Last Week in AI· Folge 235

#196 - Nvidia Digits, Cosmos, PRIME, ICLR, InfAlign

Montag, 13. Januar 2025 · 1 h 47

MP3

Shownotes

Our 196th episode with a summary and discussion of last week's* big AI news!

*and sometimes last last week's

Recorded on 01/10/2024

Join our brand new Discord here! (discord.gg/nTyezGSKwP) discord.gg/nTyezGSKwP

Hosted by Andrey Kurenkov (twitter.com/andrey_kurenkov) and Jeremie Harris (twitter.com/jeremiecharris).

Feel free to email us your questions and feedback at contact@lastweekinai.comand/or hello@gladstone.ai

Read out our text newsletter and comment on the podcast at lastweekin.ai.

Sponsors:

•
The Generator - (babson.edu/thegenerator) An interdisciplinary AI lab empowering innovators from all fields to bring visionary ideas to life by harnessing the capabilities of artificial intelligence.

In this episode:

- Nvidia announced a $3,000 personal AI supercomputer called Digits, featuring the GB10 Grace Blackwell Superchip, aiming to lower the barrier for developers working on large models.

- The U.S. Department of Justice finalizes a rule restricting the transmission of specific data types to countries of concern, including China and Russia, under executive order 14117.

- Meta allegedly trained Llama on pirated content from LibGen, with internal concerns about the legality confirmed through court filings.

- Microsoft paused construction on a section of a large data center project in Wisconsin to reassess based on new technological changes.

If you would like to become a sponsor for the newsletter, podcast, or both, please fill out this form (forms.gle/VmZDCzSBfeACe1Ke9).

Timestamps + Links:

• () Intro / Banter

• () Sponsor Break

• Tools & Apps

• () Nvidia announces $3,000 personal AI supercomputer called Digits (theverge.com/2025/1/6/24337530/nvidia-ces-digi…)

• () Meta removes AI character accounts after users criticize them as ‘creepy and unnecessary’ (nbcnews.com/tech/social-media/meta-ai-insta-sh…)

• Applications & Business

• () NVIDIA Is Reportedly Focused Towards “Custom Chip” Manufacturing, Recruiting Top Taiwanese Talent (wccftech.com/nvidia-is-focused-towards-custom-…)

• () AI start-up Anthropic closes in on $60bn valuation (ft.com/content/6f786914-734f-4b4e-b757-ab56d0d…)

• () Why OpenAI is Taking So Long to Launch Agents (theinformation.com/articles/why-openai-is-taki…)

• () TSMC Set to Expand CoWoS Capacity to Record 75,000 Wafers in 2025, Doubling 2024 Output (trendforce.com/news/2025/01/02/news-tsmc-set-t…)

• () Microsoft 'pauses construction' on part of data center site in Mount Pleasant, Wisconsin (datacenterdynamics.com/en/news/microsoft-pause…)

• () Google folds more AI teams into DeepMind to ‘accelerate the research to developer pipeline’ (techcrunch.com/2025/01/09/google-folds-more-ai…)

• Projects & Open Source

• () Cosmos World Foundation Model Platform for Physical AI (arxiv.org/abs/2501.03575v1)

• () Microsoft releases Phi-4 language model on Hugging Face (artificialintelligence-news.com/news/microsoft…)

• Research & Advancements

• () PRIME: Online Reinforcement Learning with Process Rewards (marktechpost.com/2025/01/04/prime-an-open-sour…)

• () ICLR: In-Context Learning of Representations (arxiv.org/abs/2501.00070)

• () Do NOT Think That Much for 2+3=? On the Overthinking of o1-Like LLMs (arxiv.org/abs/2412.21187)

• () METAGENE-1: Metagenomic Foundation Model for Pandemic Monitoring (arxiv.org/abs/2501.02045)

• () TransPixar: Advancing Text-to-Video Generation with Transparency (arxiv.org/abs/2501.03006)

• () The amount of compute used to train frontier models has been growing at a breakneck pace of over 4x per year since 2018, resulting in an overall scale-up of more than 10,000x! But what factors are enabling this rapid growth? (epoch.ai/data/notable-ai-models#training-compu…)

• Policy & Safety

• () InfAlign: Inference-aware language model alignment (arxiv.org/abs/2412.19792)

• () Mark Zuckerberg gave Meta’s Llama team the OK to train on copyrighted works, filing claims (techcrunch.com/2025/01/09/mark-zuckerberg-gave…)

• () Anthropic gives court authority to intervene if chatbot spits out song lyrics (arstechnica.com/tech-policy/2025/01/anthropic-…)

• () US government says companies are no longer allowed to send bulk data to these nations (techradar.com/pro/security/us-government-says-…)

• () Trump announces $20B plan to build new data centers in the US (techcrunch.com/2025/01/07/trump-announces-20b-…)

See Privacy Policy at art19.com/privacy and California Privacy Notice at art19.com/privacy#do-not-sell-my-info.

0:00 / 0:00