👀 Is OpenAI safe enough to build AGI?, Compute is key, AMA, and AI Highlights

(Source)

Ilya Sutskever, co-founder and chief scientist of OpenAI, has effectively left OpenAI.

Why? Because he most likely has seen something.

Either on the tech side or regarding AI responsibly. We know that. Especially after Jan Leike, OpenAI’s Superalignment leader, has also decided to leave. On his way out he tweeted a remarkable threat on X.com.

In short, he believes that OpenAI should dedicate much more bandwidth to preparing for the next generations of models, and stresses the urgent need for OpenAI to become a safety-first AGI company.

I’m an techno-optimist (link) and don’t want to spread fear, but this doesn’t sound good at all.

Luckily we have AI experts like Yoshua Bengio, Stuart Russell, Max Tegmark, and others that are actively thinking about AI safety in a meaningful way.

They have just released a hallmark whitepaper. They are proposing a Framework for Ensuring Robust and Reliable AI Systems.

Are you building an AI product? You should read it, and take the “Verifier“ idea into account.

MaxAI.me - Do more Faster with 1-Click AI

MaxAI.me lets you chat with GPT-4, Claude 3, Gemini 1.5. You can also perfect your writing anywhere, save 90% of your reading & watching time with AI summary, and reply 10x faster on email & social media.

Join 1M+ users on MaxAI.me to do more faster!

(Source)

As Ilya is departing OpenAI, AI experts reflect on his immense influence in the space.

Not only leaders from Google, Microsoft, OpenAI, but also Elon Musk and AI pioneer Geoffrey Hinton, have praised Ilya for his capabilities of developing intelligent systems.

In a recent interview with Joel Hellermark, one quote of Hinton stands out:

"Ilya was always preaching that you just make it bigger and it'll work better. And I always thought that was a bit of a cop-out, that you're going to have to have new ideas too. It turns out Ilya was basically right."

In essence, more computing power is crucial for achieving AGI.

In Generative AI - Navigating the Course to the AGI Future I discuss the exponential growth in computing, measured in FLOPS, driven by hardware and software advancements.

Understanding this gives an edge in predicting AI progress and needs.

On the hardware front, many computing chips form CPUs, GPUs, IPUs, TPUs, and NPUs. My book explains their differences, functionalities, and future directions, and their influence on AI (progress).

Among the increasing computing ability, one emerging trend of large tech companies is to customizing processing units, see for instance Apple's new M4 beast.

Explore other developments in the book e.g. cloud computing, quantum computing, neuromorphic computing, superconductors, and software advances supporting AI breakthroughs.

LLMs will soon respond to complex queries in fraction of seconds, analyzing terabytes of data instantly, solving complex problems.

It’ll be jaw-dropping.