Chris McCarty

We're hiring for Rails engineers on our team again, which is exciting! What resources do you like/recommend for people who may not have experience w/ Ruby/Rails but are experienced w/ other Web technologies? @eileencodes, @searls, @tenderlove (who else to ask?)

@bcherny Amazing! How quickly could an agent add a feature for the double tap hotkey desktop app to open code mode vs default? Bonus points to trigger specific subagent/flow

Released today: /loop /loop is a powerful new way to schedule recurring tasks, for up to 3 days at a time eg. “/loop babysit all my PRs. Auto-fix build issues and when comments come in, use a worktree agent to fix them” eg. “/loop every morning use the Slack MCP to give me a summary of top posts I was tagged in” Let us know what you think!

@danlovesproofs I find it helpful for communication and records (short term). It’s not hard to adjust an ai flow to think, gather context, save to system and then start.

ok i am surprised how many engineers from linear and atlassian are liking this.

@levelsio What is your Claude setup like? Custom skills and agents, plugins? Or just vanilla?

This week I decided to just permanently switch to running Claude Code on the server mostly on bypass permissions mode: c() { IS_SANDBOX=1 claude --dangerously-skip-permissions "$@"; } And for the first time in my life I think I've actually managed to outrun my todo list What happened is I simply blasted through my to do list of features I had to build and bugs I had to fix I've never shipped so fast and Claude Code almost made no mistakes, and when it did it they were tiny that weren't fatal (important because I'm mostly working on the server in production now) Before I was always known to ship fast (also because I always work alone) but while I shipped new things would always build up on my features/bug board (my users can submit them there) But this is the first week where I've been fast enough to outrun them The board is actually empty! As other people have written on here the real bottleneck is becoming myself and my creativity, not how fast I can ship. Because I think I ship faster now than I can come up with new ideas, or maybe my brain will adjust to this new speed (probably) Also I feel another limit is becoming my own mental context window, as in how many things, features, bugs, projects, I can keep in my mind in parallel while building on all of them. It's a lot and I haven't reached that limit yet but I feel I might be close I also noticed that you start going really fast the more you let it just go loose, before I was slow because I didn't trust it and I was scared it would destroy my code, now I just let it go. As @karpathy wrote, things feel like they've changed a lot around December last year when models became good enough to really code with and I feel the same When I see other friends code with Claude Code I often notice they're slow because they still check everything, which is good of course, but I feel the better way would be to create some tests and just let it run freely and see if it can pass those For me the tests are mostly just me checking out if the new feature on the site works or not, and in 99% cases it just does, and then I ask it to improve it further Because I run Claude Code on the server in production, I don't have to wait for deployment anymore (although that took only 3 seconds anyway before, that still adds up), now it's wait for it to be done coding, I refresh the site and I test it, that feedback loop is how I work and it's made me WAY faster Anyway here's what I did this week and the majority of these things were requested by people on the bug board, I'd say this is about 10x my normal output: 📸 Photo AI - Built new image viewer and mobile image viewer - Added batch remix, multi-photo import, filtering by model in gallery - Security overhaul: phased out insecure ?hash= login, migrated to session tokens - Fixed Google login loop, multi-model selection, talking scripts - Added custom audio upload for talking videos - Created dynamic model selector from server endpoint 🏡 Interior AI - Revived [ Add furniture ] feature (started 6 months ago, image models now good enough) - Added custom style upload for redesigns - Built own Gaussian Splat viewer for 3D - Made /remove_bg endpoint for furniture backgrounds - Migrated 3D walkthrough to new World Labs API - Added .skp file support, paint color masking, empty room button 🎒 Nomads - Launched weekly AI-generated newsletter from chat - Built profile edit modal, moved profile editing from /settings to profile page - Added TikTok/YouTube links, status bar, server-side API tracking - Added hundreds of new profile tags and traits - Fixed timezone filters, broken links, user avatars 🗺️ Hoodmaps - Revived write mode (before was only read for last few years because db was rekt) - Built heatmap mode using sentiment-scored tags (50K+ tags) - Fixed root cause: tags not entering DB due to wrong PRAGMA (should be WAL) - Added good/bad area detection with admin grid controls - Set up Claude Code Telegram bot for live changes - Enabled CF cache, fixed health check, fixed Brussels 📕 MAKE book - Built auto ePub/PDF generator cron worker - Added dynamic generation with personal customer watermarks - Added image compression for file size 💾 Pieter .com - Added Wikipedia text-only reader for Kindle - Exploring Windows 3.11 emulator using v86 (to replace Em-DOSBox) - Added product recommendations on homepage - Installed Wall Street Raider (1986) 👩‍💻 Remote OK - Installed Chatbase AI customer support bot - Added "report not remote" link on job posts 🏨 Hotelist (3 todos) - Fixed hotel URLs and city range bugs - Added iron amenity

Cool AI moment in a 1:1 yesterday: I was chatting w/ an engineer about a potential edge case, and we were trying to decide how/when to prioritize. I said "let's try this": and hit a hotkey to record my voice, reiterated the issue, hit enter, and 5min later we had PR to approve.

