in order to remain on the frontier of capabilities you basically have to throw out all your AI code every 6 months and build it from scratch

Chris Tate people who are refusing to adapt are going to fall behind very fast.

The goal is turning the organization’s memory into something structured and retrievable, so you stop relying on whoever happened to be in the room to (1) achieve the objective discussed and (2) decide on who else should be informed.

The winners will be companies that build the muscle of day-to-day use early enough for the gains to compound. Start learning now, or watch the advantage slip away. In short, maybe your goal isn’t just adoption, but collective AI understanding.

my body does not know the difference between a mistake at work and being hunted for sport

The two things that actually make AI useful at scale: having an enterprise context graph that connects people, docs, tickets, metrics, and systems, and wiring AI straight into the painful, repeatable workflows (status reporting, handoffs, ticket triage, CRM hygiene, release

It used to be controversial to state that there exist 10x engineers. People would get all worked up. I have news for you. AI has created 100x, possibly 1000x engineers. Time to update those hiring plans.

The incumbent SaaS vendors who survive will be those who make all of their primitives and data models easily accessible via agents. I want to rip-and-replace vendors that require using lots of in-app UI, low-code, drag-and-drop workflows. I want to keep the vendors that let me

Corollary: as users get Claude Code pilled, they won't automate UIs. They will push work down the stack until they hit something agents can operate on directly with code. Take a risk or finance org: if today's workflow requires dropping a pdf into a Drive, clicking around Google

Being "AI native" will mean a complete rebuild of the entire tech stack to run a business. Every tool. Every system. Every workflow. The companies that figure this out first will make everyone else look like they're still running on fax machines.

The Claude Code hype is real Over our Imbue winter break, I wanted to see how fast I could reconstruct the lowest level of our product as an AI engineering principles experiment—vibe engineering, not just vibe coding. In < 2 weeks, I shipped over 50k lines of high quality

At this point, claude code needs to start attending standup.

When Karpathy says this, you have to pay attention.

Nicholas Meriano Andrej Karpathy Yep here’s an example. /dedupe skill invoked on every issue github.com/anthropics/cla…

Hire developers who cannot sit idly when they notice something is broken. Extreme ownership.

Overheard in SF: Person 1: “Rome wasn’t built in a day” Person 2: “Yes but they didn’t have Claude Code”

Agentic Engineering > Vibe Coding Let’s be professionals

100%! The number one way to protect your career from AI is to be the most aggressive user of AI in your field. RUN AT IT. Humans evolve with their tools. Keep evolving.

me a month ago: claude code is a chatgpt moment. me now: claude code is a 100x bigger deal than chatgpt itself.

AI can make work faster, but a fear is that relying on it may make it harder to learn new skills on the job. We ran an experiment with software engineers to learn more. Coding with AI led to a decrease in mastery—but this depended on how people used it. anthropic.com/research/AI-as…