A light-powered chip slashes AI’s energy appetite, tiny models punch far above their weight, and the co-creator of RSS has a plan to make AI companies finally pay for the data they use. Hosts Maya Lin and Sébastien Deschamps break down the week’s biggest AI stories.

In this episode, we cover:

• 🚀 The Hardware Breakthrough: How a photonic chip from the University of Florida makes AI 100x more efficient.
• 🤖 The Rise of Small Models: Why open-source models like K2 Think (32B) and Baidu’s ERNIE are challenging the “bigger is better” paradigm.
• ⚙️ The MCP Registry: Why the “App Store for Enterprise AI” is a silent game-changer for developers.
• ⚖️ The Data Licensing Revolt: Analyzing the “Real Simple Licensing” protocol and its potential to fix AI’s biggest ethical problem.
• 💡 Startup Spotlight: AI “second brains,” companions to combat loneliness, and the state of robotaxis.
• ⚠️ The Critical Counterpoint: A discussion on a new book arguing AI is a new frontier for women’s oppression.

Key Questions We Explore:

• Would you pay for an AI companion to combat loneliness?
• Is efficiency the new battleground for AI supremacy?
• Can a new protocol finally solve the AI data licensing crisis?

Follow Byte Of Truth for your weekly dose of insightful and critical tech analysis.

#AI #MachineLearning #TechPodcast #Innovation #EthicsInAI #ByteOfTruth

This week, the AI industry faces a monumental pressure test. We break down the biggest stories that are reshaping the landscape:

• The Copyright Wars: Anthropic settles a $1.5B lawsuit over “pirated” books, and Warner Bros. sues Midjourney. Is this a victory for creators or just the cost of doing business?

• The Safety Firestorm: Over 20 State AGs warn OpenAI about child safety, and a new study links smartphones under 13 to mental health harms. Are regulators finally catching up?

• The Open-Source Counter-Offensive: New releases from Hugging Face, and Alibaba are democratizing powerful AI, challenging the closed models of giants.

• Industry Intel: Key hires at OpenAI, a cautionary tale from a battery startup’s failure, and the new tools developers can use right now.

Join host Sébastien Deschamps on Byte Of Truth for a deep dive into the ethics, policies, and technologies that matter.

Keywords: AI, Artificial Intelligence, Machine Learning, Copyright, Ethics, OpenAI, Anthropic, Hugging Face, Regulation, Child Safety, Open Source, Tech News.

Can you spot a deepfake? No, really—can you?

In this episode of Byte of Truth, host Sébastien Deschamps breaks down a fascinating new study that shows how simple digital literacy tools—like tips, visuals, and even games—can significantly boost our ability to detect AI-generated fakes. From Balenciaga Popes to phantom fingers, we explore the strange world of synthetic media and how to stay one step ahead of the algorithm.

You’ll learn:

• Why deepfakes are more dangerous (and accessible) than ever

• What kinds of training actually help people spot them

• Which techniques don’t work (and why)

• How to build digital resilience—without becoming overly skeptical

A must-listen for anyone interested in tech, trust, and the weird future of visual truth online.

Based on the paper:

Geissler, D., Robertson, C., & Feuerriegel, S. (2025). Digital literacy interventions can boost humans in discerning deepfakes. https://arxiv.org/abs/2507.23492

Behind every gripping data-driven headline is a journalist elbow-deep in messy, inconsistent, sometimes maddening data. In this episode, we explore the unsung labor of data preparation in journalism — and how it compares to the workflows of data scientists.

Based on a recent study, we uncover how journalists tackle dirty data from fragmented sources, PDFs from FOIA requests, and tables that change over time like shapeshifting monsters. From regionally inconsistent COVID stats to detective-level entity matching, this is the side of data journalism the public rarely sees — but desperately needs to understand.

🎙️ Topics include:

• What “dirty data” really means in journalism

• The 4 integration nightmare archetypes

• Why journalists can’t just impute missing values

• A fresh way to categorize data quality issues

• The tension between storytelling and spreadsheet chaos

If you’ve ever cursed a CSV or tried to make sense of government data, this one’s for you.

🎧 Hosted by Sébastien Deschamps

In recent years, there has been a growing concern about the intersection of artificial intelligence (AI), conspiracy theories, and hate crimes. Researchers have found that AI algorithms can predict hate crimes based on online search trends for conspiracy theories.

This episode of Byte Of Truth explores this complex issue in-depth, examining the ways in which AI-driven misinformation is spreading across social media platforms and influencing real-world violence.

Key Points:

• AI algorithms can predict hate crimes based on online search trends

• Conspiracy theories have been linked to real-world violence

• Researchers are exploring ways to use AI to counter misinformation

In this episode of Byte of Truth, host Sébastien Deschamps uncovers the dark art of the rug pull — the scam where your hot new token suddenly becomes worthless.

But this isn't just another crypto horror story. We explore RPHunter, a cutting-edge tool that fuses smart contract code analysis with blockchain transaction behavior to detect scams before they drain your wallet.

Learn how scammers engineer trust, manipulate liquidity, and vanish with millions — and how AI-powered graph networks are learning to fight back.

If you’ve ever minted a token, aped into a new coin, or just want to know how tech is outsmarting cybercriminals, this one’s for you.

Topics include:

• How rug pulls actually work (step by shady step)

• The science behind code and transaction graph fusion

• Why old scam detection tools fall short

• What 645 real rug pulls taught researchers

• RPHunter’s 95% detection accuracy and what that means for DeFi security

Hit play and stay ahead of the rug.

Wu, H., Wang, H., Li, S., Wu, Y., Fan, M., Jin, W., Zhao, Y., & Liu, T. (2025). Your Token Becomes Worthless: Unveiling Rug Pull Schemes in Crypto Token via Code-and-Transaction Fusion Analysis.

Deepfakes. AI-generated propaganda. Anonymous accounts reaching millions in minutes.

In this episode of Byte of Truth, host Sébastien Deschamps explores a bold new idea from Cambridge researchers: tiered anonymity for social media. Think of it as a traffic-light system for online identity — pseudonyms for everyday users, ID checks for influencers, and full-blown fact-checking for viral accounts.

We dive into why this might be the most balanced response yet to the chaos of misinformation — one that preserves privacy while demanding accountability from those with the most digital reach.

Is it the end of online freedom? Or the start of smarter speech?

Tune in for the insights. Stay for the big questions.

David Khachaturov, , Roxanne Schnyder, and Robert Mullins. "Governments Should Mandate Tiered Anonymity on Social-Media Platforms to Counter Deepfakes and LLM-Driven Mass Misinformation." (2025).

#AI #Deepfakes #DigitalPolicy #SocialMedia #FreeSpeech #ByteOfTruth