Welcome to this week’s Overclocked! A wave of AI advancements has hit the scene, with major players unveiling powerful new models that redefine what’s possible. Meanwhile, the rise of AI-generated disinformation is making it harder than ever to separate fact from fiction. Let’s break it all down. ⬇️

In today’s newsletter:
🛠️ So many new and cool AI releases, so little time
🔊 Deepfake audio controversy shakes political discourse
🎤 Famous musicians release a silent album in protest of AI-generated music
🗞️ Stay updated on everything happening in your city using AI
🎯 Can you spot the fake photo? 

🎉 This Was a Major Week of AI News

Here’s everything you missed 🧐

The AI race is heating up, and this week, four major players unveiled game-changing models that push the boundaries of what AI can do. Here’s a breakdown of the latest AI news and releases:

🧠 GPT-4.5 | OpenAI

OpenAI’s latest model focuses on natural conversations over rigid step-by-step logic. It’s smoother, smarter, and more nuanced in dialogue.

• Supports file and image uploads for deeper AI-powered analysis.

• Better hallucination control—37.1% false data rate (down from 61.8% in GPT-4o).

• Starting on March 3rd, all pro, plus, and team members will have access to GPT-4.5.

And OpenAI isn’t stopping there—Deep Research has finally been released to the masses – a new AI tool for summarizing large amounts of online data, is set to redefine digital research as we know it.

🎙️ Alexa+| Amazon

Amazon revamps Alexa with AI-driven contextual memory and smarter interactions.

• Handles conversations naturally without constant wake-word prompts.

• Understands emotion and intent for more tailored responses.

• Integrates with smart home tech and even processes images.

• Costs $19.99/month (free for Prime members).

🤖 QwQ-Max | Alibaba

A serious contender from Alibaba that’s built for reasoning and problem-solving.

• “Thinking mode” allows for step-by-step complex analysis.

• Planned open-source release for developers.

• Superior coding and multi-modal processing capabilities.

🛠️ Claude 3.7 Sonnet | Anthropic

Anthropic rolls out a hybrid AI that lets users toggle between quick responses and deep reasoning.

• Two modes: fast-response and extended thinking for in-depth analysis.

• Customizable “thinking budgets” to control costs.

• A major step forward for coding, problem-solving, and AI transparency.

🔊 Deepfake Audio Shakes Political Discourse

We are entering an era where we can no longer tell what’s real. This week, a fake audio clip allegedly featuring Donald Trump Jr. endorsing Russia over Ukraine went viral, amassing over 1.8 million views on social media.

🤷 What happened?

• AI-generated audio, falsely attributed to Trump Jr., claimed, “Honestly, the U.S. should have been sending weapons to Russia.”

• The clip was reposted by FactPostNews, an official Democratic Party rapid response account, before being deleted.

• Experts at GetReal Labs confirmed it was a deepfake, exposing the growing risk of AI-driven disinformation.

☝️ Why it matters:

• The clip circulated widely before fact-checkers intervened.

• The rise of AI-generated political content is making it harder to distinguish truth from fiction.

• Even Spotify and Telegram saw versions of the clip spread across their platforms.

AI disinformation is no longer a distant threat—it’s happening right now. Here’s something to remember, all information isn’t good information. In 2025 and beyond, we must do our research, fact check, and verify everything we see and hear. Otherwise, we’ve lost the plot. 

Weekly Scoop 🍦

📈 Nvidia says AI chip sales remain strong despite DeepSeek fears

📝 Apple AI tool mistakenly transcribes “Trump” instead of “racist”

🎶 Artists release silent album in protest against AI-generated music.

🏦 Major Asian bank to cut 4,000 jobs as AI replaces human roles

🔬 AI cracks a superbug problem in two days—something scientists spent years solving

🤯 What happens when you give AI a Rorschach inkblot test?

🎯 Local News AI Challenge

I think we’re all tired of the gloom and doom of the local news. The good ‘news’ is you can use AI to create a daily local news update that focuses on the things you care about. For example, if you have children in school, you could ask an AI model to send you daily updates about school closures and local school boards. That’s just one use case, there are so many others. 

💡 The Challenge:

Use AI to transform local news delivery. Whether you build a simple script or experiment with existing tools, your goal is to make local updates more engaging!

🚀 Tools You Can Try:

• OpenAI Scheduler – Automate when your news updates go live.

• JustSyft AI - Tell it what kind of news you want to learn about and how often to send you updates

• Perplexity – Get real-time, AI-powered news insights.

If you’re an advanced user, try these: 

• HuggingFace & spaCy – Create concise, engaging summaries of daily news.

• Local News APIs – Fetch and display real-time community updates.

🔍 How creative can you get with AI and local news? Let’s find out!

🤔 Can You Spot the Fake Photo?

Donald Trump and French President Emmanuel Macron met at the White House last week. But here’s the twist—one of these photos is AI-generated. Can you tell which one is real?

That's it for this week! As AI continues to weave into the fabric of our daily routines, from personal fitness to travel planning, we're just beginning to see its true potential. How are you using AI in your daily life? Reply and let us know!

Zoe from Overclocked