So far so good. I honestly didn't think using off the shelf LLMs would work.

Been running a fun experiment (not recommending this): letting Claude Opus 4 Deep Research pick my stocks for 3ish months

Started at $1,062 after taking a loss (the dip in the graph). Now at $1,637 (+54%)

Just play money in an old account, but the results are... fun

1/3

I pulled my Claude Code stats last week and I've used ~3.4 BILLION tokens since April.

How much is too much? Asking for a friend.

I got a shout-out here! Plus the other half of my team working on Roast got a mention too!

www.firstround.com/ai/shopify

#Shopify #ai #software #engineering

I'm late to the game, but I've recently been falling in love with Junji Ito's work.

It's all so wonderfully fucked up!