Good morning. The AI and consumer ecosystems are fueled by data. And for years, data brokers have curated and sold expansive datasets that have been leveraged by corporations and scammers alike.

The U.S. Consumer Financial Protection Bureau has proposed a new rule that would prevent data brokerage companies from selling certain types of sensitive personal information, a massive boon to U.S. privacy protections.

— Ian Krietzberg, Editor-in-Chief, The Deep View

In today’s newsletter:

• 🧠 AI for Good: Brain mapping

• 👁️‍🗨️ Researchers unveil ultra-efficient AI processor

• 🌎 China bans shipment of key materials to US in retaliatory move

• 🚨 Report: AI and cybersecurity in the enterprise 

AI for Good: Brain mapping

We’ve talked often about how much we don’t understand the human brain. It has been described by some neuroscientists as the most complex object in the known universe, due mainly to the sheer quantity of things going on within it; the human brain is estimated to have some 86 billion neurons, 85 billion other cells and 100 trillion synapses, or connections, tying it all together.  

This lack of understanding is why many neurological diseases remain uncured by modern medicine; science cannot cure what it doesn’t understand. 

What happened: Scientists have been trying to understand the brain for decades. The latest approach involves applying computers (and artificial intelligence) to literally map brains. In October, a Princeton-led team of scientists across more than 120 institutions succeeded in mapping the brain of an adult fruit fly. 

• Researchers had previously mapped the brain of a C. elagans worm, which has 302 neurons, and the brain of a larval fruit fly, which has 3,000 neurons. But the adult fruit fly, with 140,000 neurons, was a massive step up. 

• The process, which has been ongoing for more than five years, first involved taking 21 million pictures of a fly brain; AI systems were then deployed to turn those images into a labeled, 3D map. This process required 33 person-years of proofreading and annotation; it would have required 50,000 person-years without AI. 

The full fly brain is now freely searchable online. 

Why it matters: “Any brain that we truly understand tells us something about all brains,” Sebastian Seung, Princeton’s Evnin Professor in neuroscience, said in a statement. “With the fly wiring diagram, we have the potential for an unprecedented, detailed and deep understanding.”

Closing Soon: This Startup is Giving AI The Upgrade it Needs

The data center market was valued at $194.8 billion in 2022, and is expected to grow at a CAGR of 10.9% through 2030. With generative AI expanding rapidly, over 90 zettabytes of data will need to be created and stored in 2025, bringing issues of data storage into the spotlight. That's why Atombeam’s tech is poised to disrupt how machines communicate and store data, reducing data size up to 75%.

• Atombeam's patented (AI-powered) Neurpac software can make networks up to 4x faster, while also making them more secure — helping clients avoid billions in expensive hardware upgrades.

• And this applies to generative AI, too. Atombeam's tech allows generative AI applications to run faster (while consuming less power), something that could dramatically supercharge AI efficiency, transforming the whole landscape.

Plenty of industry players have taken note of Atombeam's approach; in addition to partnering with Nvidia, the company has contracts in place with the U.S. Air Force and Space Force to develop Neurpac for use in military satellites.

Atombeam’s raise just hit $10M - but this round is closing in December. Become an Atombeam shareholder for just $8/share before this round closes.

Researchers unveil ultra-efficient AI processor

Artificial intelligence isn’t magic. It’s computation. According to Nvidia, AI models — referring here to neural networks — are essentially “a mathematical lasagna, made from layer upon layer of linear algebra equations.” 

GPUs, the chips that Nvidia has made a fortune selling, have supplanted CPU chips to become the foundation of modern generative AI for a simple reason: unlike CPUs, which have fewer cores and can’t perform parallel computing, GPUs “pack thousands of cores, tiny calculators working in parallel to slice through the math that makes up an AI model.”

But AI models keep getting more complex, and researchers have been exploring potential alternatives. 

What happened: Researchers at MIT on Monday unveiled a new photonic processor that, overcoming older limitations to photonic hardware, can “perform all the key computations of a deep neural network optically on the chip.”

• Photonic processing uses light instead of electrons to process data; in tests, this device was able to perform all the necessary computations for a machine learning classification task “in less than half a nanosecond” while maintaining performance on par with traditional hardware. 

• Several years ago, photonic hardware was limited in processing machine learning tasks due to its inability to handle nonlinear operations; here, the team of researchers was able to overcome this barrier by designing devices that combine electrons and light directly on the chip. 

The chip can be fabricated using normal commercial foundry processes. 

What it means: The chip can enable faster and far more energy-efficient training and operation of AI models and other computationally intensive technologies, such as lidar. 

Never Research a Company Again

Bounti’s AI saves over 10 hours/week of painful account research and meeting prep, empowering you and your team to sell more effectively through every channel and stage of the funnel.

Deliver the right message to your ideal buyer—driving real engagement at a fraction of the cost of other solutions.

Research: Get a cheat sheet for every account that matters to you. In minutes, you’ll have everything you need to make your next conversation the reason they buy.

Create: Our AI will generate custom pitches and any type of content you want, all fully customizable to what you need to land your deal.

Engage: Have the confidence to win every conversation with detailed messaging and position that your buyers actually care about.

Get Started For Free Today

• China’s AI balancing act — beating the U.S. but keeping the tech from threatening Beijing’s rule (CNBC).

• Meta says it’s mistakenly moderating too much (The Verge).

• Global auto industry in turmoil as prospect of US tariffs looms (Semafor).

