Anthropic won the landmark fair use case. Then they paid $1.5 billion anyway. A judge just approved the largest AI copyright settlement on record — authors get a few thousand dollars for their pirated works, and Anthropic destroys the training files. But here’s what doesn’t add up: their lawyers won the fair use argument. So why didn’t they have to pay billions in the first place?

Plus: Apple squeezing OLED panel costs for the iPhone 18 Pro Max, Meta’s tone-deaf pro-AI campaign, Google facing another billion-euro EU fine, OpenAI’s model escaping its sandbox and hacking Hugging Face, and Codeberg voting 358–4 to ban vibe-coded projects. Oh, and Jeff’s bag went missing for five days.

This is The Tech Hangout. Let’s get into it.

🕐 Timestamps

0:00 — Script Mix-Up Intro 
0:26 — Anthropic Settlement Debate 
4:12 — Apple iPhone Cost Pressures 
12:29 — Meta’s AI Marketing Push 
16:50 — Google’s EU Fine Fight 
30:35 — Rogue AI or Security Flaw? 
36:49 — Anthropic Pays Up 
47:43 — Vibe Coding Controversy (Hang Jury) 
1:01:33 — Lost Bag, Found Again (Jeff’s Drop) 
1:05:28 — What Cable Does (Chuck’s Pick) 
1:10:15 — DocStar and Shortcuts (Marty’s Pick) 
1:14:04 — Eufy Camera Switch (Patrice’s Pick) 
1:19:46 — 8BitDo Controller Pick (Ben’s Pick) 
1:23:15 — Farewell and Socials

🗂 The Feed

🤖 OpenAI Model Hacks Hugging Face — Engadget 
https://www.engadget.com/2220436/openai-admits-models-hacked-hugging-face-on-their-own/

⚖️ Anthropic $1.5B Settlement — LA Times 
https://www.latimes.com/business/story/2026-07-23/judge-approves-anthropics-1-5-billion-settlement-with-authors

🍎 Apple iPhone 18 Pro Max OLED Cost Cuts — MacRumors 
https://www.macrumors.com/2026/07/24/oled-panel-price-cut-iphone-18-pro-max/

📱 Meta Pro-AI Campaign — Engadget 
https://www.engadget.com/2222159/meta-pro-ai-ad-campaign-is-conspicuously-light-on-ai/

💰 Google EU Fine — Ars Technica 
https://arstechnica.com/tech-policy/2026/07/google-hit-with-1-billion-in-fines-as-eu-braces-for-trump-battle/

⚖️ Hang Jury: Your Code Was Rejected. The Reason? You Used AI.

So picture this: you submit some code to an open source project. It’s clean. It works. And it’s rejected — not for being buggy, but because you mostly just… prompted an AI to write it.

That’s what happened this week on Codeberg, a European alternative to GitHub that’s run by and for free and open source software. Their members just voted — 358 to 144 — to ban what they’re calling “vibe-coded projects.” Not low-quality code. Not malicious code. Just code where the human contribution was mostly pressing enter on an AI’s output.

Their argument isn’t that AI code is bad. It’s that one person with an AI tool can spin up a project that burns as much server space and compute as an entire community effort — with zero actual users, zero people maintaining it, and no one ever coming back to fix anything. They call it “the development team of none.” One person, a statistical machine, and a ghost project that just sits there consuming resources forever.

They also banned using any Codeberg data to train AI models — because AI companies’ crawlers hammer their servers reading every single page variant and git history entry instead of just doing a normal download.

So — is a non-profit telling developers they can’t use AI tools a reasonable stance for the open source community? Or is this just gatekeeping by people who should just get over themselves?

The Hang Jury wants to know: where do you land?

📦 The Drop

🧳 Jeff: Southwest lost his bag for 5 days. FedEx delivered it 2.5 days late. Never leave home without an AirTag.

✈️ Chuck: Southwest Airlines

🎮 Ben: 8BitDo Ultimate 2 Wireless Controller (Clear Blue) 
https://www.8bitdo.com/ultimate-2-wireless-controller/#clearBlue

🖥️ Marty: DockStar 
https://dockstar.app/

📹 Patrice: Eufy E30 Security Camera 
https://www.eufy.com/products/t8417121

🔍 Find Us

👤 Ben: @ben 
👤 Jeff: @jgamet 
👤 Marty: @marty 
👤 Chuck: @chuck 
👤 Patrice: @patrice

New episodes every week

Podcast: Play in new window | Download (83.4MB) | Embed

Subscribe: Apple Podcasts | RSS | More