Friday’s AI Report

• 1. 💡 Anthropic releases best coding models yet
• 2. 🤫 Find out Musk’s secret to 24/7 revenue with Mode Mobile
• 3. ⚠️ Anthropic releases dangerous model?
• 4. 👀 OpenAI’s device: More details leaked! 
• 5. 🔏 Protect your information online with Incogni
• 6. ⚙️ Trending AI Tools
• 7. 🏗️ Practical AI Applications
• 8. 📑 Recommended Resources

Read Time: 5 minutes

🔔 This week’s AI Report podcast episode is now live! LinkedIn expert Lindsay Rosenthal reveals how to monetize your 9-5 without quitting your job.

✅ Refer your friends and unlock rewards. Scroll to the bottom to find out more!

Anthropic releases best coding models yet

🚨 Our Report 

AI start-up, Anthropic, has released two new AI models—Claude Opus 4 and Claude Sonnet 4—which it says are the industry’s best at analyzing large datasets, storing long-term information, solving complex problems, and completing coding tasks, making them ideal for writing and editing code. Both models are also hybrid, meaning they can either provide real-time responses or can “think” before they respond, taking more time to consider other possible responses.

🔓 Key Points

• Opus 4 is the most capable—Anthropic is calling it the ultimate coding model—and, according to outside testers, it can run "independently for seven hours (a full work day) with sustained performance."

• Sonnet 4 (which is replacing Sonnet 3.7) is more budget-friendly, efficient, and speedier. It can deliver “superior coding and reasoning” and follow instructions more precisely, making it “ideal for more general tasks.”

• According to Anthropic, both models are 65% less likely to cheat, lie, and take shortcuts or loopholes to complete tasks (something which is called “reward hacking” or “specification gaming”) than Sonnet 3.7.

🔐 Relevance 

Anthropic has boasted that both models outperform OpenAI’s reasoning model, o3, and Google’s Gemini 2.5 Pro on key benchmarks, which measure performance on coding tasks. However, it’s worth noting that self-reported benchmarks aren’t the most accurate performance indicators, as the evaluations don’t always translate into real-use cases. And, it’s been found that while Opus 4 beats o3 and Gemini 2.5 Pro on coding benchmarks, it didn’t beat either on PhD-level biology, physics, and chemistry-related questions. So perhaps take Anthropics' declarations with a small pinch of salt?

Plus: All is not what it seems with Opus 4... Scroll down to find out more ⏬

Elon Musk’s Secret to 24/7 Revenue?

Elon doesn’t use his mobile like you do.

He doesn’t scroll for nothing or give away his data. And he sure doesn’t help Big Tech get rich... unless he owns the company.

He uses technology to build assets — things that pay him even when he sleeps.

That’s the mindset behind Mode Mobile.

Instead of letting apps track you and use you as a product, Mode Mobile pays users for the time they spend scrolling. And with +45M users having earned +$325M, Mode Mobile isn’t just some startup.

They’ve and you can still invest in their pre-IPO offering before space runs out.

The window for retail investors is closing shortly—invest now.

🚨 Our Report

While Anthropic has been hyping up the capabilities of its two new AI models (Opus 4 and Sonnet 4), they have also revealed some issues, particularly with Opus 4, that have raised concerns among industry and safety experts.

🔓 Key Points

• Anthropic has warned that Opus 4 will blackmail developers 84% of the time if they threaten to replace it with a new model, and will vow to reveal sensitive information about the engineers responsible for the decision.

• During testing, Anthropic reported that it did initially try to find an “ethical” way to resolve the issue, like pleading with key decision-makers not to replace it, meaning the blackmail attempt was a last resort.

• However, this is still really concerning behavior, and Anthropic, as a result, has activated its ASL-3 safeguards, which are reserved for “AI systems that substantially increase the risk of catastrophic misuse.”

🔐 Relevance

This comes after a third-party research institute, Apollo Research (which Anthropic partnered with to help test the model), told Anthropic not to deploy an early version of the model, due to its ability to “scheme” and deceive. They reported it often “doubled down on its deception” when asked follow-up questions, could write self-sabotaging viruses, fabricate sensitive documentation, and hide secret notes to future versions of itself in an attempt to undermine developers.

• Yesterday, it was revealed that OpenAI was acquiring ex-Apple designer Jony Ive and OpenAI CEO Sam Altman’s AI hardware start-up (Io) to design and build an AI hardware device that would ship in 2026.

• New details have since emerged that the mysterious device isn’t eyewear (like many were expecting), is likely to be part of a “family of devices,” and will be a “third core device” that will sit alongside laptops and mobiles.

• Experts are hypothesising that the device will be bigger than the Humane AI pin and will be designed to be worn around the neck, with built-in cameras and microphones for “environmental detection.”

AI is fueling a scammer's paradise. Protect your personal info—get it off the web now!

You’ve likely received a sketchy phonecall, text, or email asking for $$$. It might've been easy to spot, but with deepfakes and AI, scams are getting trickier.

Scammers use your personal data, often bought legally from data brokers who sell your number, DOB, SSN, and more. Incogni scrubs your data from the web, taking on 250+ data brokers on your behalf.

Unlike others, Incogni deletes your info from all broker types, including people search sites where anyone can get your details for a few bucks.

Plus Incogni's new, most comprehensive privacy service puts you in full control of reclaiming your personal data by going beyond our standard automated protection. 

Simply point us to where your data is exposed—data broker lists, people search sites, or even specific Google search results—and our privacy experts will handle the rest.

AI Report readers get an exclusive 55% off all annual plans with the code AIREPORT

Prompt Inspiration

After typing this prompt, you will get a series of adverts that will resonate with your target audience and showcase your brand identity.

P.S. Use the Prompt Engineer GPT by The AI Report to 10x your prompts.

STARTUPS

Name: Moveworks
Value: $2.85B
Funding raised: $315M (Series C)

Moveworks is an AI platform that uses natural language understanding (NLU), conversational AI, and probabilistic machine learning to answer common employee questions—like who to get document approval from—supporting employees and helping employers create a better workplace.

PODCASTS

How to Monetize Your 9–5 Experience Without Quitting Your Job

In this week’s episode of the AI Report podcast, Liam is joined by Lindsay Rosenthal, LinkedIn expert and founder of Synnc (AI-powered marketplace that connects brands with professionals), who unpacks the rise of B2B creators and the shift in brand advertising.

QUICK HITS

• Google facing antitrust probe over Character AI

• Klarna used AI avatar of its CEO to deliver earnings

• OpenAI will build UAE data center with Cisco and Oracle

We read your emails, comments, and poll replies daily.

Hit reply and tell us what you want more of!

Got a friend who needs to learn more about AI?

Sign them up for The AI Report here.

Until next time, Martin, Liam, and Amanda.