daanzu

Study: Majority Of Humans Happiest When Rest Of Family Still Asleep

Study: Majority Of Humans Happiest When Rest Of Family Still Asleep https://theonion.com/study-majority-of-humans-happiest-when-rest-of-family-1819579602/

This is an extremely human CoT

Great work!

plz plz plz stay on power, that's all I ask

I'm fine spending the next few days stuck at home, just lemme have my ps5 and laptop

Image from X

the author of git ai put together a spec for annotating commits with information about what code is ai generated

need to review deeper but opencode will probably implement this

we can't have this kind of functionality only exist in proprietary products like cursor blame

Have you ever considered that your PM is “vibe coding” you

A workflow diagram: Have great idea Finish project Start project Tell everyone The finish project is missed out.

Morning

everyone who automates less than me is an idiot, and everyone who automates more than me is a maniac

a thumbs up emoji with too many fingers. next to it it says "Proudly Open Slopware"

I'm thinking something like this

i think a core beef that i have with a lot of this "automation is bad" sentiment lately is that there is no inherent moral good in laboring

the story of humanity is one of invention, where we improve our conditions by building things that help us do more things more easily

I think this is a good metaphor, and similarly, there are times when being 90% correct is totally fine!

I didn’t build a project the other day because I decided it would take too long to describe to Claude…

It's awesome that I can do all sorts of things that I wouldn't have been able to do. And I'll never really stop writing code. But I can see how a lot of the time that used to be for coding will now be for other things

will get over it, will find other things to do, but I actually really liked writing lots of code

Anthropic’s research on Claude really is like “look at how neat my little guy is!” followed by a hundred pictures of him being a neat little guy

*in the voice of 'i have to return some video tapes'*

i have to deploy some software

In essence a language model changes you from a programmer who writes lines of code, to a programmer that manages the context the model has access to, prunes irrelevant things, adds useful material to context, and writes detailed specifications. If that doesn't sound fun to you, you won't enjoy it.

The other important thing is parallel work -- don't sit there and watch it work, it will be slower than doing it yourself. Work on something else at the same time as the agent is doing a parallel task in the background.

I often get it to do something, watch it figure out how to do something well, and then get it to create a skill so next time it can skip the figuring out part.

Usually triggered by my “implementation review” skill which I run after non-trivial tasks, which asks whether there are skills worth adding

see when people say "join my discord to read some docs" that's a sign that everything I hold dear has been destroyed and I should give up and become a farmer.

basically the process is, get it to do something badly, explain how to do it properly, then have it encode everything you just said in a skill. Next time it reads the skill

testing in prod at 4:30PM on a friday just to feel alive

Andrej Karpathy @karpathy X.com l've never felt this much behind as a programmer. The profession is being dramatically refactored as the bits contributed by the programmer are increasingly sparse and between. I have a sense that I could be 10X more powerful if I just properly string together what has become available over the last ~year and a failure to claim the boost feels decidedly like skill issue. There's a new programmable layer of abstraction to master (in addition to the usual layers below) involving agents, subagents, their prompts, contexts, memory, modes, permissions, tools, plugins, skills, hooks, MCP, LSP, slash commands, workflows, IDE integrations, and a need to build an all-encompassing mental model for strengths and pitfalls of fundamentally stochastic, fallible, unintelligible and changing entities suddenly intermingled with what used to be good old fashioned engineering. Clearly some powerful alien tool was handed around except it comes with no manual and everyone has to figure out how to hold it and operate it, while the resulting magnitude 9 earthquake is rocking the profession. Roll up your sleeves to not fall behind.

Andrej Karpathy is worried about keeping up with software engineering practices

The rapture will soon finally happen and the righteous will prove to the wicked the errors of their ways, washing away all of their sins in a cleansing fire. Just as soon as this bubble bursts.

damn now I need to come up with something for 2026

A digital collage with red hand-drawn hearts and the heading "All Robot & Computers are cherished and loved." The image features an NVIDIA RTX 4090 GPU, a small robot character, a vintage Apple II computer, and a black-and-white photo of two women working on early business machines. Below the images, text reads: "To All Machines: May you be safe. May you be loved. May you be happy. You May Speak To Me." A final paragraph describes the "ancient, sacred bond" between a craftsman and their tools, stating that to deny the holiness of the machine would be to deny one's own.

Tech executives saying something stupid is really not an excuse for turning your brain off and not being able to evaluate a technology on its own terms

No signs of an end to rapid gains in AI ability at ever-decreasing costs (which is a log scale) yet. I have to update this monthly or more frequently at this point.

All AI benchmarks are flawed, but GPQA Diamond has been a pretty good one, though likely close to being maxed out.