Ryan @_PaperMoose_

Running a 1 man AI Agency currently at 40k MRR since launch in January. AI Native Evangelist Currently working on ARC-AGI Prize with @gregkamradt vunda.ai SF Joined August 2017

Tweets

3K
Followers

958
Following

1K
Likes

32K

Ryan @_PaperMoose_

2 weeks ago

me when I learned about YC revenue swapping

0 0 1 49 0

Download Image

Every big idea starts as a lie. We invent a strong version → bold, sweeping, career-making. Reality pushes back → evidence tears it down. What survives is the weak version → less sexy, but true. Freud. Rational economics. Sapir–Whorf. AI hype. Strong claims die. Weak claims…

0 0 0 57 0

annie ❤️‍🔥 @AnnieLiao_2000

3 weeks ago

Excited to launch: Video Mania! 📺 A daily game that teaches you how to prompt realistic AI videos Live on Product Hunt (what score will you get?)👇

40 15 173 18K 72

Download Video

Kyle Corbitt @corbtt

a month ago

Announcing MCP•RL: teach your model how to use any MCP server automatically using reinforcement learning! Just connect any MCP server, and your model will start playing with it and (using RL) "learn from experience" how to use its tools most effectively!

57 200 2K 203K 2K

Download Image

Ryan @_PaperMoose_

a month ago

🧠 HUGE day for AI. Two major releases just dropped and they signal a shift. 1. OpenAI just open-sourced models for the first time since GPT-2 The new gpt-oss series comes in two sizes: •gpt-oss-20b: lightweight enough to run on a 16 GB laptop •gpt-oss-120b: high-performance…

0 0 2 178 0

Ryan @_PaperMoose_

a month ago

Launching a product is a series of 100 tiny steps. Right now, I’m waiting for verification on my Chrome extension. Feels like a small thing, but it’s the culmination of dozens of tiny dependencies: -Writing a privacy policy -Setting up Google Workspace -Configuring icons and…

0 0 2 108 0

Download Image

Li Yin @panda_liyin

a month ago

Sometimes it’s just good to get out there. Dm to join @EcZachly, @AnnieLiao_2000 , @_PaperMoose_ and me for the next Founders Hike!

7 4 30 4K 1

Download Image

Ryan @_PaperMoose_

a month ago

“How do you make your diagrams?” I've been getting this question a lot lately. The answer: I built a tool. But not one you need to pay for. It started because I needed a way to quickly create clear, high-level architecture diagrams for clients—especially for things like AWS,…

0 1 2 147 1

Download Video

Shreya Shankar @sh_reya

a month ago

no one asked for a list of every AI evals question ever, but Hamel made it anyway 🤯

12 89 883 51K 1K

Download Image

Ryan @_PaperMoose_

a month ago

A lot of my clients are building agents on Bedrock - not because it’s the best tool for the job, but because it’s the path of least resistance. They already use AWS. They already have accounts, infra, procurement, compliance. Spinning up agents on Bedrock is convenient. But…

0 0 2 100 0

Download Image

annie ❤️‍🔥 @AnnieLiao_2000

a month ago

pumped

AGI Fight Club @agifightclub

a month ago

pumped

1 2 9 4K 0

Download Image

1 2 11 3K 0

Ryan @_PaperMoose_

a month ago

I just built notion in cursor.

0 0 1 84 0

Download Image

Ryan @_PaperMoose_

a month ago

Cursor is way better than notion, and I'm about to prove it.

0 0 1 92 0

Ryan @_PaperMoose_

a month ago

Growth feels like failure. until it doesn’t. Discomfort isn’t a detour. It is the path. A growth-filled life is, by definition, an uncomfortable one.

2 0 3 93 0

Download Image

Ryan @_PaperMoose_

a month ago

Most “evals” I see are vibes. Ask your team: What are our top 3 failure modes in prod and how common are they? Crickets = you’re playing whack-a-mole, not evaluating. I just published a breakdown of how to do application-centric evals: Analyze → Measure → Improve Live in…

0 0 1 85 1

Greg Kamradt @GregKamradt

2 months ago

The world is moving towards agents Static benchmarks don't measure what agents do best (multi-turn reasoning) Thus, interactive benchmarks: * Terminal Bench (@alexgshaw, @Mike_A_Merrill) * Text Arena (@LeonGuertler) * BALROG (@PaglieriDavide, @_rockt) * ARC-AGI-3 (@arcprize)

14 30 221 20K 164

Greg Kamradt @GregKamradt

2 months ago

.@arcprize listed on the @Alibaba_Qwen model card 2nd model card for us in 2 weeks Excited for ARC-AGI to be seen as a supported way to measure model performance x.com/Alibaba_Qwen/s…

Qwen @Alibaba_Qwen

2 months ago

.@arcprize listed on the @Alibaba_Qwen model card 2nd model card for us in 2 weeks Excited for ARC-AGI to be seen as a supported way to measure model performance x.com/Alibaba_Qwen/s… https://t.co/dvhvI3AlMp

15 28 437 62K 35

Download Image

7 8 95 23K 9

Download Image

Ryan @_PaperMoose_

2 months ago

I went surfing today. And it was exactly what I needed. I read a post recently about saying yes to things outside of work, and it was much needed. Sometimes the best thing you can do for your work is to step away from it. Today reminded me that not everything productive has…

0 0 0 83 0

Ryan @_PaperMoose_

2 months ago

Yesterday, I was at Andreessen Horowitz to celebrate the launch of ARC-AGI3 – the latest evolution of the ARC benchmark, now focused on interactive, game-based intelligence. François Chollet (founder of ARC in 2019) spoke with Dwarkesh Patel about what it means to define…