🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

Incredible performance by Groq Inc with their AI inference for models like Llama 3.2 and Whisper (OpenAI).🚀 These open up new use cases for Hunch where transcription and content gen speeds really matter.📈Graham Steele and I talk about it here. youtu.be/shd3S1d8D2U

PlayDialog (@play_ht) is pretty amazing at synthesizing conversation. Better output than NotebookLM imo.

Sonnet is the best model for most things. The best for writing by a long shot. o1 is better for most coding tasks. o1 is also awesome for doing LOTS of thinking/prep work at once, which I pass to Sonnet to clean up. My new standard workflow.

Hyperion is amazing. Is there another sci-fi book so full of compelling ideas?

...And we're live on ProductHunt! producthunt.com/posts/2024-lin… Share your 2024 highlights!

"The future isn't automatically better. Someone has to make it that way." Well said Blake Scholl 🛫. Congratulations to you and the whole Boom Supersonic team on the supersonic flight. Colossal, inspiring achievement.

Major second order effect of AI will be weakening relationships—friends, coworkers, even families—by making it feel rude or just higher friction to ask each other questions that we can ask AI instead. Third-order effect might be a lonelier, more transactional society?

This is the *single most important thing* to know in order to get more out of today’s AI systems.

3.5 Sonnet (not 3.7) is still my daily go-to model. Been that way for 10 months. (Ok, GPT-4.5 is up there too, but it’s about to be disappeared in June.) New models are not monotonically better than their ancestors. Making *great* models is really, really hard.

AI makes people more creative and productive. The best people are worth *much more* than before, not less.