AI: OpenAI's GPT-5 'Orion' this winter? RTZ #484
...seeing more hints on the path to GPT-5/GPT-NeXT and 'Stargate' AI infrastructure with Microsoft
OpenAI in this AI Tech Wave, is on a product roadmap roll, as I’ve been pointing out for months. Now that their AGI-driven AI Reasoning OpenAI-o1 has now been ‘rolled out’ to developers at least, all eyes shift of course to what comes after GPT-4o for ‘Omni’.
And that of course as we parse through OpenAI’s collection of AI product road map code-names, points to ‘Orion’. That’s presumably GPT-5, which I’ve also discussed before.
It may or not be the same as GPT-NeXT, which may end up being the next iteration, GPT-6, as OpenAI’s continues their LLM AI leadership ‘AI Scaling’ streak.
OpenAI founder/CEO continues his tradition of playfully signaling that GPT-5 ‘Orion’ is coming, with cryptic messages on X/Twitter discussed on the OpenAI developer forums:
“Orion Announcement .. to be soon in winter?”
“Sam Altman Just posted this on ‘X’ …is it an official announcement that ‘Orion’ model is going to be on winter ???”
Post:
’ i love being home in the midwest.’
’the night sky is so beautiful.’
’excited for the winter constellations to rise soon; they are so great. ’
‘Sam Altman’s official X account.’
‘So excited!’
’That would be amazing !!’”
That of course is a reference to the constellation Orion in the winter skies later this year.
No one outside OpenAI knows how much better GPT-5 is over GPt-4. Although Sam Altman himself has said in prior interviews:
“I expect that the delta between [GPT-]5 and 4 will be the same as between 4 and 3 and I think it is our job to live a few years in the future and remember that the tools we have now are going to kind of suck looking backwards at them and that’s how we make sure the future is better.”
“Look, I don’t want to downplay the accomplishment of GPT-4, but I don’t want to overstate it either,” Altman said later. “And I think this point that we are on an exponential curve, we’ll look back relatively soon at GPT-4 like we look back at GPT-3 now.”
The next generation LLM AI models, and their underlying AI compute and data center infrastructure, are all around the corner. And they’ll have meaningful implications up and down the this AI Tech Stack below as I’ve discussed around rising ‘AI Table Stakes’.
The other clues around OpenAI’s ‘Orion’ ambitions, are from OpenAI’s Japan CEO Tadao Nagasaki recently as alluded to by Techround:
“About a year and a half after the release of GPT-4 in March of 2023, users have been told that they can expect the launch of a new and improved version of Chat-GPT in the near future. In the very, very near future, in fact – before the end of 2024.”
“Indeed, Tadao Nagasaki, the CEO of OpenAI in Japan, revealed information about the much-anticipated successor of GPT-4, a model that’s been named GPT-NeXT.”
“This isn’t the first time we’ve heard of GPT-NeXT, however. The project was first unveiled in May this year at a tech event in Paris, but at that point, OpenAI was tight-lipped about how it would be different from previous models and exactly what it would be capable of doing.”
“Although Nagasaki hasn’t exactly spilt all the (metaphorical) beans, he certainly has given us a whole lot more than we knew before, and from the information he shared at the KDDI Summit a few days ago, it seems like GPT-NeXT is expected to do great things.”
Again, it’s not clear if GPT-NeXT is GPT-5 ‘Orion’, or the version after GPT-5/Orion:
“Nagasaski stated very clearly that not only is GPT-NeXT expected to be better than GPT-4, it’s anticipated to be a whopping 100 times more powerful.”
Presumably that refers to some combination of both LLM AI Training and Inference improvements. But we’ll have to see as and when ‘Orion’ is revealed.
And to get there we may need ‘Stargate’ after all.
And presumably, that level of performance is also going to need further scaled up investments in AI Compute Infrastructure. That could be the ‘Stargate’ $100 billion compute infrastructure investments hinted in March, between Microsoft and OpenAI.
Likely related to that, we got more financing plans detail on that front today in WSJ’s “BlackRock, Microsoft Partner on Massive New AI Infrastructure Fund”:
“The Global AI Infrastructure Investment Partnership will seek to raise $30 billion to invest in data centers and infrastructure.”
“BlackRock, Microsoft and United Arab Emirates state-backed investor MGX are partnering on a new artificial-intelligence infrastructure fund that aims to raise $30 billion to invest in data centers and related power infrastructure.”
“After raising the private-equity capital from investors, the partnership could deploy up to $100 billion in total capital when including debt financing, the parties said Tuesday. Most of the infrastructure investments will be made in the U.S.”
“The systems powering new AI products are highly energy intensive. The recent frenzy to build data centers to serve the exploding demand for AI is causing a shortage of the parts, property and power that the sprawling warehouses of supercomputers require.”
“Microsoft and other big tech players have been investing billions in their own data centers. The new partnership—dubbed the Global AI Infrastructure Investment Partnership—will pool capital from investors like pension funds and insurance companies that seek steady, modest returns from private infrastructure investments.”
I’ve of course discussed Microsoft’s increasing efforts to partner with financing partners in the Middle East. Now those efforts are taking shape in terms of partnerships with private equity firms in the US and abroad, to fund the massive investments involved.
This of course is all part of the ongoing AI capex investments running ahead of their eventual monetization I’ve discussed. Not to mention the rising ‘AI Table Stakes’, both in terms of the software and hardware data center infrastructure, but the underlying capabilities of the LLM AIs themselves. Not to mention the Power investments around them.
The above allusions and aspirations to ‘Orion’, ‘Stargate’ and other code-names, are. rapidly close to becoming clearer with o1 ‘AI Reasoning’ the current next step. And we’re seeing more on how these companies plan to get there and beyond.
It’ll be a cool Winter indeed. Stay tuned.
(NOTE: The discussions here are for information purposes only, and not meant as investment advice at any time. Thanks for joining us here)