Building your own AI stack

Source: https://www.instagram.com/reel/DcYqvXaJ2xY/?igsi=YzM2cWJzdXplMDVy

AI generated summary of the linked source. Posted anonymously. Transcripts are published only for public links, never for uploads, documents or pasted text.

Your library

The AI link is plain markdown. Paste it into any assistant and it can read the whole public entry.

Summary

A recap of Sequoia's closed door workshop with 80 founders and AI leaders on how companies should build their own AI stack. The talk centers on Sonya's four step framework (strategy, team, legibility, technical roadmap) and the shift from closed to open source AI now that open models are closing the gap with frontier ones.

Core ideas

One context card plus 6 ideas

Owning your AI stack matters now: Public signals from Alex Karp and Jensen Huang, including a letter signed by 25 companies, point to open source becoming the dominant posture in AI, described here as "the new punk."
Sequoia's workshop as a playbook: 80 founders and AI leaders spent a session working through how to build AI systems from scratch, and the speaker frames this as required viewing for founders, investors, and corporate leaders with an AI mandate.
Four step framework: Sonya lays out strategy (what to own vs rent), team (staffing for intelligence production), legibility (an underrated step she flags as critical), and technical roadmap (the building blocks needed).
The open source treadmill is flipping: Fine tuning open models used to get erased by every new frontier release within months. Releases like Kimi K3 and Meta's open weights are now closing that gap.
Closed stack vs open stack breakdown: The presentation maps specific companies to each layer of both stacks, giving a concrete reference point for orgs deciding where to build.
Not zero sum, but about leverage: The real question isn't picking one stack entirely. It's figuring out which layers of the stack give an organization the most leverage.

Quotes

“Clearly, open source is the new punk, sort of like the supreme hoodie for AI.”
@atlasberry008

“Step one to owning your intelligence is strategy. What parts of your AI do you want to own? What parts do you want to rent?”
@atlasberry008

“I really think this gets glossed over and is incredibly important.”
@atlasberry008

“Training open models used to be a treadmill. You'd spend like months fine-tuning, and then the next frontier release would just erase all your gains.”
@atlasberry008

“It's obviously not like a zero-sum game but it seems like it's about getting smarter on which layers of the stack actually give you the most leverage.”
@atlasberry008

Resources

Links to the tools, products, repos or reading this entry mentions. Found by a web search of what the entry names, so you can go and use them.

Sonya
newsfilter.io
Canonical Sequoia profile page for Sonya Huang, the partner who presented the four-step framework.
Kimi K3
kimi.ai
Official Kimi K3 page for the open-weight model with 2.8T parameters and 1M-token context.

Remix this knowledge

Turn this entry into something you can post. Pick a format and get a fresh, original rewrite.

Remixes are AI generated, original wording, and free to reuse.

Talk it through

Open a round table on this entry and argue with other people about what it means.

Start a round table

Drop your own

Paste any video, podcast or audio link and get the transcript, the core ideas and the quotes back in one pass.

Try Drop