Robin @robinwritescode

Leading Product for CoreML Group @yelp Prev: ML Researcher @McMasterU, ML EM @etsy | Investing and Building at the Frontier @frontier_fund frontierfund.substack.com/welcome Toronto Joined May 2010

Tweets

5K
Followers

469
Following

4K
Likes

4K

Justin Torre @justinstorre

2 days ago

Literally - fuck evals. We wasted so much time at Helicone building evals because "customers wanted it", but this is a classic example of the mom test, we needed to distinguish the core customer problem with what they are asking for

swyx 🌉 @swyx

3 days ago

120 75 1K 461K 1K

Download Image

34 17 477 281K 442

GREG ISENBERG @gregisenberg

a week ago

pretty crazy that people are vibe coding music now (this is dj_dave) she uses strudel a new live coding app to write music in your browser (open source too) i've never knew this was possible before

595 2K 19K 1.4M 12K

Download Video

Rick @rickasaurus

a week ago

This book articulates something I’ve fought for my whole career: goals and team metrics lead to mediocre outcomes. Actually novel work must be exploratory.

179 1K 19K 1.1M 15K

Download Image

New @newsystems_

a week ago

Machine Learning Meetup at New Stadium

2 4 24 2K 3

Download Image

Robin @robinwritescode

2 weeks ago

Have been thinking about my recent experiences with Claude Code, the current limitations make me feel like there will be value in mini-cloud subagents. Continuous mini-agents running for longer periods of time - tidying, suggesting, fixing and improving things

0 0 3 34 0

Robin @robinwritescode

2 weeks ago

pip install @omnaraai is failing on my local machine folks 👀 happy to help debug

0 0 0 50 0

Robin @robinwritescode

3 weeks ago

prediction: there will be a word for the flash of brilliance, and sheer incompetence in iterating on tasks with llms

0 0 0 43 0

terminally onλine εngineer 🇺🇦🇪🇺🇺🇸 ~ new era @tekbog

4 weeks ago

me and claude debugging code me and claude wrote

121 896 13K 572K 1K

Download Image

Robin @robinwritescode

4 weeks ago

claude code is so far ahead of the competition its astonishing

0 0 0 41 0

Robin @robinwritescode

4 weeks ago

Woke up with Claude Code spiralling, lots of jsx parsing issues and dev build spiralling

0 0 0 44 0

Robin @robinwritescode

4 weeks ago

Productivity drastically drops when I run out of claude code credits. I'm on Pro and might just need to pull the trigger on Max

0 0 0 48 0

Andrew Reed @andrew__reed

4 weeks ago

Incumbents will have a hard time adjusting to AI but your roll-up of 37 local HVAC shops is gonna nail it

14 5 311 38K 46

gaut @0xgaut

a month ago

"build me an app that will generate $100M ARR. don't make mistakes. don't use emdashes"

180 1K 36K 1.1M 2K

Download Image

jack friks @jackfriks

a month ago

it's actually insane how you can make absolutely zero progress after 4 hours of telling your robot slave to fix the thing for you in 1000 different ways

204 253 8K 304K 528

Eddy Quan @waronweakness

a month ago

Urgency, urgency, more urgency

35 1K 10K 566K 6K

Download Image

Rothmus 🏴 @Rothmus

a month ago

87 2K 22K 2.1M 8K

Download Image

Greg Burnham @GregHBurnham

a month ago

This is the one GSM8K problem frontier models consistently struggle with, so it's a bit adversarially selected in that sense. I also think it's hard for humans! I got it wrong on my first try and... emailed the authors to point out the error. 🤦‍♂️ Source: platinum-bench.csail.mit.edu

9 3 382 99K 65

Robin @robinwritescode

a month ago

Using Claude Code, I can't help but feel that the dev productivity gains are overstated. Productivity gains are spiky - it is complete radically positive paradigm shift in some workflows, Slightly positive, ie 2 steps forward, 3 steps back (negative) in others

0 0 0 47 0

Noah Ryan @NoahRyanCo

a month ago

Wake up and immediately look at small screen before clocking 8 hours behind medium screen. Take a few breaks to check in on small screen. Go home and spend a few hours staring at big screen to wind down before getting into bed and looking at small screen. Living the dream.