- The Alignment
- Posts
- Is Elon Musk creating an OpenAI Rival ?
Is Elon Musk creating an OpenAI Rival ?
And Mark Zuckerberg bets big on AI
Welcome to The Alignment, where we help you understand what’s happening in the world of AI without any of the jargon or bs.
Here’s what we have lined up for you today -
Is Elon Musk creating an OpenAI Rival ?
Projections about OpenAI's Foundry model
Mark Zuckerberg is setting up a dedicated AI product team for Whatsapp, Instagram and Messenger
Is Elon Musk creating an OpenAI Rival ?

In recents months Elon Musk has become a vocal critic of some of OpenAI’s decisions with regards to the deployment of their AI models. Musk who helped co-found OpenAI nearly 8 years ago firmly believes that development of AGI should be completely open sourced instead of being in the hands of companies. This was also one the reasons why he chose to leave OpenAI as they began to pivot from being a completely open source organisation.
As reported by The Information, now it seems as though Musk is trying to have another go at spinning up an AGI initiative, except this time to rival OpenAI the very organisation he helped found. It’s being reported that he is recruiting ex Google and Deepmind researcher Igor Babuschkin to work on this initiative.
If there is one thing we know about Elon, it’s that if he is even slightly serious about this, there is a very real chance that we will see a radically different approach to AGI research than what is being done by any other organisation.
Projections about OpenAI's Foundry model
Nathan Labenz, founder of AI video production company Waymark took a deep dive into OpenAI’s proposed Foundry platform which it plans to launch in the coming months. Here are the highlights of Nathan’s analysis -
Foundry is the platform for serving OpenAI's "latest models", which will "soon" come with "more robust fine-tuning" options.
GPT4 is not a single model, but a class of models, defined by pre-training scale & parameter count, and perhaps some standard RLHF. Microsoft's Prometheus (used in Bing) is the first public GPT4-class model.
OpenAI's Foundry will be where businesses build AIs to perform their cognitive tasks – also known as *services*, also known as ~75% of the US economy.
The "Foundry" price range will be from ~$250K/year for "3.5-Turbo" (ChatGPT scale) to $1.5M/yr, suggesting that OpenAI can readily demonstrate GPT4's ability to do meaningful work in corporate settings in a way that inspires meaningful financial commitments.
This looks eerily similar to the TSMC approach, the foundry based out of Taiwan that runs the global semiconductor industry.
Mark Zuckerberg is setting up a dedicated AI product team for WhatsApp, Instagram and Messenger
Mark Zuckerberg is going all in on Generative AI. In an internal announcement he shared that various AI teams will be consolidating their efforts into a single team dedicated to building AI for Whatsapp, messenger and instagram.

The short term focus will be to build creative and expressive tools and the long term focus will be on building AI personas. Not long ago Mark went all in on the Metaverse as the company transitioned from being called Facebook to Meta. One could think that this new move might be in coordination with Meta’s ambition to create new generation virtual experiences in the “Metaverse” although sceptics would argue that this seems to be an attempt for Meta to strike gold as they continue to build their new identity post the Facebook social media era.
Product Watch Booth AI-
Pro quality product photography with AI.
Around the industry -
Snapchat has launched its own ChatGPT powered by ChatGPT.
State of the GenAI industry report by A16Z (and where $$$ value will accrue for AI)
Top Universities in Asia are open to integrating GenAI for educational purposes
FTC warns tech: ‘Keep your AI claims in check’
Anthropic begins supplying its text-generating AI models to select startups
John Oliver’s AI Monologue
You know AI is going mainstream (and probably is in a large inflated bubble) when the leading late night show dedicates a 30 min segment on it. How do you think John did explaining AI ?