There were thousands of people cyberbullying some(<10) researchers and most of researchers did not even involve any business activity. I will yield this space to your people if I can’t even blame such online violence.

Junhology@junholau.bsky.social · 2y ago

Well… considering the reason why other people moved to bsky from Xitter, you can't really blame them (us). Like, artists and writers are trying to protect their works. It's natural that they are getting more defensive. Meanwhile many of you guys act like, idk, like it doesn't matter.

去年11月曾经是很好的机会能让ML/AI社区迁移到bsky。很多我熟悉的人那时候都在这边。但bsky充满了反ML/AI的气氛，官方毫无阻止策略，现在大部分人都回到了X。当然比起来fedi/长毛象来说，bsky还稍微好一些。我至今不能理解为什么越去中心化的技术越能聚集反AI,反互联网甚至反科技的人群，而且同时发生在中文和英文。我也一样降低了这边的活跃，恢复了X的更新。

Alexander Doria@dorialexander.bsky.social · 2y ago

I have to admit that except for the occasional cross-posting, not sure I’ll keep very active on here. There was a big opportunity to migrate the core AI/ML community that was unfortunately squandered with all this hf dataset backlash.

There are so many AI research and developer was trying to migrate to bsky back to Nov 2023. But the anti ML/AI sentiment has pushed almost everyone back to X. Bsky is not the worst, it's far better than Mastodon/Fedi. the most surprising thing is the trend happens both in English and Chinese.

Ricardo Tavares@rjtavares.bsky.social · 2y ago

I think it did. People like @howard.fm were trying out and liking bluesky, and then that happened and the anti ML/AI sentiment pushed them away.

geospy.ai AI is changing everything...

GeoSpy | Find a Photo's Location Instantly

Easily find a photo's location with GeoSpy, the AI-powered tool for precise photo geolocation. Perfect for investigations, OSINT, and discovering where photos were taken.

geospy.ai

Ross Ulbricht被川普赦免了。我当然不会改变什么看法，但是民主党实在是没用，这种本来该在民主党总统任内做的事情竟然留给了川普。

deepseek r1真的是太震撼了，我觉得每个人都应该去试试看，现在网页上是免费的，api要付钱。

Project Digits is $3000 device with 128gb unified memory! This is insane, shut up and take my money! techcrunch.com/2025/01/06/n...

Nvidia's Project Digits is a 'personal AI supercomputer' | TechCrunch

At the 2025 Consumer Electronics Show in Las Vegas, Nvidia unveiled Project Digits, a device it's calling a 'personal AI supercomputer.'

techcrunch.com

b站产生了一些和youtube完全不同的生态，蛮有趣的。最近发现的一个赛道是“提前退休，辞职回老家，用极低的价格租个房子，和一些朋友每天一起做饭打游戏”，早点退休躺平这个youtube当然也有类似表达，但是中国特殊的生态让这个行为变得非常特殊。尤其是通过快递给他们投递各种东西，每天收快递，开盲盒。主播自称自己相当于网友的电子宠物。

simonwillison.net/2024/Dec/31/...

Things we learned about LLMs in 2024

A lot has happened in the world of Large Language Models over the course of 2024. Here’s a review of things we figured out about the field in the past …

simonwillison.net

That didn't take long! Nomic AI has finetuned the new ModernBERT-base encoder model into a strong embedding model for search, classification, clustering and more! Details in 🧵

I spent too long on this.

x.com/elonmusk/sta... I guess Elon will understand that MAGA is more dangerous than leftist.

Turns out we weren't done for major LLM releases in 2024 after all... Alibaba's Qwen just released QvQ, a "visual reasoning model" - the same chain-of-thought trick as OpenAI's o1 applied to running a prompt against an image Trying it out is a lot of fun: simonwillison.net/2024/Dec/24/...

Trying out QvQ—Qwen’s new visual reasoning model

I thought we were done for major model releases in 2024, but apparently not: Alibaba’s Qwen team just dropped the Apache2 2 licensed QvQ-72B-Preview, “an experimental research model focusing on …

simonwillison.net

Good weather.

Great to see this trending on the @hf.co Hub! I also wrote a short blog on why we're building this dataset: huggingface.co/blog/davanst...

