A good thing for frameworks like Effect, which push TS to its limits.

I don't 🥲

I do what's interesting and hope for the best 😭

What I understood: sucking wires improves sleep

👏🏻

A line chart plots global capacity of solar power measured in gigawatts. From 2010 capacity starts to grow exponentially, way beyond a series of predictions drawn as lines in yellow. Actual installations have been more than 3x higher than their five year forecasts.

The world’s solar capacity reached 1,419 gigawatts in 2023, way beyond any predictions. 1 gigawatt = power for a medium sized city

“Where’d you serve?”

“Aisle 30 — bath, plumbing, faucets, kitchens and electrical.”

When politicians talk about keeping the children safe from "pornography", this is what they mean

The Billionaire Hoarders: How the Wealthy Became Our Biggest Threat How a sick obsession is killing workers, democracy, and our planet.

Many of our billionaires are richer than any pharaoh or king in the history of the world, while a family lifestyle that could be comfortably supported by a single income in 1980 takes two people working full-time to maintain today. https://trib.al/X9l4PzM

I had pleasant experiences with form and chart code generation. Not complex, but usually very tedious work.🤔

But I also had to slap some virtual wrists to keep the AI from duplicating code all the time 🥲

Same as with image generation.

People assume you write 10-20 word prompts and get something good. Yet, it takes at least some artistic knowledge to even formulate a prompt that doesn't lead to slop.

Sure, you don't need the motor skills to draw it, but if you can't even explain what you want... 🫠

Yeah, it's ridiculous.

If a technology makes me 10-20% faster, I think it's good.

But it some people seem to think that if AI doesn't make them at least 100% faster, it's not worth it 🙄

First, simplicity over novelty. Early on, Scott Forstall told us Arc felt like a saxophone - powerful but hard to learn. Then he challenged us: make it a piano. Something anyone can sit down at and play. This is now the idea behind Dia: hide complexity behind familiar interfaces.

ah yes the piano… a famously easy instrument that everyone knows how to play

browsercompany.substack.com/p/letter-to-...

The “lets tell anyone we dont like to kill themselves” crowd are not welcome here sorry

we *will* get people to wear an electronic device that's actively surveying 24/7. this time for sure

Anyway I tried out Cursor last night and was shocked at how bad it was out of the box. And only marginally better after some tweaks.

I continue to find the complete randomness of experiences with LLMs to be pretty baffling.

Memorial Day tomorrow

Boolean all the things!!

And then bitmask them for maximum performance 😎

GitHub - permaweb/ucmWiki: A decentralized and permanent wiki for Arweave docs. A decentralized and permanent wiki for Arweave docs. - permaweb/ucmWiki

don't know how to deploy one, but here is the code:

github.com/permaweb/ucm...

Also, with Dyon there seems to be a pretty comprehensive solution for the game-scripting issue 🤔

Asynchronous Svelte · sveltejs svelte · Discussion #15845 This is a long document; get yourself a cup of tea. tl;dr You can now use the await keyword in Svelte — in your <script>, inside $derived expressions, and in your markup — by installing the async b...

Async Svelte is coming! github.com/sveltejs/sve...

And? 🤔

Most of my business opportunities come from X 😭

@svelte.dev 5 being backwards compatible is a big deal.

We migrated most of loopback.works to Svelte 5, but the things we haven't migrated yet continue to work just fine!

I can't express the gratitude I feel towards the Svelte contributors 🙏!

absolutely bodied by luxembourgish

#jj-vcs: different approach to versioning

thisalex.com/posts/2025-0...

I'm so glad Aider requires adding editable files 🫠

Is there any public @ipfs.tech node (not gateway - means something with port 5001 open)? For sure only reading stuff enabled. Would need it for a CI job where it would not be feasible to start a node every time the CI runs

Anthropic tend to do this research and then say, "See! The AI lies to us!" In reality, it is just evidence that LLMs work exactly how the skeptics claim they work (i.e., without deep understanding). CoT helps because it creates second-order token groups in attention, nothing more.

Such hard feelings in a documentation

www.gnu.org/software/gui...