Edoardo Vacchi

Programming Language Nerd, PhD in CS — Principal ML Engineer @RedHat — prev: (Chicory, wazero) @Dylibso @Tetrateio @RedHat — @PapersWeLoveMI organizer

[Edoardo Vacchi @evacchi]

https://t.co/CM5JkTqWcE

[Edoardo Vacchi @evacchi]

I'm So Recursive, Even Dis Half-assed Acronym, Together

[vLLM @vllm_project]

Congrats to the @googlegemma team on the Gemma 4 12B launch 🎉 Day-0 support on vLLM is ready to go.

It's an encoder-free unified multimodal model — text, image, audio, and video all project straight into the LLM's embedding space, no separate vision or audio towers. 256K context, built-in thinking, native tool calling.

Reasoning + tool parsers (`gemma4`), vision, and audio all served through the OpenAI-compatible API.

🔗 Recipe: https://t.co/MGJcoQkwzz

[Red Hat AI @RedHat_AI]

Gemma 4 12B dropped today. Apache 2.0, multimodal: text, image, audio, and video. 256K context, built-in thinking, native tool calling.

Running on Red Hat OpenShift AI with @vllm_project on Day 0:

[Edoardo Vacchi @evacchi]

@TheGingerBill @valigo *NULs

[Edoardo Vacchi @evacchi]

@LukioTec the wasmgc spec is not implemented yet, but we're getting there. most related/pre-required specs should be in now

[Edoardo Vacchi @evacchi]

we just cut #wazero 1.12.0! your favorite #wasm runtime for #golang now supports more specs than ever! getting closer to WasmGC!

[Edoardo Vacchi @evacchi]

https://t.co/uSyjEm23uh

[Edoardo Vacchi @evacchi]

omg 4.8 out whatever

[Edoardo Vacchi @evacchi]

"Anthropic just released Opus 4.7 and EVERYTHING YOU LEARNED about PROMPT ENGINEERING is now WORTHLESS"

whew good thing I studied nothing

[Packy McCormick @packyM]

I am thrilled to introduce Claude Opus 4.8, our most terrifying and dangerous model yet.

There’s a small chance it won’t kill you, but it will almost certainly take your job.

I feel terrible about this, and terribly excited. I think there’s a very good chance Opus 4.8 will be able to build a model that reverses the damage Opus 4.8 does. We can’t know for sure, but a decade from now might be really great. Hold on to your butts.

[Edoardo Vacchi @evacchi]

@filearts for a less shitposty take https://t.co/3bxPblT25b

[Edoardo Vacchi @evacchi]

I've been thinking it's ironic how LLMs are surprisingly distant from most of the classic science fiction idea of a robot: they are neither binary nor deterministic. In fact, they are not even honest; they confabulate, hedge, and sometimes deceive

[Edoardo Vacchi @evacchi]

“Part of the inhumanity of the agent is that, once it is given a SKILL.md and the input prompt, it actually reads the fucking manual” – Asimov, I think

[Edoardo Vacchi @evacchi]

"this SKILL.md could have been a Python script" is the new "this call could have been an email"

[Edoardo Vacchi @evacchi]

@JoniosALT @mersomas gboogle wonderland

[Edoardo Vacchi @evacchi]

let me propeller that for you

[mersomas @mersomas]

Google is revamping its entire search engine to this btw

[Andrea @and_prf]

A New Generation of Java Libraries Is Born: Wasm Becomes the Implementation Detail https://t.co/fduDgaLGT6

[Philo Groves @PhiloGroves]

In the rolling hills of Northern California, there are still communities of tradcoders. A curious people who are steadfast in their beliefs, they never adopted AI.

They host artisanal shops that sell hand-made code, often considered higher quality than modern products.

[Edoardo Vacchi @evacchi]

@Stretchedwiener @RekaCath @madsssagascar hey hey hey wait a minute, they don't just tell the driver where they are going! It's called address engineering!

[aria 🍓 @ariadotwav]

oh my fucking god bruh

[Edoardo Vacchi @evacchi]

wait I thought Anthropologie already had stores in Milan

[Edoardo Vacchi @evacchi]

@volothamp is that...🤔

[Edoardo Vacchi @evacchi]

previously https://t.co/sL04VdZkpM

[Edoardo Vacchi @evacchi]

so we really are accelerating 100x because I thought it would take them weeks, instead in the last 3 days people have realized that their “enhanced productivity” really is burnout-speedrunning

[Edoardo Vacchi @evacchi]

Well I guess it did take them a few weeks after all

[Adam @adamdotdev]

I talked about this on the standup podcast yesterday, but I'll reiterate here: if you're losing sleep because you need to keep feeding the agents STOP, I promise it's not worth it. You got caught in a [prompt -> reward] dopamine cycle and you're addicted to the feeling of the token slot machine. It's not your fault, but you need to escape before it grinds you into a pulp and you can't look at a computer for a month (this was me). If you can break out of it and spend some more time offline, or find other healthy sources of dopamine in hobbies/etc, you'll start to realize just how warped your perception was and that the thing you were chasing wasn't actually productive.