1229
Harry

@htormey #1229

☘️ Eng at @coinbase. I specialize in web3, Native iOS & #ReactNative. Formerly @apple @facebook. Tw: @htormey🏠 MIA | PDX. PFP @crypto_coven
1523 Follower 190 Following
It’s nice to touch grass from time to time.
How I feel on the first day of uptober.
Decided to get myself a WHOOP for prime day, obviously I had to try out the AI coach feature. Happy to say it’s great at sleep tracking and basic react native tasks.
/React
Clip from my interview with @nickcherry
where we talk about React Native performance lessons he learned from building the Coinbase Mobile App. Also, if you want to chat with me about React Native, I'll be at ChainReactConf this week in PDX.

https://youtu.be/dAliQd1FaNs?si=fYUqdTRjSIBquitL
Clip from my interview with @nickcherry. Why expo was wrong for React Native at Coinbase in 2018 but is essential today.

https://youtu.be/jX1xx5nrMGY?si=BDEPcKnfaOdHX4cw
/dev
My interview with @nickcherry about migrating Coinbase's mobile apps to React Native is live. We talk about React Native perf, his work building @warpcast, crypto + mobile, tooling & wrapping native libraries
https://stepchange.work/blog/migrating-coinbases-56-million-users-to-react-native-key-lessons-and-takeaways
/dev
Motto I live by.
GM! Anyone having issues with scaling Postgres? I can help.
AI engineers are compared to humans based on SWE-bench, it mainly covers Python tasks with ≤15 line changes evaluated by unittests. I wrote this article to give you a framework to assess if SWE-bench is relevant to you https://stepchange-blog.ghost.io/why-do-ai-software-engineers-like-devin-struggle-to-fix-bugs/
/AI
Interesting that Gemini pro 1.5 with its large 1m token context window seems to perform poorly on programming tasks with AI agents. https://twitter.com/htormey/status/1776288547148124179?s=46&t=Bc9m1uknKGxrhqisx2EO2g
/AI
What’s your startups plan for high quality Glycine? If your CTO says anything other than Donghua Jinlong, run: https://www.instagram.com/reel/C5QBc57uDlg/?igsh=MTc4MmM1YmI2Ng==
Apparently posted this meme to Facebook 10 years ago. Still seems relevant today.
Happy St Patrick’s day.
First attempt at Greek Chocolate cake.
Cognition Labs just released their report explaining Devins 13% score on the SWEBench benchmarks. Mandatory reading if you are interested in AI writing code https://www.cognition-labs.com/post/swe-bench-technical-report
My instagram reels are being overrun with Australian mining jobs propaganda.. Anyone else seeing this? https://www.instagram.com/reel/C31HUPFI5Rm/?igsh=MTc4MmM1YmI2Ng==
Does farcaster have a pinned cast feature yet? I like this Twitter/X feature.
I wrote up some of my thoughts on Cognition Labs AI software engineer Devin. In this article I talk about how LLMs are benchmarked for coding tasks and why I’m not worried about Software engineers being replaced in the near future https://www.htormey.org/can-you-replace-your-software-engineers-with-ai/
/AI
One of my favorite lines from @jessepollak is to focus on working on things that give you energy. You can really see this with Base and how successful it’s become. It takes a lot of energy to power through the frustrating parts of building a product like that. The farcaster team has the same energy. Hats off to both!
Hard to believe $coin was trading around $54 on April 28th last year and is trading at $207 today. I still think this statement is true today. I also think now is a great time to work at Coinbase if you have the right skills.

https://twitter.com/htormey/status/1762864807995416584?s=46&t=Bc9m1uknKGxrhqisx2EO2g
Thinking that the next article I’m going to write will be about llm evaluation frameworks. Currently digging into docs.ragas.io. Anyone have any they would recommend? Interesting aside, I’ve come across multiple promising python frameworks and no JavaScript ones.
/AI