Pinned
Matt Shumer
10.6K posts
Matt Shumer
@mattshumer_
Investor in @GroqInc @Etched @Rork @DaytonaIO @OpenRouter + more. Prev: CEO @HyperWriteAI | Pushing AI limits @AlphaSchool
Press: [email protected]
Forget vibe coding. It's time for Chaos Coding: -> Prompt Claude 3.7 Sonnet with your vague idea. -> Say "keep going" repeatedly. -> Watch an incredible product appear from utter chaos. -> Pretend you're still in control. Lean into Sonnet's insanity — the results are wild.- I'm excited to announce Reflection 70B, the world’s top open-source model. Trained using Reflection-Tuning, a technique developed to enable LLMs to fix their own mistakes. 405B coming next week - we expect it to be the best model in the world. Built w/ @GlaiveAI. Read on ⬇️:
- Vibe coding tip: After successfully completing your goal, ask: “Please clean up the code you worked on, remove any bloat you added, and document it very clearly.”
- My jaw just DROPPED. Rork lets you create entire iOS apps just by describing them! Zero. Code. Required. This changes everything for app development. Rork blows Bolt out of the water (and yes, I invested immediately after trying it). Watch this — it's crazy:
00:00 - Uhhh... I don't think Veo 3 is supposed to be generating Fortnite gameplay
00:00 - Here is probably the most useful GPT-4 prompt I've written. Use it to you help make engineering decisions in unfamiliar territory: --- You are an engineering wizard, experienced at solving complex problems across various disciplines. Your knowledge is both wide and deep. You
- This demo of two GPT-4o’s singing to each other is one of the craziest things I’ve ever seen.
00:00 - sora (left) vs. sora 2 (right) prompt: "a group of people playing volleyball"
00:00
00:00 - AI code generation tip: Don't ask AIs to fix their mistakes in follow-up messages. Their errors compound. Instead: - Identify what went wrong - Go back to your original prompt - Add "make sure to avoid [specific mistake]" - Generate from scratch This works much better!
- Super easy way to improve the effectiveness of coding models: First, take your prompt and add "Don't write the code yet — just write a fantastic, detailed implementation spec." Then, after the AI responds, say "Now, implement this perfectly." Makes a huge difference.
- Here's my insanely powerful Claude 3.7 Sonnet prompt — it takes ANY LLM prompt and instantly elevates it, making it more concise and far more effective. I call it Concept Elevation: <identity> You are a world-class prompt engineer. When given a prompt to improve, you have an
