James J. Womack

I've learned a lot on Twitter these past 10 years. Thank you to everyone I've stalked & learned from :). Especially in the early days when I was developing solo in San Diego, Twitter allowed me to learn how world-class developers & designers communicated & thought about problems.

I dig this comment from realcygnus youtube.com/watch?v=6oHNzc…

@RobertStock6 I'm excited for you @RobertStock6. I *am* a long-time software engineer, but the breadth and scale of the baseball stats projects I can now achieve on my own—it's been an absolute game changer.

I’m a baseball player. I’ve never written a single line of code in my life. But over the last few months, I used AI to build a pitching analytics platform from scratch. 8.9M pitches. Custom ML models. Total transparency. Here is a look at what happens when you give an interested athlete access to AI:

most steals by 30+ homer sluggers from california ballparkgenius.com/search?q=most%…

@4DogFart @DNIGabbard apnews.com/article/fact-c…

Contrary to the blatantly false and slanderous accusations being made against me by Members of Congress and their friends in the propaganda media, the Office of the Director of National Intelligence has and will continue to take action under my statutory authorities to secure our nation and ensure the integrity of our elections. My response to Congress:

Amazing 90s style website for a tennis club in Borrego Springs thecourts.net

In a team meeting today, one of my engineers asked "What are those things that *aren't* AI?". The words they were grasping for were "humans" or "human users". So great a % of computer application usage is initiated by AI agents, that usage by a human becomes aberrational.

@LangmanVince @elonmusk You have data that shows otherwise?

Please reply to this post if you’re not receiving 𝕏 notifications or if the notifications are delayed more than a few minutes

@greg16676935420 @marcrandolph He doesn’t work at Netflix now and didn’t then either

@marcrandolph Do you still agree with this?

Hi. I'm Marc Randolph, co-founder and first CEO of Netflix. Ask me anything.

@_SDot22 Definitely bumped up the chances a good bit

Glasnow out and Heaney in. I can smell #61 coming soon for Cal.

@FoulTerritoryTV The engagement thirst is tangible

If the season ended today, does Tanner Scott have a spot on the Dodgers playoff roster?

ProTip™ If you want to manage or run two different version/branches of a repo at the same time, use git worktree instead of copying the directory or cloning it twice git-scm.com/docs/git-workt…

@akshay_pachaar What makes memory human-like? The human ability to store near infinite memories is nice, but not a single one of those memories is lossless.

knowledge graphs are INSANELY good at giving agents human-like memory! today, we're building an MCP-powered memory layer that can be shared across all your AI apps like Cursor, Claude Desktop etc. it's built using a real-time knowledge graph. 100% open-source and self-hosted.

Today we're changing how people build for the web. Introducing MagicPath Libraries. Your component library is a living doc that AI uses to design in your style. No more slop. Build with defaults or custom libraries, share them with teams. Always synced with your code.

@LakeShowYo The Lakers aren’t in the NFL

Who are the Lakers of the NFL? 🤔

@ThePrimeagen yes, as did mine—the “drop” hit different

my day looked very much different

@tyler_agg anyone interested in creating great experiences, great art, great company culture etc.—they wouldn't be trying to automate 100% anyway. anyone who's used automated customer support (or listened to music on suno) of any sort could tell you that

everyone's obsessed with building fully automated AI systems but if you want good outputs especially for creative tasks, you're gonna need humans in the loop AI should make things faster not fucking replace your brain entirely here's an example of what i mean: say you're building a system to generate ad images and scripts using nano banana and claude you can automate parts like: - customer research - brainstorming - creating prompts to send to claude/nano banana but before it actually generates anything have a human review the prompts and ideas first because most of the time, the output won't be exactly what you had in mind especially if you don't have really clean, structured performance data backing up those decisions and that's totally fine if it requires a human review these systems don't need to be 100% automated as creative tasks are meant to be iterative anyways and remember automating 70% of the work while still getting great results is way better than doing everything manually or having it fully automated and generating mid ass results again the goal is to make your overall process faster with minimal loss of output