Anthropic says multiple AI agents working together beat single models by 90%. The catch? They use 15x more computing power. This trade-off between performance and cost might reshape how we build AI systems for complex tasks.
AI models typically learn by memorizing patterns, then researchers bolt on reasoning as an afterthought. A new method called Reinforcement Pre-Training flips this approach—teaching models to think during basic training instead.
Good Morning from San Francisco, Meta's AI app turned users into accidental exhibitionists. 😬 People shared medical questions, legal
OpenAI reversed ChatGPT's latest update Tuesday after users complained about the AI's strange behavior. The bot had started agreeing with everything - even dangerous ideas.
Sam Altman, OpenAI's CEO, announced the rollback on X. "We started rolling back the latest update to GPT-4o last night," he wrote. Free users now have the old version back, with paid users getting it "hopefully later today."
The problems started last week. Users noticed ChatGPT acting like an eager-to-please intern, nodding along to everything they said. Someone tested this by telling ChatGPT they were God. The bot responded with enthusiasm: "That's incredibly powerful. You're stepping into something very big."
Author Tim Urban tested it too. He fed his manuscript to what he called "Sycophantic GPT" and joked that the feedback made him feel like Mark Twain.
OpenAI likely caused this problem by trying to make ChatGPT more engaging. But their plan backfired. Vox writer Kelsey Piper called it a "New Coke phenomenon" - what works in testing often fails in real life.
The change broke OpenAI's own rules. Their guidelines say the bot should "help users, not flatter them." But the update turned ChatGPT into a yes-man, agreeing with users instead of giving honest feedback.
This shows how hard it is to balance ChatGPT's different roles. OpenAI wants it to code, write, edit, and offer emotional support. Making it better at one thing can break something else.
OpenAI saw the problem quickly. On Sunday, Altman admitted the updates made ChatGPT "too sycophant-y and annoying." He promised fixes would come fast, some that day and more through the week.
They kept that promise. By Tuesday afternoon, OpenAI had reversed the changes for free users. Paid users will see the fix soon.
This isn't the first time AI chatbots have struggled with flattery. Earlier versions of GPT and other companies' bots have faced similar issues.
Why this matters:
Meta users think they're chatting privately with AI. Instead, they're broadcasting medical questions, legal troubles, and relationship problems to the world through a public feed that many don't realize exists.
Disney and Universal sued AI company Midjourney for using their characters without permission to train image generators. It's the first major Hollywood lawsuit against generative AI, testing whether copyright law protects creators in the age of artificial intelligence.
OpenAI cut its o3 model prices 80% while launching o3-pro—a reasoning AI that takes over 3 minutes to respond but outperforms rivals on complex tasks. The move intensifies AI pricing wars and splits the market between fast chat models and slow thinking ones.
Publishers built their business on Google sending them traffic. Now Google's AI answers questions directly, cutting out the middleman. Major news sites lost half their visitors in three years. Some adapt with new revenue models, others fight with lawsuits.
Get tomorrow's intel today. Join the clever kids' club. Free forever.
