The Roman Wars and the Tech Wars

The history of Roman Empire and its legendary emperors like Julius Caesar, Augustus Caesar, Nero Claudius Caesar etc. never fail to fascinate the public. The tsar of big- tech world like Mark Elliot Zuckerberg is no exception in this regard.

Facebook’s (Meta’s) Mark Zuckerberg has earlier told the New Yorker about his particular fascination with the Roman emperor, Augustus Caesar. “Basically, through a really harsh approach, he established 200 years of world peace,” Zuckerberg explained. “What are the trade-offs in that? On the one hand, world peace is a long-term goal that people talk about today …” “On the other hand, that didn’t come for free, and he (Augustus) had to do certain things” Zuckerberg said.

Augustus Caesar reigned as the first Roman emperor from 27 BC until his death in AD 14. His reign was at one of those extraordinary junctions in history – when Rome’s republic teetered, crumbled, and reformed as the empire. The reign of Augustus initiated an imperial cult, as well as an era of imperial peace (the Pax Romana) in which the Roman world was largely free of armed conflict. “He had to do certain things” is a relaxed way of Zuckerberg describing Augustus’ brutal and systematic elimination of political opponents. The Roman historian Tacitus had something to say in this regard: “solitudinem faciunt, pacem appellant”. They make a desert and call it peace. 

As Alter Egos go, Augustus Caesar is a bold one for Mark Zuckerberg. Henry Tricks writes in his Schumpeter column: “Both stopped at nothing to build empires—though unlike the impetuous Mr Zuckerberg, Augustus’s motto was “make haste slowly”. Both gave the illusion of sharing power (Augustus with the Senate, Mr Zuckerberg with shareholders) while wielding it almost absolutely. The Roman emperor is Mr Zuckerberg’s role model. In a recent podcast he used the 200-year era of stability ushered in by Augustus to illustrate why he is making Meta’s generative artificial-intelligence (AI) models available in a way that, with some poetic licence, he calls open source.”

Like Augustus conquered the Roman world, Zuckerberg wants to conquer the Tech world and the AI world with his generous and predatory give away of Meta’s crown jewel Llama 3.1, a freely available large language model (LLM) whose most powerful version, he says, rivals the top offering from the Open AI, maker of Chat GPT, a closed-source AI stack.

Llama 3.1 consists of Meta’s largest generative AI model to date, 405B, and updates to the 70B and 8B versions. Creating a large model like 405B comes at a steep cost. Meta said that to train Llama 3.1 405B, it used more than 16,000 of Nvidia’s H100 GPUs. Those AI chips cost between $25,000 to $40,000 each depending on the configuration, meaning Meta spent up to $640 million to train the new model. There will be many other development costs. Meta’s Llama 3.1 is certainly an eye-catcher. The biggest version has 405 billion parameters (a common definition of LLM power), almost six times those in its predecessor version. 

As per Meta, Llama 3.1 405B is in a class of its own, with unmatched flexibility, control, and state-of-the-art capabilities that rival the best closed source models. Our new model will enable the community to unlock new workflows, such as synthetic data generation and model distillation. 405B is the first openly available model that rivals the top AI models when it comes to state-of-the-art capabilities in general knowledge, steerability, math, tool use, and multilingual translation.

In his letter of July 23, Zuckerberg says open-source Llama is good for Meta, good for the world and therefore a platform that will be around for the long term. “One of my formative experiences has been building our services constrained by what Apple will let us build on their platforms. Between the way they tax developers, the arbitrary rules they apply, and all the product innovations they block from shipping, it’s clear that Meta and many other companies would be freed up to build much better services for people if we could build the best versions of our products and competitors were not able to constrain what we could build. On a philosophical level, this is a major reason why I believe so strongly in building open ecosystems in AI and AR/VR for the next generation of computing.”

Free access to Llama 3.1 is no philanthropy by Meta or Zuckerberg. Meta’s business and revenue model is based on advertisements and not on subscription revenue. So, Meta has no risk of cannibalising itself by making its state-of-the art AI stack open source. By making its large language models available for free, Meta will be able to undercut the prospects of its rival tech giants.

Like his hero Augustus, Mark Zuckerberg has battles to fight to make his position secure in the tech world.

[Photo by Anthony Quintano, via Wikimedia Commons]

The views and opinions expressed in this article are those the author.