AI Inside: Tech Agnostic
For Wednesday, January 15, 2025
Welcome to this week’s AI Inside newsletter, where we break down the latest stories in artificial intelligence and technology from yesterday’s podcast. Let’s dive into the highlights!
Tech Worship and a Call for a "Tech Reformation"
This week, we spoke with Greg Epstein, author of Tech Agnostic: How Technology Became the World’s Most Powerful Religion, and Why It Desperately Needs a Reformation. Greg argues that our relationship with technology mirrors religious devotion and calls for “tech agnosticism” to critically rethink this dynamic.
Biden Administration’s Final AI Moves
As the Biden administration wraps up, two major developments could shape AI's future:
Export Restrictions on GPUs: A new rule introduces a three-tier system for GPU exports, heavily restricting access for adversarial nations like China while allowing freer access for allies. Nvidia criticized the rule as harmful to U.S. competitiveness, while Microsoft expressed cautious compliance.
Executive Order on AI Infrastructure: The order focuses on building large-scale data centers and clean energy facilities to support AI operations while addressing national security concerns. However, it lacks measures to tackle water consumption by energy-intensive data centers.
OpenAI’s Economic Blueprint
Ahead of Biden’s executive order, OpenAI released its vision for U.S. AI regulation. The blueprint calls for significant federal investment in infrastructure like chips and energy, critiques fragmented state-level regulations, and advocates for streamlined federal policies—aligning closely with Biden’s announcements.
Copyright Battles Heat Up
The legal landscape around AI training data is intensifying:
NYT vs. OpenAI & Microsoft: Publishers accuse OpenAI of copyright infringement for training on their data without compensation. The case could force ChatGPT to destroy its dataset if ruled against OpenAI—a potentially catastrophic outcome.
Meta and LibGen Controversy: Meta reportedly used LibGen (a repository of pirated books) to train its generative AI models, raising questions about legality and fair use. This case could set a precedent for how copyrighted material is handled by AI companies.
Amazon’s Alexa Overhaul
Amazon is working on significant upgrades to Alexa but faces challenges like improving response accuracy, reducing hallucinations, and maintaining Alexa’s familiar personality. Despite being an early player in voice interaction, Amazon lags behind competitors in modern AI advancements—but Alexa's influence remains undeniable.
Adobe Firefly Bulk Create
Adobe introduced Firefly Bulk Create, a generative AI tool capable of processing up to 10,000 images simultaneously. It automates tasks like background removal and resizing, making it ideal for large-scale projects. Currently in beta, it requires a subscription to Adobe’s Premium Firefly plan.
Slopaggedon: AI Content Polluting Search Results
A Google Image search for “does corn get digested” recently revealed a flood of low-quality AI-generated imagery—dubbed “slopaggedon.” This highlights how generative content can pollute search results, making it harder to find accurate information.
Google NotebookLM Gets Friendlier
Google has updated NotebookLM with “friendliness tuning” after users found its responses to interruptions adversarial. Now, the AI hosts respond politely when interrupted—modeling human-like surprise instead of frustration.
Thanks for reading! Don’t forget to subscribe at aiinside.show and support us on Patreon at patreon.com/aiinsideshow for ad-free episodes and exclusive perks! See you next week!