rogerbai

Cloud architecture | AI Startup. Opinions my own.

[Inherent @inherent_labs]

1/ Today, we introduce Faraday, a 27B-parameter AI Scientist that extends the capabilities of coding agents with a layer of scientific intuition. Trained via long-horizon RL, Faraday outperforms Claude Opus 4.8 and GPT-5.5 on the task of replicating research papers. 🧵

[Jeff Dean @JeffDean]

Announcing Discovery Loop!

I am very excited to announce that, along with my longtime friends and collaborators @Sanjay_Ghemawat, @OriolVinyalsML and @quocleix, we are founding Discovery Loop (@DiscoLoopAI), a Public Benefit Corporation whose mission is to automate machine learning, science, and engineering to accelerate discoveries and progress. The four of us have worked together for 14 to 30 years, and have helped build some of the world’s most used products, infrastructure and AI models, and we’re excited to turn our attention to this ambitious endeavor.

♾

Learn more at: https://t.co/Rv3LMdLluK

[Yu Su @ysu_nlp]

We trained a ~frontier Deep Research Agent on academic budget

> 32 H100s
> 8K synthetic samples
> fully open training infra + recipe (SFT, mid-training, RL)
> models of diff sizes (2B -> 35B) ready to use out of the box

This is yet another demonstration of how the frontier of AI is changing. We have reached a point where open models + a small capable team + a few hundred Ks can produce specialized models with ~frontier capabilities. The future of AI doesn’t have to be held in a chokehold by a handful of closed models.

We've open-sourced everything we've built and learned from this project. Hope it helps the community build more!

📌 Project: https://t.co/ZTsYE4JPqF
📌 Paper: https://t.co/UW0A9i4jA3
📌 Code: https://t.co/Bxyf12Q4j6
📌 Model Weights and Data: https://t.co/W15CLnEO1T
📌 Demo: https://t.co/pEx2FOqGZd

Amazing effort led by @jianxie_ (our 1st year student!!), Tianhe Lin, Zilu Wang. joint with @hhsun1 and the @osunlp team. thanks @amazon Xiangjun Wang for a gift that covers the compute and fruitful discussion.

[Zach Lloyd @zachlloydtweets]

https://t.co/cjyoRlzgxO

[Oliver Prompts @oliviscusAI]

this game on steam teaches math and coding better than 99% of the uni courses out there.

[Sherry Jiang @SherryYanJiang]

here are some cool companies doing stuff in world models and physical ai - anyone im missing??

@amilabs
@theworldlabs
@gen_intuition
@moonlake
@physical_int
@bifrost_ai
@reactorworld
@odysseyml
@NianticSpatial
@Figure_robot
@SkildAI
@DecartAI
@runwayml
@wayve_ai
@agilityrobotics

[AA @measure_plan]

I've open sourced this project!
- use voice commands to change interaction mode (drag, rotate, scale, animate)
- use hand gestures to control the 3D model
- drag/drop to import other models (only GLTF format supported rn)

created using @threejs, @Rosebud_AI, mediapipe computer vision, web speech API, and Quaternius 3D models

🔗 github repo below

[Alexs🥇 @Real_Alexs_]

This afternoon, I watched the recently popular movie #NeZha2.

After watching it, I felt that rather than telling a Chinese mythological story, the movie is actually reflecting the current international reality.

In the film, the Immortal Lord Wuliang talks about vanquishing demons and upholding morality, but actually engages in dirty deeds. Similarly, some countries in the international community advocate "freedom, democracy, and human rights" while actually exporting turmoil, war, and manipulating coups.

The Immortal Lord Wuliang suppresses potential rivals, even allies like the Dragon Clan, to maintain the dominance of the Chan Sect. Likewise, a certain country suppresses potential competitors, whether they are enemies or friends, to maintain its global hegemony.
#NeZha #animation

[CalCo @calco_io]

Nomic Embed Text V2 is now available

- First general purpose Mixture-of-Experts (MoE) embedding model
- SOTA performance on the multilingual MIRACL benchmark for its size
- Support for 100+ languages
- Truly open source - open training data, weights, & code
- Apache 2.0 License

[Yixin Ye @BLeavesYe]

🤔 How many examples does an LLM need to learn competition-level math?
Conventional wisdom: 100,000+ examples
Our discovery: Just 817 carefully chosen ones 🤩
With pure SFT, LIMO achieves:
57.1% on AIME
94.8% on MATH

LIMO: Less is More for Reasoning 📝
🔗 https://t.co/FE1lhsuAgA

[Alex Miller @alexlmiller]

Deepseek’s greatest breakthrough is not having a butthole logo

[Ammaar Reshi @ammaar]

Just built a Perplexity clone using Gemini 2.0 + Grounding, and the wildest part? @Replit's Agent wrote ALL the code in 2 hours!

Search anything, get sources, ask follow-ups.

Google has all the pieces to make AI search incredible. Hope they productize it soon!

Demo + code 👇

[Andy Pavlo (@andypavlo.bsky.social) @andy_pavlo]

Buckle up because we're crashing into the new year with my annual database retrospective: License change blowbacks! @databricks vs. @SnowflakeDB gangwar! @DuckDB shotgun weddings! Buying a college quarterback with database money for your new lover! https://t.co/NnFHGElFNy

[Elon Musk @elonmusk]

Absolutely true

[rogerbai @rogerbai6]

倒是跟第一点一致，也是一家都没有投😀😀😀

[卫斯理 @imwsl90]

到年底了 ，看到朱啸虎年初对AI领域的一些讲法，你同意多少？我觉得很多看的很准的

[rogerbai @rogerbai6]

Pretty good Tech Stack:
CrewAI: The leading multi-agent platform
Weaviate: The AI-native database for a new generation of software
Groq: Fast AI Inference
ExaSearchTool: The search engine for AI
WebsiteSearchTool: Built in web rag tool from crewai_tools

[Hassan @nutlope]

Introducing LogoCreator!

An open source logo generator that creates professional logos in seconds using Flux Pro 1.1 on @togethercompute.

100% free and open source. Demo + code:

[rogerbai @rogerbai6]

reliable comrade in the right place

[Elon Musk @elonmusk]

It is morning in America again

[Anderson Mancini @Andersonmancini]

Realistic Water Simulation with #threejs is live:
https://t.co/XcmCC4hQnW

Now, you can configure scenes, objects, colors, and light positions. You can also dive to see the effects from underwater 🌊💧. It works great on mobile 🤩.

The breakdown video will be on my YouTube channel soon. Be sure to subscribe so you don't miss it: https://t.co/AZNU3lXhy4

Thanks again, Evan Wallace, for the inspiration 🙏🏻
#threejsjourney #creativecoding #r3f

[Vaibhav (VB) Srivastav @reach_vb]

Fuck it - it’s raining smol LMs - SmolLM2 1.7B - beats Qwen 2.5 1.5B & Llama 3.21B, Apache 2.0 licensed, trained on 11 Trillion tokens 🔥

> 135M, 360M, 1.7B parameter model
> Trained on FineWeb-Edu, DCLM, The Stack, along w/ new mathematics and coding datasets
> Specialises in Text rewriting, Summarization & Function Calling
> Integrated with transformers & model on the hub!

You can run the 1.7B in less than 2GB VRAM on a Q4 👑

Fine-tune, run inference, test, train, repeat - intelligence is just 5 lines of code away!