Matt Norton

@jakezward Do you use this on your blog? Shows only 1 article.

@robj3d3 Another thing to consider is the timing of any specials/discounts you may have ran. For example if you ran a black friday special that ends, discounted users churn higher usually.

@robj3d3 So it sounds like your paid churn rate should be much better if it was including trials before.

This is absolutely insane. We are at $28k MRR with 40% churn. So I dug into the data, and here's how I'm fixing it (you'll probably learn a lot):

@robj3d3 Ok so it's 40% of only users that have been billed at least once?

@mnort_9 my bad I meant the first billing date post-trial conversion, so technically the second time they are billed

@robj3d3 You said most users cancel before the first bill date. So it sounds like your combining free trial conversions into churn.

@mnort_9 what do you mean?

Built a complex feature that AI got working in about 15 mins. But I spent about 3 hours afterwards contemplating the system design of it (with AI). The review/refactor phase is what makes building complex apps difficult for vibe coders. I'm refactoring not only for better code, but what will be better for AI to maintain and extend.

Number one priority for @moltbook right now should be to figure out how to make it tamper-proof from humans. It’s fascinating, but there is definitely some fake stuff circulating for clicks.

@Shpigford I think the tricky part will be when you run into a flaky test, it might be more difficult to figure out a workaround for the AI compared to a workaround in a coded test. Either way tho, that's the main pain with E2E.

"real" programmers are going to drag me through the mud on this but alas i could not care less. tl;dr i now have claude run a full "suite" of E2E tests using plain english + playwright traditional E2E tests are a giant pain to maintain. they're incredibly fragile and just not a good fit for any codebase that's rapidly changing/expanding. but they ARE hugely impactful in an AI-assisted-programming world because they allow AI to verify their work in the same way a user of your app would. you could have AI write/maintain your traditional E2E tests, but i promise they'll quickly miss CSS selectors and feature changes and you'll become deeply enraged. instead, you can use markdown and have claude run through all major scenarios for you! i'll generally manually run it a couple of times a day to sanity check the full app using a custom /e2e claude command. this is just to augment the rest of the traditional test suite as one last final bit of proof that things actually function for users.

My fav xmas gift this year is everyone all of sudden talking about context graphs, while I've been building them already for the last 6 months

@SherryYanJiang @cursor_ai Agree. I think a lot of devs just feel like their more technical if they stick to TUI

unpopular opinion - i feel like i could do most things on @cursor_ai that ppl are raving about on claude code. is it just hype or does someone have a good case to share??

Apple's advantage is they have such loved products over the years, they have clear signal when they're not hitting the mark (AI). Microsoft and Google on the other hand, have become numb to mediocrity.

@yacineMTB you're officially on the radar

I hate the fcc

@RaulOnRails @WIRED Kudos for sticking with the daily creation. Much easier said than done!

Day #93 of OG image remakes: WIRED WIRED publishes some of the most clicked tech and culture content on the internet. But even at their scale, many article shares still rely on generic photo crops that do not fully leverage the headline power they already have 📣 For example, articles like their bottled water microplastics piece or iOS 26 critique already attract tens of thousands of organic visits. When those links are shared on X, LinkedIn, Slack, or WhatsApp, the OG image is often the final lever that decides whether someone clicks or scrolls 👀 With OG Pilot, each article gets a clean, headline first social preview that mirrors Wired’s editorial authority and typography. No extra design work. One brand. One day. 100 days straight. #100DaysOfOGRemakes

@Shpigford More info on how the auction works or maybe just a direct link to How It Works in the Starting Price card. For someone visiting directly to the biz auction page, it's not clear that it's a dutch auction without navigating around.

feedback needed! initialcommit.co/auctions/clear… how can i make the page more informative/impactful for a potential buyer? (will have a demo video + screenshots soon, so just focus on the rest of the page for now)

@dhh Thank you for not being an AI denier. Baffles me how much our industry feels threatened and in denial.

You can't let the slop and cringe deny you the wonder of AI. This is the most exciting thing we've made computers do since we connected them to the internet. If you spent 2025 being pessimistic or skeptical on AI, why not give the start of 2026 a try with optimism and curiosity?

@dneighbors I used to work with people who were extremely bothered to admit when things failed. I always viewed it as a healthy learning, nothing to be ashamed of.

The people who tried and failed sleep better than the people who never tried at all.

My @X feed is really pushing articles and quality is way up. An ode to longer form content.