Notices by Daniel (djh@chaos.social)

@obrhoff

"Edit prediction won't be free forever, but right now we're just excited to share and learn."

https://zed.dev/blog/edit-prediction

🦧

Not sure these small AI models are there yet in terms of quality / latency to run them offline. Especially when most folks don't have powerful GPUs in their dev laptops. Mac folks are probably best positioned here.

#SamAltman of #OpenAI fame on a panel discussion live right now at TU #Berlin

Very first question if folks on the panel say "please" and "thank you" to #ChatGPT.

Anthropomorphising #LLMs — let's maybe not?

Overall I would have loved for a more critical discussion than giving Sam a platform for unchallenged narratives and takes.

@tuberlin

@obrhoff Ah that sucks. I can feel (!) it's getting worse, too, these days. But unless there are tents and needles on the streets and toilet paper getting thrown around, I would in no way compare it to SF / NY.

@obrhoff How is it comparable here in Berlin? Care to elaborate? Do you have numbers on this or is it just a feeling or perchance a hyperbolic mid-day rant? 🌝

@obrhoff 450M but not a single market; and not much access to capital. I believe those two specifically make up for most of our differences.

@obrhoff Yeah the smaller distill models you can locally are great, too, it's just important to understand that you won't get the proper deepseek-r1 everyone is hyped about.

What models to pick for running locally depends on: What you can fit in memory, if you run on cpu or gpu, if you want to optimize for latency e.g. for interactive use cases, and what kind of quality output you're looking for.

Just try a bunch and see what happens 😛

@obrhoff It's all open research

https://arxiv.org/search/cs?searchtype=author&query=DeepSeek-AI

For details on deepseek-r1 and the qwen / llama distilled models, see

https://arxiv.org/pdf/2501.12948

for the distilled model benchmark see table 5.

They're qwen / llama model architectures and different compared to their main contribution.

@obrhoff because it's not deepseek-r1.

Only the 671b / 404gb model is deepseek-r1 and most folks won't be able to run that locally.

All other variants are the qwen / llama distilled models. It's highly confusing that folks like ollama name them deepseek-r1. They're completely different models.

@obrhoff talk is cheap, send pull request 🤓

@obrhoff It's gonna be interesting what will happen to all these voluntary ESG initiatives and climate pledges at scale. Will they get cut similar to DEI initiatives and were they used for signalling only?

There's no accountability behind most of those voluntary climate pledges and in top they're in the way of pushing LLMs.

I expect lots of changes there with the new administration taking over and we're probably seeing just the beginning now.

@obrhoff I listened to this only because it's such a high staked episode and otherwise can't make myself tolerate this podcast.

But when Lex went on about a BBQ in Texas with his bro Joe Rogan I couldn't believe how cringe and delusional all of this is.

@gedankenstuecke I don't know about your grant process but I've seen EU grants where folks hire agencies to write for them turning 20 page proposals into 200 page proposals meeting the EU grant requirements. These agencies take a cut of the grant money when they succeed. In cases like those using AI really is just the 🍒 on top of the 💩

@derek There's a high chance that if you're thinking about these topics you're already doing more than most. In that case I wouldn't sweat the details or get lost in details and take a case by case decision based on how you feel like.

In the end we must remember that we need systemic change at scale holding corporations accountable. If we're stressing over every lightbulb in our every homes we're micro-optimizing only to feel better individually.

@forteller @thunderbird Some ten+ years ago there was a way to set up sieve rules in Thunderbird with an addin jf I recall correctly — and once set up that worked quite well. But it's a bit of the hacky kind.

There really should be a way ro have Thunderbird's junk scoring move mails automatically.

@forteller @inkscape This is auch a great little idea! 🤯

I believe there are already window modifiers such as "always on top" so why not having a "locked aspect ratio" one 🖼️

#Bewohnerparkausweis in #Berlin: 20,40€. Einmalig. Gültig für zwei Jahre.

Für den Preis von drei Döner mal fix zwei Jahre lang parken. #Mobilitätswende wann? #fuckcars

https://service.berlin.de/dienstleistung/121721/

@GossiTheDog The very recent zstd fork branch updates agree with the assessment that the compression ecosystem as a whole was the domain this threat actor was playing in:

https://github.com/JiaT75/zstd/branches/all

Can't make it to this weekend's #OpenStreetMap meetup in #Berlin unfortunately, so here's something I played around with a while ago instead 🤓

A map that fetches Mastodon tags dynamically as you're interacting with the map. Think: you're zooming into Berlin and content tagged with the #Berlin tag pops up! ✨

Attribution: map data #OpenStreetMap, map tiles #Protomaps, #MapLibre map sdk

#Mapstodon #Maps #Mastodon

@forteller Such an amazing achievement! 👏 Thank you for all your contributions to OpenStreetMap 🙇♂️

What happens when #Android manufacturers prefer battery life over proper functionality 😮

This is from the android issue tracker about OEMs stopping background workers unexpectedly ✋

https://dontkillmyapp.com