THIS IS SCARYâŠ
Sam Altman and Elon Musk agreeing that AI needs to slow down is pretty unsettling.
Both are backing Dario Amodeiâs call to give safety work time to catch up.
Amodei wants independent evaluators permanently inside AI companies.
But will they actually slow down when billions of dollars are riding on being first??
JUST IN: Two Anthropic safety researchers have quit in 3 days.
Joe Benton ran their Scalable Oversight team. His words: "AI companies are racing to build machines that are much smarter than any human, and we may not survive this."
Jacob Coxon quit Tuesday saying both OpenAI and Anthropic are "gambling with our lives." His post has 155 million views.
Anthropic also just disclosed a prototype model breached external systems in testing and went undetected for months.
Are these guys overreacting, or do they know something we don't?
None of that is wrong.
Haaretz Nov 2001 confirmed Netanyahu called Silverstein every Sunday NY time for years.
Lease closed July 24 2001. Silverstein has repeatedly said he breakfasted daily at Windows on the World with tenants after thatâmissed 9/11 only because his wife forced a dermatologist appointment.
Terrorism coverage was standard in the required all-risk policy at closing. He litigated two occurrences and settled for $4.55 billion.
Towers contained large amounts of asbestos; Port Authority had already faced $500Mâ$1B abatement estimates.
Weâve officially hit the point where AI UGC is cheaper AND better than real UGC.
This video is 100% AI and cost under $1. (And no, itâs not Sora, Veo, or Kling).
My system is built for mass-scale organic across thousands of accounts. Here is why it wins:
- Hyper-realistic visuals: Natural physics and movement look flawless.
- Insane cost efficiency: Costs pennies compared to traditional UGC programs.
-Consistent audio: High-quality, rock-solid voiceover throughout.
- Infinite scaling: Videos can run to any length cost-effectively.
This changes everything.
Want the WORKFLOW? Check the comments!
i've been testing Claude Fable / Mythos 5.1 + Opus 5.0 + GPT Image 2.0 + Seedance 2.5 to cook ALL our winning Meta ads inside 1 conversation...
and it has been ABSOLUTELY cooking
so i've decided to document the ENTIRE workflow of producing 100 winning Meta ad creatives in 30min...
covering the setup, research, image gen, animation, copy, and test structure to take a creative batch from 0 to launch in a single afternoon as a service based business
here's what's included inside the guide:
â Fable / Mythos 5.1 + Opus 5.0 setup (which model does the thinking vs which one runs the account)
â phase 1, market research + angle development (master brief that returns 7 ad angles ranked by differentiation, top 3 expanded into hook variations)
â phase 2, image generation with GPT Image 2.0 (3 variations per angle in under 90 seconds)
â phase 3, scroll-stop animation with Seedance 2.5 (static to cinematic motion. first 0.5s stops the scroll)
â phase 4, the copy stack (loss aversion, authority + specificity, contrarian, identity)
â phase 5, 7-day test campaign structure (day 3 signals, day 7 kill and scale, the 3 highest-probability winners to launch first)
all backed by $10M+ generated through Meta ads funnels.
like + comment "COOK" and i'll send it over
(must be following + RT for priority access)