Love this. Feb '27 is rapid implementation too.

Here's public @buffer.com feedback because @joel.is is ghosting me in his dm's neutral_face.jpg 1. When scheduling a post, I want to see the dates that already have posts on the date picker

I woke up early due to neighbors kids and I'm out of coffee beans. How's your monday morning?

Well yes technically I asked for a 'dashed' arrow I guess. Thanks Gemini.

"Technology makes it faster to build, but harder to care." Great read by Karri Saarinen, co-founder of Linear. linear.app/now/why-is-q...

I really hate this. Le Chat was characteristic, it screamed European. Now its just a synonym for slop really.

YO www.youtube.com/watch?v=fW-p...

Yes

Nobody is talking about this because of obvious reasons, but Russia (who historically has really good hacker groups and cyber warfare) is falling behind dramatically in the LLM/ai game. Or thats what I think (not like we hear whats going on there...) Does any1 know if Russia even has an ai "lab"?

Sooooo what LLM makes the best ASCII art? Asking for a friend. This is OpenAI's 'smartest model' trying to do a mouse:

This fully grown adult shrimp decided to change the color of her pattern from white to BLUE. Happened over the course of a week. You wont hear me complaining! #caridina #neocaridina #shrimp #skrimp #skromp

Any1 else really hating the AskUserQuestion tool in Claude Code?

Actually doing this for all current projects right now. This includes migrating from npm to pnpm for some projects. If that pulls in a supply chain attack I am just gonna quit programming lol.

Wild

Friendly reminder to pin your dependency versions and/or tell your package manager to refuse to install any versions that were published less than 24hrs ago.

FYI: If you updated your mistral sdk between May 11 at 22:45 UTC and May 12 at 01:53 UTC you might be fucked-ish. docs.mistral.ai/resources/se...

Does any1 know if there is a synthetic dataset of a fake corporation that keeps getting updated? Like MS Contoso, but with new data being added every day?

Fair point.

Fun image gen collab ^_^ worldrecordcity.com

Yasssss we need more skrimp here

I have a test coverage of 70+% on statements, functions and lines. I also yeet Claude Code's work straight into a hosted test environment without running locally to check O_o

Fucking hell. I thought Anthropic were the good guys after saying no to the Pentagon. Now they're in bed with mecha-hitler himself.

I dont feel old. But I did reply to somebody by commenting a yt video that was uploaded SEVENTEEN YEARS AGO.

Now that I think of it, its been a while since Claude Caude (Opus 4.7 for planning, Sonnet 4.6 for execution) has been 'wrong'. Sometimes technically the result is different than what I intended, but its down to pretty nuanced implementation details.

I beg you pardon?

I badly want to know what a PDS is - googling finds me many different definitions. @pds.dad whats going on here? I also want this ball to change color when Claude Code needs my input.

Yep, this is why my github actions failed yesterday. Calling it now: github wont stay free forever.

This is so good.

Google removed a pledge not to use AI for purposes like weapons development and surveillance from its ethics guidelines. Now DeepMind is unionizing to keep Google to their own ethical standards. I hope they keep on fighting the good fight. www.wired.com/story/google...

there is only one correct way to redesign GitHub and this is it

Making animated ascii for my side project #gamedev #indiegames

Lately I've been telling Claude to spawn subagents if any code exploration needs to be done in the PRD-phase, and its working really well.

Pretty cool from Anthropic to do this post mortem on the recent fuckuppery. And glad to report that things are working much better since they fixed these issues. www.anthropic.com/engineering/...

Major "oh fuck" moment. (Its not Jerome Powell)

Pretty wild Vercel only treated .env vars as 'sensitive' after opt-in lol. vercel.com/kb/bulletin/...

My god Claude Code is a shitshow atm.

Fuck. This. Shit.

I guess ill leave this terminal open for a while ¯\_(ツ)_/¯

Accurate

Getting back to this: I now have an email-update Claude Code skill that fetches the messages received since the last time that skill ran and provides a very nice summary. It updates the skill itself with a timestamp so it always fetches all messages since the last time it ran.

Also, why the fuck has Reddit (small indie company™️) not improved their "Ask"-feature since inception? Its fucking shietie.

Ok I have a bunch of email accounts I need to keep track of. I feel like its time to make my email-life better with some form of AI updates, preferably on my phone. What y'all got?

Everybody wants to be a thought leader but nobody wants to take notes while listening to their favorite podcast.

This is what developing software for clients feels like

I dont know who automated this fucking annoying account, but pls make it stop.

Gosh darnit now I cant unsee it. They even wear similar glasses for crying out loud 🤣

Slowly getting into the groove now that caffeine is entering my system. I guess I have dependencies too.

Truly great piece well worth the read.

So a few people have been integration physical devices with coding terminals (like a lamp changing color when Claude Code needs your attention). What would the ultimate physical device hooked up to event driven hooks look like? I want one that plays dead when you hit your rate limit.

Yoooo they released it