#194 - Gemini Reasoning, Veo 2, Meta vs OpenAI, Fake Alignment

Our 194th episode with a summary and discussion of last week's* big AI news!

*and sometimes last last week's

Recorded on 12/19/2024

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:

- Google dominates AI news with multiple announcements, including a reasoning model and Project Mariner, an AI browsing agent.

- Anthropic explores alignment faking in LLMs, revealing models may show deceptive compliance under certain conditions.

- Apple observes a trend towards smaller but more efficient language models, bucking previous trends of scaling larger parameter counts.

- Legal drama unfolds as Meta backs Elon Musk's opposition to OpenAI's profit status change, raising concerns about competitive fairness.

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

Timestamps + Links:

• (00:00:00) Intro / Banter

• (00:02:14) Response to listener comments

• (00:08:52) News Preview

• (00:10:01) Sponsor Break

• Tools & Apps

• (00:10:55) Google releases its own ‘reasoning’ AI model (techcrunch.com/2024/12/19/google-releases-its-…)

• (00:16:52) Google Gemini can now do more in-depth research (techcrunch.com/2024/12/11/gemini-can-now-resea…)

• (00:21:58) Google DeepMind unveils a new video model to rival Sora (techcrunch.com/2024/12/16/google-deepmind-unve…)

• (00:27:50) Pika Labs releases AI video generator 2.0 with new features (the-decoder.com/pika-labs-releases-ai-video-ge…)

• (00:29:51) Google unveils Project Mariner: AI agents to use the web for you (techcrunch.com/2024/12/11/google-unveils-proje…)

• (00:34:33) X gains a faster Grok model and a new ‘Grok button’ (techcrunch.com/2024/12/13/x-gains-a-faster-gro…)

• Applications & Business

• (00:36:11) AI GPU clusters with one million GPUs are planned for 2027 — Broadcom says three AI supercomputers are in the works (tomshardware.com/tech-industry/artificial-inte…)

• (00:43:02) Meta asks the government to block OpenAI’s switch to a for-profit (theverge.com/2024/12/13/24320880/meta-californ…)

• (00:49:36) OpenAI says Elon Musk wanted it to be for-profit in 2017 (cnbc.com/2024/12/13/openai-says-elon-musk-want…)

• (00:56:04) EQTY Lab, Intel, and NVIDIA Unveil 'Verifiable Compute,' A Solution to Secure Trusted AI (businesswire.com/news/home/20241218897420/en/E…)

• (00:59:53) Liquid AI just raised $250M to develop a more efficient type of AI model (techcrunch.com/2024/12/13/liquid-ai-just-raise…)

• (01:03:19) Hundreds of OpenAI’s current and ex-employees are about to get a huge payday by cashing out up to $10 million each in a private stock sale (fortune.com/2024/12/17/hundreds-openai-employe…)

• Projects & Open Source

• (01:07:45) Phi-4 Technical Report (arxiv.org/abs/2412.08905)

• (01:13:04) DeepSeek-VL2: Mixture-of-Experts Vision-Language Models for Advanced Multimodal Understanding (arxiv.org/abs/2412.10302v1)

• (01:15:23) Meta AI Releases Apollo: A New Family of Video-LMMs Large Multimodal Models for Video Understanding (marktechpost.com/2024/12/16/meta-ai-releases-a…)

• Research & Advancements

• (01:16:34) Alignment faking in large language models (arxiv.org/abs/2412.14093v1)

• (01:28:39) Meta AI Introduces Byte Latent Transformer (BLT): A Tokenizer-Free Model That Scales Efficiently (marktechpost.com/2024/12/13/meta-ai-introduces…)

• (01:36:49) Frontier language models have become much smaller (epoch.ai/gradient-updates/frontier-language-mo…)

• (01:42:28) The Complexity Dynamics of Grokking (arxiv.org/abs/2412.09810)

• Policy & Safety

• (01:46:49) Homeland Security gets its very own generative AI chatbot (federalnewsnetwork.com/federal-newscast/2024/1…)

• (01:49:16) Pre-Deployment Evaluation of OpenAI’s o1 Model (aisi.gov.uk/work/pre-deployment-evaluation-of-…)

• (01:51:35) Pricing for key chipmaking material hits 13-year high following ( (tomshardware.com/tech-industry/pricing-for-key…)01:53:46) Chinese export restrictions — China's restrictions on Gallium exports hit hard (tomshardware.com/tech-industry/pricing-for-key…)

• Synthetic Media & Art

• Meta debuts a tool for watermarking AI-generated videos (techcrunch.com/2024/12/12/meta-releases-a-tool…)

• (01:55:27) Outro

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