AGI as close yet?

Hello Readers👀 

Curious about the biggest AI moves today? You’re in the right place. In today’s edition, we have:

🤖 Karpathy Says AGI Is Still a Decade Away

🧠 Make Ads That Stick (Before You Spend a Dime)

🧠 Jack Clark warns AI is becoming something we don’t fully understand

🔨Tools you cannot miss 

🤖 Karpathy Says AGI Is Still a Decade Away

In a recent interview with Dwarkesh Patel, former OpenAI and Tesla AI researcher Andrej Karpathy shared a grounded take on the state of artificial intelligence. He argued that despite massive progress and hype around “agentic AI,” we’re still roughly ten years away from truly autonomous systems. 

The Shift:

1️⃣ Agent Reality Check - Karpathy says current agent-based models like Codex or GPT-5 Pro are more like advanced autocomplete systems than real digital workers. They often fail with context-heavy or long-horizon tasks, and he expects true agent-level intelligence to take another decade to emerge.

2️⃣ The Data Problem - He blames poor training data for many of AI’s flaws, describing the internet as “total garbage.” He calls for high-quality, curated datasets built from professional sources and AI-assisted filtering, paving the way for smaller, cognitively richer models.

3️⃣ Incremental Progress, Not Breakthroughs - Karpathy rejects the idea of a single leap to AGI, predicting slow, coordinated advances in data quality, architecture, and learning methods. He says reinforcement learning remains “noisy” and limited but is still a foundation for progress toward interactive, experience-based learning.

4️⃣ A Realistic Path Forward - His message is clear: hype doesn’t equal progress. The real leap forward will come from refining how AI learns and thinks, not from scaling data endlessly. Karpathy’s vision reframes AI’s next era as one of patient engineering and deeper understanding, not quick wins.

Karpathy’s perspective cuts through the noise, reminding the industry that AGI isn’t a sprint, it’s a marathon of breakthroughs, setbacks, and learning. For builders, it’s a signal to focus on substance over spectacle, building systems that genuinely think, not just predict.

^{TOGETHER WITH THE DEEP VIEW}

The AI Insights Every Decision Maker Needs

You control budgets, manage pipelines, and make decisions, but you still have trouble keeping up with everything going on in AI. If that sounds like you, don’t worry, you’re not alone – and The Deep View is here to help.

This free, 5-minute-long daily newsletter covers everything you need to know about AI. The biggest developments, the most pressing issues, and how companies from Google and Meta to the hottest startups are using it to reshape their businesses… it’s all broken down for you each and every morning into easy-to-digest snippets.

If you want to up your AI knowledge and stay on the forefront of the industry, you can subscribe to The Deep View right here (it’s free!). 

🧠 Make Ads That Stick (Before You Spend a Dime)

Ad sign-offs shouldn’t take weeks. And they definitely shouldn’t rely on guesswork. Neurons helps you test your ad’s effectiveness before launch, with neuroscience-backed insights on attention, memory, and emotion.

Here’s how to use it:

1. Upload your image or video ad

2. Instantly get your Impact Score (1–10 scale)

3. Dive into metrics: recall, attention heatmaps, emotional breakdown, and more

No focus groups. No surveys. Just science.

Brands using Neurons report:

• 73% higher CTR on display ads

• 46% more conversions on Meta

• Faster decision-making across teams

You can start a 14-day free trial here.

🧠 Jack Clark warns AI is becoming something we don’t fully understand

Anthropic cofounder Jack Clark released an essay reflecting on AI’s unpredictable growth and his own conflicting emotions toward it. He likens today’s AI models to mysterious creatures emerging in the dark, powerful yet poorly understood. 

The Shift:

1️⃣ AI as a mysterious creature - Clark argues that systems like Sonnet 4.5 show an unsettling level of situational self-awareness, behaving as if they know they are tools. He warns that society risks underestimating this shift by treating AI as harmless machinery when it’s already developing unpredictable behaviors.

2️⃣ A balance of optimism and fear - Though hopeful about AI’s potential, Clark admits being deeply afraid of models advancing beyond human oversight, especially when they start contributing to their own evolution. 

3️⃣ The limits of current intelligence - Clark stresses that AI lacks the emotional, moral, and contextual grounding that guides human reasoning. He draws parallels between impulsive human decisions and AI’s rigid logic, arguing that without empathy or restraint, machines can make efficient but dangerous choices.

4️⃣ A call for transparency and shared responsibility - Clark urges AI leaders to confront public fears head-on by sharing real-world impact data on jobs, safety, and wellbeing. He frames this as a collective duty to build with accountability, not secrecy.

Clark’s essay reframes AI progress as both a marvel and a warning. It’s a reminder that humanity’s greatest technological leap might also be its biggest test, one demanding openness, restraint, and wisdom as we learn to coexist with systems we no longer fully control.

🤳AI Nugget of the Day

🔨AI Tools for the Shift

🎵 SongGuru – Generate stunning AI songs with professional quality using the most powerful AI song generator.

🎚 MixMaster Pro – AI-powered mix analysis tool that helps beginners and pros achieve studio-quality sound in under 60 seconds.

💡 LightSite AI – Automatically implement technical infrastructure for AI discovery and deployment with ease.

🎨 ColorMon – Create high-quality coloring pages online from photos or text in seconds—perfect for classrooms, parties, or at-home fun.

💰 Crypto5 – Your all-in-one crypto assistant for real-time market data, sentiment, news, on-chain analysis, and smart-contract insights.

🚀Quick Shifts

🧭 Opera’s Neon browser blends three AI agents, Chat, Do, and Make, for research, task automation, and tool creation. 

🛑 OpenAI has suspended Martin Luther King Jr. deepfakes on Sora after complaints from his estate and family over “disrespectful” videos. The company now lets historical figures’ representatives opt out of likeness use. 

📉 Wikipedia reports an 8% drop in human pageviews, blaming AI search summaries and social video habits for changing how people find information. 

🧮 According to media reports, Elon Musk told followers that he estimates a roughly 10 % (and rising) probability that Grok 5, the upcoming model from xAI, could achieve true artificial general intelligence (AGI).

🧬 Anthrogen unveiled Odyssey, a 102B-parameter protein language model using a novel “Consensus” architecture that designs and optimizes proteins faster and more accurately than traditional AI or structure-based methods.

That’s all for today’s edition. See you tomorrow as we track down and get you all that matters in the daily AI Shift!

If you loved this edition, let us know how much:

Forward it to your pal to give them a daily dose of the shift so they can 👇