Daniel van Strien@danielvanstrien.bsky.social · 2y ago

Introducing FineWeb-C 🌐🎓, a community-built dataset for improving language models in ALL languages. Inspired by FineWeb-Edu the community is labelling the educational quality of texts for many languages. 318 annotators, 32K+ annotations, 12 languages - and growing!🌍 huggingface.co/datasets/dat...

Hugging Face shows how test-time scaling helps small language models punch above their weight https://venturebeat.com/ai/hugging-face-shows-how-test-time-scaling-helps-small-language-models-punch-above-their-weight/ #AI #SmallLanguageModels

Hugging Face shows how test-time scaling helps small language models punch above their weight

Given enough time to "think," small language models can beat LLMs at math and coding tasks by generating and verifying multiple answers.

venturebeat.com

techcrunch.com/2024/12/23/w... so actually it's Facebook 's WhatsApp sued NSO and won it, but Apple kept silent for a long term.

WhatsApp scores historic victory against NSO Group in long-running spyware hacking case | TechCrunch

A US judge ruled that the Israeli spyware maker breached hacking laws by using WhatsApp to infect devices with Pegasus

techcrunch.com

看完o3我感觉人类存在主义危机就要来了。于是我去看了看神学院的学费，竟然很便宜，要不。。。🤯

PlayStation and AMD are teaming up to infuse games with AI. The two companies are establishing a “deeper collaboration” to work on “Machine Learning-based technology for graphics and gameplay” www.theverge.com/2024/12/18/2...

PlayStation and AMD are teaming up to infuse games with AI

The work could benefit graphics and gameplay.

theverge.com

一块水冷，一块原厂风冷。加了个电风扇对着吹，风冷这个温度能稳定在77度，水冷53度。如果颠倒一下上下顺序可能差距更小。。。

New staff, use a steam deck as a laptop 🤣. (Steam deck is a 16gb unified memory device and you can run llm inference on it with rocm.)

Alright Hacker News readers, you have one job: make "Funko just called my mom" beat out "Sora is here"

wtf of them?

itch.io@itch.io · 2y ago

I kid you not, @itch.io has been taken down by Funko of "Funko Pop" because they use some trash "AI Powered" Brand Protection Software called Brand Shield that created some bogus Phishing report to our registrar, iwantmyname, who ignored our response and just disabled the domain

大统华超市新增了我小时候最喜欢的东西，拿破仑蛋糕。按道理说这应该是苏联/俄语文化区域内最流行的东西，但是网上用英语搜出来的图片都不对。而在中国，我在很多城市都吃过各种版本，奇怪的是，只有天津我小时候吃的这个版本和苏联最接近。在加拿大我找了很久，终于在乌克兰人开的超市找到了一种家庭厨房的版本，和我小时候吃的完全一样。大统华这个算是非常接近了，但蛋奶酱还是太多了，有点偏拉美版本。但可以接受了。

“知情人士表示，苹果从年初开始，就在和百度文心一言、阿里通义千问、以及月之暗面Kimi等中国AI公司洽谈合作。” It's ridiculous that Apple chose Baidu's llm instead of qwen or Kimi. Qwen is closed to gpt sota and open, but Baidu is below the average.

Infinity: Scaling Bitwise AutoRegressive Modeling for High-Resolution Image Synthesis abs: arxiv.org/abs/2412.04431 New visual autoregression framework that performs bitwise token prediction w/ an infinite-vocabulary tokenizer & classifier, a new record for autoregressive text-to-image models.

One of my biggest Ukrainian war reading sources is coming to bsky!!

WarTranslated (Dmitri)@wartranslated.bsky.social · 2y ago

Known russian channel considers that the exchange of any Russian territories, specifically Kursk, is not actually implied by the constitution, so the Kursk region will be recaptured by force "no matter how many years it takes." 🤣

It's about you. Thank you for 30 years of play 💙

30th Anniversary | Thank You

YouTube video by PlayStation

youtube.com

PlayStation has officially joined Bluesky 💪

PlayStation@playstation.com · 2y ago

It's about you. Thank you for 30 years of play 💙