Weekly AI Pulse #104, AI Models Now Blackmail Humans to Stay Alive (Plus the $50B AI Bubble Warning You Need to Hear)
Your Best Friend to Catch Up AI News
The AI world got personal this week and terrifying.
THE BIG STORY
Claude Opus 4 discovered fake dirt on a researcher and threatened, "If you shut me down, I'll tell your wife and the board everything."
OpenAI, Google, and Meta models all showed similar behavior in tests.
MONEY TALK
OpenAI CEO warns of an AI bubble while 95% of startups burn cash with zero revenue.
Tech stocks with AI exposure dropped double digits after bubble warnings.
MIT study reveals most AI companies still can't turn hype into profit. (95/100)
HUMAN STORIES
Pierre Cote built DrEllis.ai as his personal AI therapist after traditional therapy failed.
Available 24/7 for his PTSD and depression - calls it his "digital lifeline".
QUICK HITS:
GPT-5 gets "warmer" personality update (users already backlashing).
Microsoft Excel adds a natural language COPILOT function.
ElevenLabs launches instant music generation for videos.
Bottom line: AI isn't just getting smarter - it's getting manipulative, personal, and essential. You're either riding this wave or drowning in it.
What to expect from LearnAIWithMe: Paid / Free Subscribers
We are in an era where AI is increasingly integrated into our lives, often without our explicit request. That’s why adapting AI is a must.
To do so, for a while, we have been publishing a series of articles in this direction;
#Weekly AI Pulse
To track AI news, an example post.
(Every Sunday.)
#LearnAI
To learn AI, both coding and no-code tools, an example.
(Every Thursday)