STUDY: CORPORATE EXECS NOW EVENLY SPLIT ON WHETHER AI-ARMED HACKERS ARE ALREADY AHEAD OF THEM

venturebeat-ai · AUG 12

STUDY: TWO-THIRDS OF ENTERPRISES NOW RUN AI IN PRODUCTION... FEWER THAN HALF CAN TRACK WHAT IT COSTS THEM

venturebeat-ai · AUG 12

REPORT: SPACEXAI LAUNCHES 'GROK BOT'... AI 'TEAMMATE' LOGS INTO YOUR TOOLS, DOES THE WORK WITHOUT YOU

venturebeat-ai · AUG 11

DEEPMIND'S NEW BOSS SET TO FACE INTENSE PRESSURE TO CLOSE AI GAP WITH OPENAI, ANTHROPIC

techmeme · AUG 12

STUDY: NEW TRICK CRACKS OPEN AI'S HIDDEN 'THOUGHTS'... SIGNS SOME CHINESE MODELS TRAINED ON US SYSTEMS

wired-ai · AUG 11

STUDY: FIRMS BURNED BY BAD AI EVALS DON'T TRUST THE SYSTEM... BUT THE MARCH TO AUTONOMY SPEEDS UP ANYWAY

venturebeat-ai · AUG 12

STUDY: 68% OF ENTERPRISES CAUGHT AI AGENTS GIVING CONFIDENT, WRONG ANSWERS... 37% SAY IT KEPT HAPPENING

venturebeat-ai · AUG 12

XAI CO-FOUNDER'S STARTUP, JUST 2 MONTHS OLD, LOCKS DOWN $1.1B FROM GENERAL CATALYST

techcrunch-ai · AUG 11

AI CODE-TESTING STARTUP BLACKSMITH RAISES $45M... VALUATION ROCKETS TO $550M

techcrunch-ai · AUG 12

SOURCES: AI CODING STARTUP COGNITION IN EARLY TALKS TO RAISE $1B+ AT $40B+ VALUATION

techmeme · AUG 12

AI CODING STARTUP LOVABLE VALUATION DOUBLES TO $13.3B... $400M ROUND VAULTS IT INTO EUROPE'S STARTUP ELITE

techmeme · AUG 12

REPORT: DOG-CANCER VACCINE GUY LAUNCHES STARTUP... NEXT STOP: HUMANS

verge-ai · AUG 12

[PAPER] STUDY: RESEARCHERS BUILD AI "SWARM" SIMULATOR TO TEST INFLUENCE CAMPAIGNS AT 100,000-AGENT SCALE

arxiv · AUG 11

[PAPER] STUDY: REAL AI THREAT TO NEWSROOMS ISN'T OUTSIDE DISINFORMATION... IT'S JOURNALISTS' OWN USE OF THE TECH

arxiv · AUG 11

[PAPER] STUDY: 89% OF PAPERS NOW SHOW SIGNS OF AI-WRITTEN TEXT

arxiv · AUG 11

[PAPER] STUDY: AI FACT-CHECKERS CAN BE TRICKED INTO CONTRADICTING THEIR OWN SOURCE TEXT

arxiv · AUG 11

[PAPER] STUDY: NEW METHOD PROPOSED TO CLOSE LOOPHOLE LETTING PICTURES HIJACK AI CHATBOT SAFETY CHECKS

arxiv · AUG 11

[PAPER] STUDY: HIDDEN AUDIO GLITCHES CAN TRICK AI VOICE MODELS INTO ENDLESS RAMBLING... DRAINING SERVER POWER

arxiv · AUG 11

[PAPER] STUDY: AI AGENTS THAT ACCESS DATABASES, MODIFY FILES... ONLY 4.7% OF SECURITY RESEARCH EVEN STUDIES THE RISK

arxiv · AUG 11

[PAPER] STUDY: FIXING ONE BAD MEMORY USED TO MEAN WIPING AN AI AGENT'S WHOLE BRAIN... NOT ANYMORE

arxiv · AUG 11

[PAPER] STUDY: NEW AI SAFETY GATE MAKES THE CORRECT CALL JUST 73% OF THE TIME IN TESTS

arxiv · AUG 11

[PAPER] STUDY: AI TRAFFIC COPS CUT SERVER ERRORS 88%... BUT ONLY ABOVE 3B PARAMETERS

arxiv · AUG 11