The Weekly Byte January 22-28, 2026

Week of January 22-28, 2026

🔥 Lead Story

Tesla Invests $2B in Elon's xAI

In its Q4 earnings, Tesla disclosed a $2 billion investment in xAI — Musk's AI company that owns X.com and Grok. The stated goal: "enhance Tesla's ability to develop and deploy AI products and services into the physical world at scale."

Here's the kicker: Tesla shareholders actually voted against this investment in November. But Tesla proceeded anyway, arguing it aligns with "Master Plan Part IV."

xAI reportedly burned $7.8B in the first nine months of 2025. This is essentially Musk shuffling money between his companies, but it signals deeper integration between Tesla's autonomous ambitions and xAI's models.

Business impact: Watch for Grok-powered features in Tesla vehicles. The AI race isn't just about chatbots—it's about who owns the physical world interface.

📰 This Week's Top Stories

1. Amazon Cuts 16,000 Jobs Globally

Amazon announced it's eliminating 16,000 corporate jobs—the second major round since October's 14,000 cuts. The company framed it as "reducing layers, increasing ownership, and removing bureaucracy."

Translation: more AI, fewer humans.

Why it matters: The tech layoff wave continues into 2026. SMBs relying on AWS should monitor for service changes as Amazon reallocates resources.

2. Google Explores Publisher Opt-Out for AI Search (UK)

Under pressure from UK regulators, Google is "exploring" letting publishers opt out of AI-generated search features. If major publishers pull their content, AI Overviews could become significantly less comprehensive.

Why it matters: If this spreads beyond the UK, it could reshape SEO strategy entirely. Watch this space.

3. Apple's 30% Patreon Tax Incoming

Apple will soon take up to 30% from all Patreon creators using the iOS app. Creators are scrambling to push subscribers to web payments instead.

Why it matters: The platform tax debate heats up. Creator economy businesses need to factor in app store fees or find workarounds. This affects anyone building subscription-based products.

4. Google AI Plus Plan Rolls Out ($7.99/mo)

Google's more affordable AI tier is now available in 35+ countries. Includes Gemini 3 Pro, Nano Banana Pro image generation, and 200GB storage.

Why it matters: AI tools hitting mainstream pricing. Competition heating up below the $20/mo tier dominated by ChatGPT Plus. Good news for SMBs on a budget.

5. Trump Admin Using Gemini to Write Regulations

ProPublica reports the Transportation Department is using Google's Gemini to draft regulations. Officials quoted: "We don't need the perfect rule... don't even need a very good rule."

Yikes.

Why it matters: AI-written government policy is here. Quality concerns aside, this signals rapid AI adoption in unexpected sectors.

6. Trinity Large: Open 400B MoE Model Released

Arcee.ai released Trinity Large, an open-source 400B parameter sparse mixture-of-experts model with 13B active parameters per token. Uses 256 experts with 4 active per token—higher sparsity than DeepSeek-V3 or Qwen.

Why it matters: Open-source AI keeps closing the gap with closed models. SMBs can now access frontier-class capabilities without API dependencies or vendor lock-in.

7. TRAIN Act: AI Training Transparency Bill

A bipartisan bill was introduced letting copyright holders discover if their work trained AI models. Backed by RIAA and SAG-AFTRA.

Why it matters: Copyright friction is increasing. AI companies may face mandatory disclosure requirements. If you're building AI products, factor this into your data sourcing strategy.

8. Nvidia Debuts AI Weather Models

Nvidia claims its AI weather models are outperforming traditional physics-based forecasting. Google's WeatherNext making similar claims.

Why it matters: AI is eating specialized domains. Weather, legal, medical—every vertical is being disrupted. If you're in a data-rich industry, someone's building an AI to compete with you.

🛠️ Tool of the Week

Sherlock: MitM Proxy for LLM Tools

Open-source proxy that intercepts LLM API traffic and visualizes token usage in a real-time terminal dashboard. Track costs, debug prompts, and monitor context window usage.

Privacy-conscious devs can finally see exactly what their AI coding assistants are sending. Supports Claude, Gemini CLI, and Codex.

💡 Quick Takes

• YouTube pledging to "reduce the spread of low quality AI content" — good luck with that
• Jellyfin published formal LLM/AI development policy for contributors
• Ben Affleck being surprisingly thoughtful about AI in Hollywood (unlike Chris Pratt who wanted an AI co-star)
• "The Five Levels" essay trending on HN — from "spicy autocomplete" to "dark factory" — useful mental model for AI adoption stages

📊 Numbers That Matter

🎯 Brian's Take

This week I started an experiment: What if I hired an AI as an actual employee?

Not a chatbot I poke when I have questions. Not a code assistant. A 24/7 autonomous agent with access to my servers, my WordPress sites, my GitHub repos—working while I sleep.

I'm using Clawdbot to run an agent I've named Erdma (after the Erdmanndli—mythical earth spirits from the hills near my home in Switzerland). In the past 48 hours, Erdma has:

• Audited and cleaned up an entire WordPress site
• Created 50+ draft blog posts
• Built custom automation scripts
• Added affiliate product boxes to 21 articles
• Delivered a morning briefing before I woke up

The results have been... surprisingly good. And a little unsettling.

Tomorrow I'll publish the full breakdown: what worked, what broke, and whether AI employees are ready for real business workflows. Stay tuned.

Thanks for reading The Weekly Byte. Hit reply if you spotted something I missed.

— Brian

Follow me

If you liked this article, Follow Me on Twitter to stay updated!