• AI Odyssey
  • Posts
  • Sora's first BRUTAL review is here...

Sora's first BRUTAL review is here...

Plus: Meet the NEW AI safety & security board

Welcome back to AI Odyssey! 👋

Today, we’re uncovering the truth about Sora, some heat brewing up in the OpenAI vs xAI drama and lastly, you’ll never guess what big names are a part of the new AI safety and security board.

Let's dive in!

P.S. Have you ever seen a Laniakea Supercluster? The AI art will surprise you. ;)

Is Sora as good as it looks? Here’s an honest review by a post-production artist…

When things look too good to be true, sometimes there are crevices to be filled on closer inspections.

Since the announcement of Sora, nobody has forgotten its video content creation powers that looked miles ahead of its game. Now, a filmmaker who got early access to shoot with it is here to fill up the real strengths and limitations of Sora.

Shy Kids, a digital production team was picked by OpenAI to produce short films. In an interview with visual effects news outlet fxguide, post-production artist Patrick Cederberg described “actually using Sora” as part of his work on airhead.

The truth:

While OpenAI’s post lets us assume that Sora is fully formed, the reality is that these were professional productions, complete with robust storyboarding, editing, color correction, and post-work like rotoscoping and VFX. 

Just as Apple says “shot on iPhone” but doesn’t show the studio setup, professional lighting, and color work after the fact, the Sora post only talks about what it lets people do, not how they actually did it.

Cederberg’s interview is interesting and quite non-technical, so if you’re interested at all, head over to fxguide and read it

Some specific cons:

  • AI routinely generates unwanted elements which is another time-consuming process, if they can’t get the prompt to exclude them.

  • Precise timing and movements of characters or the camera aren’t really possible

  • The generated clips were also often in slow motion for no particular reason.

As a result, the team did hundreds of generations, each 10 to 20 seconds, and ended up using only a handful. 

Here’s a behind-the-scenes video:

SPONSORED BY HUBSPOT

Streamline your work with the power of AI! Dive into HubSpot’s guide to using ChatGPT at work and learn how to elevate your productivity and creativity. 

  • Efficiency Unlocked: Enhance productivity with daily ChatGPT integrations

  • Everyday Solutions: ChatGPT for streamlined email and project management

  • Practical Insights: Guidelines for effective ChatGPT application in work

  • Ethical Guidelines: Best practices for responsible ChatGPT use

  • Future Ready: Skill enhancement and learning with AI tools

  • Access 100 prompt ideas to help you unleash the power of generative AI.

xAI, Elon Musk’s OpenAI rival, is closing on $6B in funding

You know Elon Musk has been going hard against OpenAI lately, right?

Well, here’s some news. xAI was initially looking to raise $3 billion at a $15 billion valuation. But…

There was so much investor interest that they bumped it up to $6 billion at an $18 billion pre-money valuation! The deal is expected to close in the next few weeks unless terms change again.

Some of Musk's buddies are getting in on this investing round - including Sequoia Capital, Steve Jurvetson's Future Ventures, Valor Equity Partners, Gigafund. A lot of these firms have close ties to Musk's other ventures like SpaceX and Tesla.

Musk's grand vision is for xAI 

To connect the digital and physical worlds by incorporating training data from his various companies like Tesla, SpaceX, Boring Company, and Neuralink.

The idea is that xAI's chatbot Grok, which is already integrated into Twitter/X, can then supercharge things like self-driving cars and Tesla's robot Optimus.

X itself owns a stake in xAI, potentially benefiting from xAI's growth despite X's recent struggles under Musk's ownership.

Musk vs OpenAI

This is Musk's way of taking the battle to OpenAI after their fallout. 

He's already open-sourced some xAI code and has been trolling OpenAI publicly. With this funding round, he's doubling down on overtaking his old startup. 

Sam Altman, Satya Nadella, and other tech execs join the government’s AI safety board

The Biden administration is forming a new Artificial Intelligence Safety and Security Board.

This is to develop recommendations for the secure use of AI within U.S. critical infrastructure sectors like energy, transportation, and manufacturing.

The board includes leaders from major tech companies like OpenAI's Sam Altman, Nvidia’s Jensen Huang, Microsoft’s Satya Nadella, Alphabet’s Sundar Pichai, academics, civil rights leaders, and executives from critical infrastructure companies.

The board, chaired by Homeland Security Secretary Alejandro Mayorkas, aims to address potential risks of embedding AI into vital systems while also harnessing AI's benefits.

It is part of the Biden administration's efforts to establish oversight of powerful AI systems amidst growing concerns, as the government has largely avoided regulating the tech industry so far.

The board will hold its first meeting in May and convene quarterly to provide guidance to critical infrastructure operators on safely deploying AI technologies.

💣 Latest AI hits: Quick recap

  • Apple has reportedly resumed talks with OpenAI to build a chatbot for the iPhone. [Link]

  • Watch it and weep (or smile): Synthesia’s AI video avatars now feature emotions. [Link]

  • What will Instagram’s chatbot creator look like? Alessandro Paluzzi posted these screenshots in a thread where he’s been tracking the feature since January. [Link]

  • Google adds AI conversation practice for English language learners.[Link]

  • Samsung and Google tease “exciting things” for AI on Android and Galaxy hardware. [Link]

  • Nvidia CEO hand-delivers world’s fastest AI system to OpenAI, again — first DGX H200 given to Sam Altman and Greg Brockman. [Link]

  • Apple’s new OpenELM teases the future of AI on the iPhone. [Link]

  • Microsoft Edge now lets you rewrite text using Copilot without opening. [Link]

  • ‘To the Future’: Saudi Arabia Spends Big to Become an A.I. Superpower. [Link]

AI art to sip your coffee with…

Prompt: Laniakea Supercluster 

AI tools to explore:

  • Layla: AI trip planner. Partnered with Skyscanner and Booking.com, Layla curates personalized ideas for you based on what you actually like, showcasing trip destinations through Insta-worthy videos. 

  • Web3summary: Translate hard-to-read crypto jargon. “Our AI is trained to answer questions about DeFi protocols & DAOs. Simplify any web3 project or your research.”

  • Play AI: The voice interface of AI. Real-time conversational voice AI platform for creating human-like voice agents. It contextualizes conversations and handles turn-taking, interruption, voice energy, and emotion modulation.

That’s a wrap! See you on Thursday for a fresh dose of AI Odyssey. 😎

Enjoy the newsletter? Please forward to a pal. It only takes 18 seconds.