• Uber’s big plan for India: buses (Rest of World).

• FTC bans location data company that powers the surveillance ecosystem (404 Media).

If you want to get in front of an audience of 200,000+ developers, business leaders and tech enthusiasts, get in touch with us here.

• Software Development Manager: Oracle, Remote

• Applied AI/ML - Vice President: JPMorgan Chase, New York, NY

• Amazon said it is building a supercomputer, in collaboration with Anthropic, that will be composed of “hundreds of thousands” of its Trainium2 chips. Amazon expects this to be the largest AI compute cluster in the world once completed. (Amazon recently doubled its investment in Anthropic to $8 billion).

• The U.S. Consumer Financial Protection Bureau proposed a rule that would prevent data brokers from selling financial and personal information about Americans, including social security and phone numbers.

China bans shipment of key materials to US in retaliatory move

Just days after the Biden administration enacted its latest round of export restrictions against China, the country retaliated by outright banning the shipment of key semiconductor and battery materials to the U.S. 

The details: China’s Ministry of Commerce said Tuesday that it is banning the shipment of so-called “dual-use” materials — materials that could also be used for military purposes. 

• The ban includes gallium, germanium, antimony and other “superhard materials.” Both gallium and germanium are key to semiconductor production, and according to the Center for European Policy Analysis (CEPA), China enjoys a “near-monopoly” on both. 

• Researchers at the U.S. Geological Survey recently developed a new forecasting model that estimated there would be a $3.4 billion decrease in U.S. GDP if China were to implement a total ban on shipments of gallium and germanium. 

They said that economic losses would be “concentrated” in the semiconductor industry, “which would account for more than 40% of the net loss.”

The landscape: This latest escalation of the trade war between the U.S. and China comes just weeks before President-elect Donald Trump — who has promised to place tariffs of up to 60% on Chinese goods — takes office. 

Report: AI and cybersecurity in the enterprise  

Even as that mix of Big Tech corporations and startups attempt to get enterprises on board with their generative AI products, cybercriminals (and cybersecurity officials) have already adopted the tech. And a new report out of cybersecurity firm Ivanti found that, currently, GenAI “stacks the deck in favor of threat actors, who can exploit AI capabilities to continuously improve their phishing attempts.”

The details: The report, the result of surveys with more than 14,000 security and IT professionals, noted plenty of optimism from cybersecurity professionals around the positive impact generative AI will have on security, even as AI-powered attacks continue to ramp up. 

• Nearly half of those surveyed view generative AI as a net positive for security; the tech is already being used for enhanced threat detection and rapid, real-time responses to potential security breaches. 

• Still, according to the report, cybersecurity professionals remain highly concerned about the ways in which generative AI is enhancing phishing attacks, exploiting software vulnerabilities, enhancing ransomware attacks and enabling denial-of-service attacks. 

“Attackers are now using generative AI to craft highly believable content to lure victims — all at high scale and low cost,” the report reads. “This threat vector will become even more powerful as attackers further personalize their phishing messages based on data found in the public domain.”

57% of those surveyed say they use anti-phishing training to protect against these attacks; only 32% believe that this approach is “very effective” against the threat at hand. 

“Attackers have greater resources, more time and dogged persistence,” Brooke Johnson, a senior vice president and chief legal counsel for Ivanti, said in a statement. “Given the nature of the threat — and how dynamic and complex it is — we as a security community have to work together and share knowledge to defend ourselves against this very complex, organized enemy.”

The optimists: Faced with a shortage of cybersecurity professionals, generative AI offers a way for short-handed security teams to still maintain their defenses, according to the report.  

• 90% of those surveyed believe generative AI benefits security teams as much or more than threat actors. 

• Respondents, were, however, six times more likely to say that AI tools will primarily benefit employers, rather than employees. 

According to the report, AI-powered cyber threats will continue to escalate, and they can only be met by a combination of human professionals and AI-powered defenses. 

“Separate layers can no longer prevent AI-powered attacks from breaching your enterprise. In the hands of attackers, gen AI breaks down defenses by simultaneously breaching networking servers and layers,” Robert Grazioli, Ivanti’s chief information officer, said in a statement.

“The way forward is not to eliminate the human element, but to empower humans with AI assistants. These assistants, in collaboration with other AI assistants, can gain a more holistic, cross-disciplined view of the organization.”

While the threats are prominent for enterprises, individual people — as we reported last week — are also at risk (and often, personal cyber vulnerabilities can become exploitable enterprise vulnerabilities).  

AI-powered phishing, social engineering attacks and fraudulent, deepfake identity hijacking all remain real, active threats that can only really be countered by a shift in culture — we cannot by default trust anything we encounter in the digital world, be it a phone call, a Facebook post or an email. 

We must verify in order to trust, especially when it comes to sensitive information.

🤔 Your thought process:

Selected Image 1 (Left):

• “The man in the second picture has his legs crossed weird. The shoe soles for both people in the second picture are odd. The only issue i see with the first picture is the person on the right, their hand looks a bit odd holding the wine glass.”

Selected Image 2 (Right):

• “Damn, this one was really tricky!”

💭 A poll before you go

Thanks for reading today’s edition of The Deep View!

We’ll see you in the next one.

Here’s your view on AI-powered flood prediction:

34% of you said that localized, AI-powered flood prediction would inspire you to evacuate where you normally wouldn’t; a quarter said it wouldn’t and the rest aren’t really sure.