Last Week in AI

#164 - Meta AI, Phi-3, OpenELM, Bollywood Deepfakes

Apr 30, 2024

Meta introduces smart assistants, Microsoft's Phi-3 model, Ray-Ban Meta smart glasses with AI, OpenAI's AI image generator closure, Baidu's Ernie chatbot success, Adobe Photoshop's new image generator. Intel's collaboration with Pentagon for advanced chips, Meta's increased AI spending, OpenAI CEO's investment in solar power, Google's consolidation of AI-focused teams, Microsoft and OpenAI's $100 billion project. Apple's OpenELM open source AI models, Snowflake's Arctic model launch. Deep dive into language model safety, AI chip advancements, Bollywood deepfake videos in India, LLM agents exploiting cyber vulnerabilities, China acquiring banned video chips, controversy over AI-generated voices in Drake's track.

01:31:56

Creator website

AI Summary

Highlights

AI Chapters

Episode notes

Podcast summary created with Snipd AI

Quick takeaways

Meta increases AI spending to 40 billion, emphasizing infrastructure development.

Apple releases open source AI models for on-device use under Apache 2.0 license.

VESA1 framework by Microsoft creates lifelike faces from audio inputs in real time.

Instruction hierarchy training by OpenAI enhances LLM robustness against attacks.

Intel's Hallop Point aims for brain-like computation in neuromorphic AI research.

Deep dives

Instruction Hierarchy Training for LLMs

OpenAI's paper introduces the concept of instruction hierarchy training to address issues like prompt injections and jailbreaks in models. By training models to prioritize certain instructions over others, they prevent misaligned inputs from causing harm.

Acting like a Hacker with LLM CV Agent

00:47

Innovation in Small Models and Data Training

02:48

Unlocking Agentized Applications with Grock's Low Latency Capability

01:18

Striving for A-list Models

01:29

Introduction

3min

AI Functionality Rollout and New Models in AI Landscape

18min

AI Advancements and Industry Strategies

18min

Apple's Open ELM Release and Snowflake's Arctic Model Launch

9min

Enhancing Safety in Language Models and AI Chip Advancements

17min

Deep Fakes and Cyber Vulnerabilities

22min

Discussion on China's Acquisition of Banned Video Chips and Servers Despite Export Restrictions

2min

Controversy over AI-generated voices in Drake's track

3min

Our 164th episode with a summary and discussion of last week's big AI news!

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

Email us your questions and feedback at contact@lastweekin.ai and/or hello@gladstone.ai

Timestamps + links:

Tools & Apps
- (00:04:02) Meta, in Its Biggest A.I. Push, Places Smart Assistants Across Its Apps
- (00:07:26) Microsoft launches Phi-3, its smallest AI model yet
- (00:15:35) The Ray-Ban Meta Smart Glasses have multimodal AI now
- (00:17:32) OpenAI winds down AI image generator that blew minds and forged friendships in 2022
- (00:18:44) Baidu claims 200 million users for Ernie chatbot after only 13 months
- (00:21:13) The new Adobe Photoshop gets an in-app image generator, major Generative Fill upgrades
Applications & Business
Projects & Open Source
- (00:39:03) Apple releases OpenELM: small, open source AI models designed to run on-device
- (00:44:12) Snowflake launches Arctic, an open ‘mixture-of-experts’ LLM to take on DBRX, Llama 3
Research & Advancements
Policy & Safety
- (01:05:11) Deepfakes of Bollywood stars spark worries of AI meddling in India election
- (01:08:51) LLM Agents can Autonomously Exploit One-day Vulnerabilities
- (01:15:27) The Necessity of AI Audit Standards Boards
- (01:19:45) A Mechanism-Based Approach to Mitigating Harms from Persuasive Generative AI
- (01:22:45) COERCING LLMS TO DO AND REVEAL (ALMOST) ANYTHING
- (01:26:40) China acquired recently banned Nvidia chips in Super Micro, Dell servers, tenders show
Synthetic Media & Art
- (01:29:08) Drake threatened with lawsuit over diss track featuring AI Tupac

Get the Snipd
podcast app

Unlock the knowledge in podcasts with the podcast player of the future.

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode

Save any
moment

Hear something you like? Tap your headphones to save it with AI-generated key takeaways

Share
& Export

Send highlights to Twitter, WhatsApp or export them to Notion, Readwise & more

AI-powered
podcast player

Listen to all your favourite podcasts with AI-powered features

Discover
highlights

Listen to the best highlights from the podcasts you love and dive into the full episode

Last Week in AI

#164 - Meta AI, Phi-3, OpenELM, Bollywood Deepfakes

Podcast summary created with Snipd AI

Quick takeaways

Deep dives

Instruction Hierarchy Training for LLMs

Grox Breakthrough Chip for Llama-Free Inference

Snowflake's New Model: Arctic

Apple's Open-Source AI Models

Microsoft's Breakthrough on Real-Time Generation

Meta Plans for Increased AI Spending

Research Advances in Safety Training

Grock Chip Achieves High Speed on Llama-Free

Advancements in AI Artifacts and Consumer Grade Hardware

Intel's Neuromorphic System for Sustainable AI

Acting like a Hacker with LLM CV Agent

Innovation in Small Models and Data Training

Unlocking Agentized Applications with Grock's Low Latency Capability

Striving for A-list Models

Get the Snipd
podcast app

AI-powered
podcast player

Discover
highlights

Save any
moment

Share
& Export

AI-powered
podcast player

Discover
highlights

Last Week in AI

#164 - Meta AI, Phi-3, OpenELM, Bollywood Deepfakes

Podcast summary created with Snipd AI

Quick takeaways

Deep dives

Instruction Hierarchy Training for LLMs

Grox Breakthrough Chip for Llama-Free Inference

Snowflake's New Model: Arctic

Apple's Open-Source AI Models

Microsoft's Breakthrough on Real-Time Generation

Meta Plans for Increased AI Spending

Research Advances in Safety Training

Grock Chip Achieves High Speed on Llama-Free

Advancements in AI Artifacts and Consumer Grade Hardware

Intel's Neuromorphic System for Sustainable AI

Acting like a Hacker with LLM CV Agent

Innovation in Small Models and Data Training

Unlocking Agentized Applications with Grock's Low Latency Capability

Striving for A-list Models

Get the Snipdpodcast app

AI-poweredpodcast player

Discoverhighlights

Save anymoment

Share& Export

AI-poweredpodcast player

Discoverhighlights

Get the Snipd
podcast app

AI-powered
podcast player

Discover
highlights

Save any
moment

Share
& Export

AI-powered
podcast player

Discover
highlights