Something I just told a founder: Stay as small as you can for as long as you can. People who come to visit your office should always be surprised that such an important company has so few employees.
+1 for "context engineering" over "prompt engineering".
People associate prompts with short task descriptions you'd give an LLM in your day-to-day use. When in every industrial-strength LLM app, context engineering is the delicate art and science of filling the context window with just the right information for the next step. Science because doing this right involves task descriptions and explanations, few shot examples, RAG, related (possibly multimodal) data, tools, state and history, compacting... Too little or of the wrong form and the LLM doesn't have the right context for optimal performance. Too much or too irrelevant and the LLM costs might go up and performance might come down. Doing this well is highly non-trivial. And art because of the guiding intuition around LLM psychology of people spirits.
On top of context engineering itself, an LLM app has to:
- break up problems just right into control flows
- pack the context windows just right
- dispatch calls to LLMs of the right kind and capability
- handle generation-verification UIUX flows
- a lot more - guardrails, security, evals, parallelism, prefetching, ...
So context engineering is just one small piece of an emerging thick layer of non-trivial software that coordinates individual LLM calls (and a lot more) into full LLM apps. The term "ChatGPT wrapper" is tired and really, really wrong.
@janniksin@carlosalcaraz@rolandgarros You will stand back up stronger. Lots of interesting moments you could have closed this. Each of these is a great learning opportunity. The most positive outcome is that you are up there with Alcaraz even on clay. You have to believe that!
@petergyang I think that MCP will change things. It will be easier for LLM to execute their plans when they have tools to extract real time info and take actions.
@jasonfried How to you blend “audio only” with the mention in the video that AI is a creativity force multiplier? ai needs some kind of canvas to be truly creative. No?
Building a successful startup requires to have the rare ability to run faster and harder as the mountain gets steeper. Most slow down. Some turn back. The best entrepreneurs accelerate—with an attitude—when it’s hardest, loneliest, and uphill all the way.
Great AI demos hide all the hard parts: edge cases, bugs, human input behind the scenes. A real product survives chaos. The best AI won’t impress in a demo. It’ll quietly deliver in production.
AI is stuck in a chat-first mindset, and that’s a dead end. The next AI breakthroughs aren’t going to be about bigger models. They’ll be about tools that help us build, create, and integrate AI into real workflows.
A good day to revisit my chat with the amazing @RaazHerzberg, the first PM at @wiz_io.
The craziest part of the story is that the $0 to $100m ARR in 18 months timeline included a pivot.
Here's the full chat: https://t.co/qM5ebj9oPl
100% - empowered design at every stage of product AND company building is a cheat code of sorts. tight feedback and iteration loops, speedy alignment around strategy and implications, and solutions to problems that often outperform what PMs / biz leaders **thought** they wanted.
Trusting yourself isn’t about confidence. It’s about courage. Confidence comes after you prove yourself right. Courage is trusting yourself before the proof exists.