The path to superintelligence is unclear. But Meta is chasing it anyway.

The tech giant has jumped headfirst into building a superintelligence lab, seeking to create an AI system that can far outperform human intelligence in all domains. CEO Mark Zuckerberg has attempted to court hundreds of talented engineers and researchers for the lab with eye-popping pay packages, according to The Wall Street Journal.

With Meta facing powerful competition in the red hot AI industry, the company’s commitment to superintelligence may be a bid to play catch-up and a “case of one-upmanship,” said Bob Rogers, chief product and technology officer of Oii.ai and co-founder of BeeKeeper AI. “They probably have a fear that someone’s going to end up with the ultimate model that wipes out everybody else, and they’ll end up being beholden to one organization,” he said

But without a precise vision, chasing superintelligence may be a pipe dream, he said. “Going after a vague, undefined goal is not usually the path to success,” said Rogers. “To immediately jump from where we are now – which is (large language models) being really cool language factories – and skipping (artificial general intelligence) to go straight to superintelligence is bold.”

Superintelligence, not ‘Superwisdom’

As it stands, superintelligence is only theoretical. There are a number of hurdles in the way of creating an all-knowing AI model:

• As it stands, researchers haven’t figured out how to make AI stop hallucinating, especially as we continue to challenge models with reasoning. The harder the task, the more likely it is to make a mistake, said Rogers. For an AI model to be considered superintelligent, “It’s got to be really good at not hallucinating,” he said.
• In addition to the technical obstacles, creating a system that’s that intelligent is incredibly risky, said Rogers, and could be riskier depending on who is in charge of it. The larger models get, the more unpredictable they become, he said. Plus, even if developers create kill switches, the models may be powerful enough to work around them.

Beyond the barriers and risks, the pot of gold at the end of the superintelligence rainbow may not be worth it, said Rogers. Enterprises are currently struggling to manage the large language models of today, and often find success with implementing small, niche models and agents, something which Meta’s Llama family of models happen to be well-suited for, he noted.

“I don’t think there’s a huge amount of merit in having an agent also happen to be the smartest being that has ever existed” Rogers said. “It’s not clear to me that superintelligence solves a problem that we need to solve. They’re not building superwisdom.”

AI data centers are clustered in the northern hemisphere.

Data from Oxford University, reported by The New York Times earlier this week, found that only 32 nations house AI data centers. The US, the EU and China harbor most of them, and 90% are operated by US and Chinese companies.

The reason isn’t simple geographic preference: With the colossal expenditure required for constructing such centers, it makes sense that wealthier economies have denser AI infrastructure than others, said Trevor Morgan, COO of OpenDrives: “AI is a resource.”

“There are certain parts of the world that have lots of oil, there are certain parts of the world that have a lot of uranium, but in order to get to a resource, you have to have resources,” said Morgan. “You’re talking about the tech-heavy centers of the world that also have the resources to develop a new resource.”

That doesn’t mean that enterprises everywhere else are down for the count, however “Don’t overlook creativity and the desire for some of these emerging markets to do more with less,” said Morgan.

‘Natural Progression’

And the disparity may not exist forever, regardless. Large enterprises that have the necessary cash may seek to invest in emerging markets.

• Morgan compared it with the development of the Stargate AI data center project in the small town of Abilene, Texas: “Economic centers are changing. If it’s shifted to Abilene, why can’t it shift to somewhere else?”
• Expanded overseas development may already be taking place, too. In late May, Amazon pledged to develop AI-focused server farms in Chile, New Zealand, Saudi Arabia and Taiwan.
• Building out infrastructure in those markets would likely be a boon for the AI industry broadly, said Morgan, creating a “diversity in the evolution of the technology.”

“Those who have the resources are the ones who can go and speculate and probably corner future markets. That hits pay dirt,” Morgan said. “It’s almost like the natural progression of market buildout.”

Not all quantum bits are created equal.

Amazon is seeking to patent a system for the “readout and reset of fluxonium qubits.” For reference, “fluxonium” isn’t a term out of science fiction – it is a kind of quantum bit that’s known for being more stable and having longer coherence times than others used in quantum computing.

Amazon’s system seeks to harness these bits by reading their state and resetting them after each quantum-computing operation. It uses what’s called a “readout resonator,” which extracts information from a qubit to a “quantum metamaterial” that’s made to interact with signals coming from quantum components, allowing the system to understand the condition of the qubit without disturbing it.

While it may sound esoteric, the goal is to make quantum algorithms more precise, faster and more efficient. Patents like these signal a broader interest from tech giants in overcoming the major barriers present in scaling quantum technology.

Amazon isn’t new to the quantum game. Earlier this year, the company joined the likes of Microsoft and Google in debuting its own quantum chip called Ocelot. The company claims that Ocelot can cut the costs of error correction by up to 90% compared with current approaches, representing another step towards scalable quantum computing.

And though it doesn’t operate its own quantum computer, Amazon Web Services has long offered access to quantum computing through a program called Amazon Braket, remotely allowing customers to use quantum devices from a number of different firms.

Though the quantum race is quickly picking up the pace, it’s still too nascent to see who the frontrunners will be. By filing patent applications that seek to make the tech more stable and usable, however, Amazon may be trying to get its foot in the door early.