andersco

Because I think it’s fair to say it’s becoming a standard format for ai agent specs docs etc. It’s not going away anytime soon.

Input: this project was a complete shitshow.

Output: This project provided us with numerous unexpected learning opportunities and allowed us to demonstrate significant agility in a highly dynamic, fast-paced environment.

Watched the video. Yeah, it’s cheesy, it’s not perfect, but it’s pretty f-ing good nonetheless. IMO, this article reads like a stodgy old English professor berating a freshman essay because it wasn’t world class literature.

https://www.anders.co/

Rob Pike is my hero

Also seeing outages relating to App Store Connect, currently unable to submit builds to TestFlight.

I have found enabling the codebase itself to be the “Claude.md” to be most effective. In other words, set up effective automated checks for linting, type checking, unit tests etc and tell Claude to always run these before completing a task. If the agent keeps doing something you don’t like, then a linting update or an additional test often is more effective than trying to tinker with the Claude.md file. Also, ensure docs on the codebase are up to date and tell Claude to read relevant parts when working on a task and of course update the docs for each new task. YMMV but this has worked for me.

I didn’t read the title closely enough and thought this was Uncle Bob recommending static sites.

Yes, there is a “Hello” in the noise, made visible through contrast and animation.

For this to truly be funny, it needs to be installed on on an unsuspecting user’s laptop, preferably some C-level type about to join an in-person board meeting.

Can you provide specific examples of threads that drift into politics?

https://archive.is/aISbM

https://archive.is/VSvpv

https://archive.is/4jyjL

Still an extraordinary experience after all these years and possibly the best use of horizontal scrolling I’ve seen. Lots of previous discussions and posts on HN: https://hn.algolia.com/?q=if+moon+only+1+pixel

Yes, thanks for the clarification. One of this law firm’s main clients was one of the big tobacco companies. Also this was in the mid 90s before the big tobacco settlements.

Yes, after working as a dba for a law firm and discovering that I was the administrator for thousands of lawsuits against tobacco companies. In many cases the victim had already died of lung cancer and the suit was being brought by family members. Many of the families’ income was near poverty level. At the same time, the law firm was in a plush nyc office building. I spent day after day feeling like I wanted to throw up and finally quit.

Love shadcn, only gripe I have is I still need to use —-legacy-peer-deps due to using React 19.

I recently had an experience where I was trying out the AWS Bedrock service using their knowledge base quickStart. I followed their flow for creating a knowledge base and then played with it a bit but decided it wasn’t for me. That was in April. On May 1, I received a bill from AWS for several hundred dollars. It turns out this index I created was costing me about $16/day. I had set up a budget alarm for $20 but it was never triggered. I contacted AWS support and they were very apologetic but would not offer any kind of refund. On the one hand, sure I should have deleted those indexes after trying them out, but there was no info or any indication provided that they’d incur a daily expense even when not in use. In fact, even after I received the bill I was not able to find the source of the expense because it was described as coming from OpenSearch, which I had not used. It was only after contacting AWS support that I was able to determine that the source of the expense was Bedrock.

Fail: I tried “knulla”, Swedish for “to fuck” which it caught immediately. However, I then tried “prutta”, Swedish for farting and it did not catch that one. (Though it did flag it as Australian slang for having sex, so I guess it caught it through dumb luck.)

https://archive.is/7PL2a

I was at the talk Jeff gave earlier this evening at Cooper Union and I came away with a strong conviction of the power of this idea. To learn more, check out https://www.givedirectly.org/rural-us/

https://archive.ph/1T298

One more reason why technical interviews should be in person and, better yet, using a whiteboard only.

https://archive.is/ihdR6

Need solace? Consider that you are at this very moment traveling through space at 66,000 mph

Linus Torvalds != big tech

I’d say only complete coding challenges that are sent directly from a prospective employer or that you yourself access via a coding challenge site. I’d never click on a coding challenge link posted in a social media link.

This could be for one or multiple monitors. I’d want to ideally only have to describe my snapping rules once and then just have windows snap automatically, but also be able to tweak my instructions as I maybe discover edge cases.

Prompt-based window snapping. For example, I would describe how windows should snap based on the type of app, the number of windows and more. Example: single terminal window should snap bottom left, second window should snap next to it. Browser window full screen unless other apps are open in which case half screen left. Etc.