Today is OpenStack's 15th birthday! What an honor to be part of this community from the beginning. I've met so many friends all over the world and have enjoyed seeing others fall in love with open source and make careers and companies out of it, in OpenStack & beyond! To 15 more!

You're Probably Breaking the Llama Community License > If you're distributing or redistributing a LLM model that is under the "Llama 3.3 Community License Agreement", you might be breaking at least one of the terms you've explicitly/implicitly agreed to. notes.victor.earth/youre-probab...

Who are some cool open source AI folks in Paris I should meet up with?

OpenInfra Summit coming to Paris-Saclay, France Oct 17-19, 2025! OpenStack, Kata Containers, StarlingX, Zuul + many other open source infrastructure projects will be discussed along with OpenInfra for AI and the mass migration from Vmware to OpenStack Mark your calendars and practice your French!

“New infrastructure is being designed specifically for AI. GPU support has been part of #OpenStack for years, making it the standard in HPC & AI is the next evolution...users like the Dawn Supercomputer in the UK are already building AI clouds on OpenStack" - @markcollier.me tfir.io/openinfra-fo...

DeepSeek Open Source Optimized Parallelism Strategies, 3 repos [Discussion]

DeepSeek goes beyond “open weights” AI with plans for source code release https://buff.ly/4gT62Nd

Arne talked about that in one of the CERN presentations 😁 indico.cern.ch/event/150867...

Are we joy maxing?

Wallace Shawn Emerges As Frontrunner To Replace Daniel Craig As James Bond

OK! My Google colleague Thang Luong shared some exciting updates about AlphaGeometry2! AG2 now has surpassed the average gold-medalist in solving Olympiad geometry problems, w/ a solve rate of 84% compared to 54% previously! Paper: arxiv.org/abs/2502.03544 See full list of authors on link

📢 Today we're releasing a new highly detailed dataset for video understanding: HD-EPIC arxiv.org/abs/2502.04144 hd-epic.github.io What makes the dataset unique is the vast detail contained in the annotations with 263 annotations per minute over 41 hours of video.

Another seed could be a wikipedia style data curation effort. In parallel, research data-efficient models that can be trained on 100M tokens. This is comparable to both what humans need to develop speech and to wikipedia, proving that it's possible to curate this amount of data as a community.

First day at FOSDEM and I’m loving it. So much open source and interesting people. Just met a guy who wrote code for F1 to make the cars faster including physics simulation models for designing carbon fiber spoilers and better gear efficiency.

My first FOSDEM

I’d like to learn more about Merriam and Webster. What was that partnership like? I wonder if they ever fought, battled, combatted, wrestled, dueled, or skirmished?

I used the DeepSeek R1 reasoning model to prepare for a new course proposal. These screenshots show with and without the "DeepThink" option turned on--strikingly different. R1 does a lot more synthesis and offers clearer suggestions. It also accepts pdf files, o1 doesn't. Crazy this is open source.

it really tied the room together

Hello world

Major @openinfra.org 2024 highlights: surging @OpenStack.org adoption driven by AI infrastructure & digital sovereignty; successful migrations from VMware; two on-time releases enhancing AI/ML workload support; + upcoming OpenStack 15th anniversary celebration! www.openstack.org/blog/opensta...

I was interviewed at CES this morning. We mentioned that consumers will decide whether they want to be on a platform that is “unfiltered free speech “ or “ moderated”. I mentioned that I loved @bsky.app , and spontaneously people started clapping 🔥.

Does anyone have a contact for Ann ? Would love to discuss with her the possibility of partnering with her to publish her cartoons exclusively on @bluesky

Like scarves wrapped around the neck of Johnny Depp, so are the days of our lives.

While a low-temperature universe theoretically provides an ideal environment for reducing noise, the lack of available energy ironically makes reliable computation—and potentially intelligent life—unsustainable in the long run. So don't even worry about it.

My instinct is that Uber may be the wrong comparison because there's no reason for AI power to stay concentrated in the hands of a few. Every closed model has been matched in capability by an open source one within a few months, freeing anyone in the world to run it.

OH: wow, Lenny Kravitz is older than slash

stop lying: if you just died in my arms tonight, you would not be singing right now

One of my favourite words to say is "infinitesimal". You get the one-two internal rhyme of "in" and "fin" to start, and then, at the end, the delicious long-short-long rhythm of "tes", "i", "mal". Five star word

all the kings horses seem like a bad first choice for repairing an anthropomorphic egg

Nakatomi Plaza is Aubrey’s sister. Fun fact.

Merry Christmas, Bluesky!!!

An open source ai model, even if it’s half a step behind SOTA, will have a far bigger impact on the world and will accelerate progress faster

guy who shoved me at the bar last night didn't know he was pushing a trained speed walker but he sure as hell found out

Always was

Honestly, the Let's Encrypt folks don't get nearly enough credit for basically protecting the entire fucking internet, by making it absolute bog standard to encrypt everything. It happened so fast and so many people were skeptical.

An underrated feeling is when you meet someone, and you realize you have a mutual friend, so you take a picture together and send it to that friend with no context. What a rush!

same tbh

If you read nothing else about AI this week, read this - absolutely fascinating, ties together a whole bunch of recent developments in the field Wrote my own notes here but really this piece defies summarization, you should absorb the whole thing simonwillison.net/2024/Dec/19/...

good job online today, everyone. i just saw all the posts

Fully open sourced Apache 2.0 licensed physics engine is brawndo (it's got what robots crave) github.com/Genesis-Embo...

me: you dim sum, you lose some waiter (in chinese to other waiter): poison this man

"For GitHub Copilot Free users, the data collected by GitHub Copilot may be used for AI Model training where permitted and if you allow in your settings." IMHO this is probably a reasonable expectation but also good to be aware of (and I adjust settings to your liking)

Databricks raises $10b Series J at $62b valuation, the largest venture round ever. www.databricks.com/company/news...

Tired: AI is the new oil Wired: data is the fossil fuel of AI

What’s good?

Good quotes by @markcollier.me on this article, desperately trying to correctly call Open Source what is really Open Source www.cio.com/article/3616...?

Google really cooked today. Realtime AI and Deep Research are both fantastic! Still long Gemini.

Wrote up my initial impressions of the new Google Gemini 2.0 Flash model - it's really good, and the streaming mode (where you can stream video and audio to it and get audio streamed right back) is pure science-fiction simonwillison.net/2024/Dec/11/...