pngwn

(^)﹥ 🦋

[pngwn @evilpingwin]

@harrisonju it would at least be simple for the agent to generate the crops from the collected data but maybe we can have some kind of fast path for these usecases. I'll take a look and see if we can do it!

[pngwn @evilpingwin]

ML Intern in HuggingChat, is now a great place to iterate on data!

Explore existing datasets. Create your own. Iterate on them with the agent!

Try it yourself: https://t.co/JWnihMas5j

[🍉 Abubakar Abid @abidlabs]

Our thinking behind ML Intern is that there are parts of the machine learning workflow that should be automated. And there are other parts that you should continue to do yourself -- like looking at the data!

[pngwn @evilpingwin]

@harrisonju Just the bounding box coords currently but if you think the actual crop would be useful we could potentially add that as an option!

[pngwn @evilpingwin]

@probably_coding @pierrecomputer Highlights is faster as it is handwritten wat (wasm)! I haven't done a full bench but a quick test showed it was 1-4x faster.

I then wrote a quick poc of the twinkleplop tokeniser in wat and saw some speedups, so a wasm version is possible. will see.

[pngwn @evilpingwin]

@Mudit__Gupta https://t.co/OIbXodw8dM

[John Althouse @4A4133]

Sorry everyone... We're announcing a 0-Day vuln in QUIC that allows anyone to remotely turn off a server running these libraries. I swear we didn't plan this.

https://t.co/fopoKsPulV

[John Althouse @4A4133]

Sorry everyone... We're announcing a 0-Day vuln in QUIC that allows anyone to remotely turn off a server running these libraries. I swear we didn't plan this.

https://t.co/fopoKsPulV

[pngwn @evilpingwin]

@dummdidumm_ i go middle, big, then little.

- Make sure the rough architecture/ design isn't far off.
- Make it functional (do all of the things)
- add acceptance/ e2e tests
- then refine the details.

Works well.

[pngwn @evilpingwin]

After today, I'm not sure we have a security industry, more of an insecurity industry.

[pngwn @evilpingwin]

Luke strikes again with another instant hit!

[Luke Edwards @lukeed05]

Astro: 14s 🐢
bongo: 2.2s 🥁

Same 10k Markdown pages, a sixth of the memory.
Edits rebuild in ~250ms.
Incremental builds.

Introducing bongo – static sites with a good beat~!

Hit it ��
$ bun create bongo my-site

Docs: https://t.co/RrKploMrCi
GitHub: https://t.co/DPb5F8Chyp

[DeepSeek @deepseek_ai]

🚀 Introducing DeepInfer, our in-house inference framework, now fully open source!

Powering DeepSeek-V4.1-Flash at scale with high throughput and low latency. We’re sharing our production stack to help the community lower compute costs, reduce energy use, and make AI more affordable.

🔗 https://t.co/JGAik1kgk0

1/4

[hannah @erhannah]

omg hit tweet, go build something awesome for the oss folks with gr.workflow on the face hugging app <3 https://t.co/YgFcN9Syuv

[hannah @erhannah]

one day you wrote console.log("here") for the last time and didn't even know it

[hannah @erhannah]

flux 3 klein vs flux.2 klein comparison, same prompt, same seed 🐭

[hannah @erhannah]

a trendy new model drops and it's immediately butchered to death by 101 benchmarks. i call that cruella d'eval

[pngwn @evilpingwin]

Bad take. Saying that OSS is ‘arguing over minutiae’ a few days after saying a vibe coded tool is ‘harmful to OSS’ is a special kind of double standard.

And bun very recently got a trendy rust rewrite too.

Baffling.

[Stanislav @khromov]

🌭 Love the progress Bun is making! While others are arguing over minutiae or rewriting existing tools in Rust or some other trendy language, Bun just casually ships a ESLint/Prettier replacement. #bun #javascript
https://t.co/A0wv4lGG8J

[hannah @erhannah]

for when you miss the OG gradio view <3

[🍉 Abubakar Abid @abidlabs]

Do you prefer a comfy ui-style "workflow view" or a regular gradio "app view"?

Starting from @Gradio 6.30, you don't have to choose!

[Gradio @Gradio]

The model you need isn't on the Hub? Make it yourself!

ML Intern turns a prompt into a trained, evaluated model on the Hub. It plans, asks for a budget, smoke-tests, trains, evaluates and publishes.

6 models trained in a weeks time at ~$100 compute. Link in reply👇

[Luke Edwards @lukeed05]

Introducing pocketty_
SSH for iPhone & iPad, made for herdr~! 🎉

🔸 Get notified when an agent on any host is blocked
🔹 Tap → you're in that exact pane
🔸 Supports Tailscale/Bonjour natively
🔹 Diffs for every agent turn
🔸 No account or subscription

https://t.co/EoI8HTJCdX