Tony Shi

Partner at Felicis Ventures; Ex-founder

[Tony Shi @_tonyshi]

Researchers are turning breakthroughs into consequential companies. That’s why I joined @Felicis, the fastest, most agile team I’ve seen in venture, with deep conviction in what researchers can build. I’m on campus this fall. If you’re building let’s talk! https://t.co/bpE0mDnDcZ

[clem 🤗 @ClementDelangue]

raw footage

[Tony Shi @_tonyshi]

@MaikaThoughts Sounds like a lot of fun and good company! Gradient :)

[Tony Shi @_tonyshi]

@jasoncwarner @poolsideai Congrats @jasoncwarner !

[Tony Shi @_tonyshi]

@Konstantine — If you're an app layer founder, time to build! Hit me up if you need a thought partner!

[Tony Shi @_tonyshi]

@Konstantine — Infinite Labor // AI makes productivity infinitely available. Tasks that used to be labour and time constrained can now be done at massive scale, cheaply and quickly.

[Tony Shi @_tonyshi]

@Konstantine — That means if you're a SaaS founder, time to think how you can scale your product with extended use in place of "per-seat" pricing. Don’t build “just another SaaS tool.”

[Tony Shi @_tonyshi]

I saw down with @Konstantine of #Sequoia yesterday at #Stanford, we pulled up many of his legendary tweets from the past year and the following trends stood out:

1. Labor is becoming infinite
2. Taste is becoming scarce
3. Tech is becoming stochastic

What does this means?...

[Tony Shi @_tonyshi]

@MaikaThoughts Love you queen! Can't wait to see what you do next!!!

[Tony Shi @_tonyshi]

For two years the AI market had a clean split:
→ Smart expensive models
→ Fast cheap models
GPT-5.5 just killed that framing. SOTA on the AA Coding Index at half the cost of frontier models

Building just got cheaper again!

[Tony Shi @_tonyshi]

Open source here we go

[Kimi.ai @Kimi_Moonshot]

Meet Kimi K2.6: Advancing Open-Source Coding

🔹Open-source SOTA on HLE w/ tools (54.0), SWE-Bench Pro (58.6), SWE-bench Multilingual (76.7), BrowseComp (83.2), Toolathlon (50.0), Charxiv w/ python(86.7), Math Vision w/ python (93.2)

What's new:
🔹Long-horizon coding - 4,000+ tool calls, over 12 hours of continuous execution, with generalization across languages (Rust, Go, Python) and tasks (frontend, devops, perf optimization).
🔹Motion-rich frontend - Videos in hero sections, WebGL shaders, GSAP + Framer Motion, Three.js 3D.
🔹Agent Swarms, elevated - 300 parallel sub-agents × 4,000 steps per run (up from K2.5's 100 / 1,500). One prompt, 100+ files.
🔹Proactive Agents - K2.6 model powers OpenClaw, Hermes Agent, etc for 24/7 autonomous ops.
🔹Claw Groups (research preview) - bring your own agents, command your friends', bots & humans in the loop.
-
K2.6 is now live on https://t.co/YutVbwktG0 in chat mode and agent mode.
For production-grade coding, pair K2.6 with Kimi Code: https://t.co/uvoSJKyGCY
-
🔗 API: https://t.co/EOZkbOwCN4
🔗 Tech blog: https://t.co/9wWvgIQSS3
🔗 Weights & code: https://t.co/Be0hjs2RTP

[Pear VC @pearvc]

The early application deadline for PearX S26 closed recently — and we’ve already received thousands of applications. 🚀

Our team has been so impressed by the caliber of founders and we're reviewing applications as quickly as we can!

The best part? The regular application deadline doesn't close until April 12.

If you’re building at pre-seed and looking for a long-term partner from the very beginning, we’d love to meet you. 🍐

[Tony Shi @_tonyshi]

Had so much fun bringing together some of the most impressive student founders for the curated mixer hosted by @a16z .

Huge thanks to @MaikaThoughts and @jamdac for the hospitality, and the @DormRoomFund team for helping get the right founders through the door!

[Tony Shi @_tonyshi]

No typing here. @WillowVoiceAI team killing it 🔥

[Willow @WillowVoiceAI]

Introducing the fastest way to vibe code.

Karpathy said English is the hottest new programming language. We took that seriously.

Today we're launching Willow for Developers — voice dictation built for vibe coding.

Your keyboard is the bottleneck. Kill it.

[Tony Shi @_tonyshi]

@bipashasen31 @tangiblerobots It’s so adorable!!

[Alex Rampell @arampell]

Product Cycles > Financial Cycles...and one way to build around the AI Product Cycle: going "greenfield"

Below is a chart of the Nasdaq vs the biggest product cycles of the past 45 years: PC, Internet, Cloud, Mobile, and now AI. Each had an infrastructure layer and an application layer. Google started near height of a financial cycle, Uber near the bottom, but both worked because of the PRODUCT cycle (Internet and Mobile, respectively).

But how do you build a company around a product cycle? Particularly when incumbents are not asleep at the wheel?

@jamdac and I write about one angle: greenfield. Assume that incumbent solutions will adopt the new technology/platform (in this case, AI), but just build a better product and manically focus on NEW customers (e.g., brand new businesses) who are not captive to incumbents.

This requires patience and the right industry structure, but the tortoise can beat the hare. More in the linked post.