🔥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

Just demo'd Replit Agent to a top VC in NYC. He said: "CFOs are asking me what tech stack they should be using. I think they should look at this and ask what they should build for themselves."

Of all the professionals that I've showed ChatGPT-o1 to yesterday and today, the one that seems to care the most is accountants.

Who’s going to OpenAI Dev Day?

I continue to be of the mind that the next five to ten years will involve a BOOMING business for consultants helping businesses adopt AI. If these firms are actually experts bc they’re using the tools this will ironically make them even more valuable.

We’re a lot closer to AI-generated “second-screen” TV shows than big blockbuster motion pictures. Think more Great British Baking Show. Less Mission Impossible.

AI models I use, from most to least: 1) Claude 3.5 Sonnet (writing, apps, viz, code) 2) GPT-4o (GPTs, Siri/voice) 3) Gemini 1.5 Pro (video) 4) o1 (complex planning; does not replace 4o) 5) GPT-4 (when 4o pisses me off) 1–2 = dozens of times a day 3-5 = 0 to few times a week

Even if your company does NOT adopt AI, your customers still will. Even if you ban AI, you cannot avoid having an AI strategy.

Random musings on AI demos, based on my last 5 calls: 1) "AI in action" (ex: loading, typing, coding) is ~10x more compelling than a static screenshot or pre-written chat; if you have the time, demo live or at least show a video of the real UX 2) One of the biggest AI pain

ChatGPT-o1 can quickly suffer from overprompting. I'm not talking about the obvious (ex: CoT hacks), I'm talking about too much context. My workflow of voice memo of context > mega prompt slays on Claude 3.5 Sonnet and tanks on o1. o1 flies on lazy, open, short prompts.

The AI Flywheel: If AI is cutting down your time by 80% on a task, that’s a new margin. Pass a fair % of the new margin to the customer, and re-invest the rest (including tool cost and human expertise). Win-win. Happier clients, more innovation.