@mcuban Trajectory is another thing to consider. If it gets 80% cheaper and/or 5x productive every few quarters

This is the smartest counter I’ve seen to ai taking over jobs, in the short term. Is the ((aggregate tokens cost to do what an employee does + plus fully encumbered developer and maintenance costs ) / (fully encumbered employee cost ) )<= productivity ? If it takes 8 Claude agents, at $300 for tokens, per day, plus $200 per day in dev/maint , to do what an employee does per day, at a fully encumbered cost of $1200. That’s 2600/1200. But then you need to factor in the productivity rate. Is it more than 2.16 x productive ? Are there qualitative issues like morale, morality, whatever , that can’t be quantified, that need to go into the decision? What is the going forward progression of burdened costs for the tokens ? Curious what people think about this ?

@SahilBloom AI summarize this for me in a single emoji so I don’t need to read or learn it

@justinskycak I put my workout plans on the calendar. Easier to just do the workout than regretfully/sheepishly delete the calendar event.

Every move that a chess player makes functions to get the board in a state where good plays emerge. Athletes do the same thing -- "skate to where the puck is going to be" type of stuff. But few people realize that this is a general strategy to apply in daily life. So much of making good choices comes down to putting yourself in a position where it's easier to make the good choice. In life you get points for the good choices you make, not the difficulty you overcame to make those good choices. So put your environment on easy mode. You don't get any extra points for having your environment on hard mode. All that happens is your performance suffers. Doing your best means not just doing your best in each moment, but also doing your best to put yourself in a position where it's easier to make the best choices. Personally, I know that my likelihood of exercising each day is inversely related to 1) how long it's been since I woke up, and 2) how much effort it takes to begin the exercises. If evening comes and I haven't exercised, I'm not going to do it. If I shower and start my day without exercising, I'm not going to do it. If I have to exit my apartment and drive to a gym, I'm not going to do it. So I exercise in the morning, before I shower, before I start my day, and I do it at home. And that works great to keep me on the rails exercising consistently, every day. You can play the same kind of strategy with anything.

@minordissent Hmm idk. Kids are some of the most interesting people…

The reason most ambitious men don’t find time with their kids that appealing is because kids are honestly pretty boring. the solution is to make hanging with them exciting. Mostly by doing sufficiently stimulating (usually dangerous) things together. Wrestle. Throw them into the air as high as you can. do tricks on a jungle gym. take the car to an empty parking lot and do donuts. etc. You and them will have a blast.

Lots of people posting this info since it’s now public. I haven’t seen anyone who has been on X longer than I have. Can anyone beat May, 2013?

@stephsmithio @Olohiremee Depends on context, but a varied team can have more strengths than 1 person. Also this means the team just executes but can’t improve as it grows because it would be constrained my the managers individual abilities. Taken to the extreme: this means the CEO can do *every* job?

@Olohiremee How come? Genuine q, why wouldn’t you want a capable boss?

No pls

@petergyang Aura digital frame

What purchase under $500 has significantly improved your life (or made you happier)?

@aakashgupta That month has 44 days in it

Show up:

@GeorgeMayer Right

@cmccarty Like ai pr solutions?

The bottleneck is code review, so let’s say (generously) that you review 200 lines every 10 minutes. In an hour you review 1200 lines. In ten you review 12k! You have not paused to prompt, test, debug, go to the bathroom.

@ctjlewis I think the point is still valid—It’s a game changer. An idea that might take 2 weeks of x deep work sessions can be an afternoon while sipping a beer. Maybe it’s not enterprise prod level (not everything needs too be), but it’s becoming quicker to build and test than to spec

sorry Paul, but you guys still do not understand anything about codegen. totally bullshitting. nicest way i can say it

@nateberkopec Of course AI wrote my bash / git log scripts and the python/pandas code to graph it… Still have to think a lot and give detailed instructions but I think it’s a good think more time is reinvested there vs typing code.

@nateberkopec I ran some numbers this week (caveat that new lines, net lines, etc all have their flaws) and have seen 3x since leaning AI first and pushing through even when it felt slower (compared to copilot mode that is). A lot might be tests, comments, etc but that’s important too.

Everyone’s shooting too high with this shit. LLMs only need to make you 50% more productive in order to be the most significant innovation in programming since FORTH. This 10x, 5x etc stuff is just social media noise. You don’t add to the conversation with hyperbole.

@_HMSP How much was a cheeseburger with added bacon?

I just shook my head, laughed, and walked out Y’all, an incompetent front of house will kill a restaurant far faster than subpar food because people may forget a disappointing meal, but they never forget terrible service. 4/4

Went to a new burger joint. Cheeseburger: $10 Bacon cheeseburger: $12 Order bacon cheeseburger Waitress: “FYI, bacon is extra” “Well, yeah, it costs $2 more” “No, there’s an upcharge” She says getting bacon on a bacon cheeseburger is an extra $2.50 on top of the price 1/4

@samuelstroschei @YeetMeTwice @sanjeevn72 Why stop there and not just go all the way to family as a service?

@YeetMeTwice @sanjeevn72 It's unsustainable. short term adjustments of sleep/work schedules are doable, but not over multiple years. take kids as an example. bringing or picking up kids from school at a certain time is a fixed schedule that cant be shuffled at someones will

That was me. I said that. Time zones played a huge role, though. Here’s the full story 👇

@SergeyStavisky @McKnightFdn Loving column D

I’m very honored to join the remarkable @McKnightFdn Scholars community! With bonus joy to see so many friends in this year’s cohort (and amongst prior awardees - yay lifetime community and meetings!). mcknight.org/news-ideas/202…