Good morning. Wishing you all a very happy Thanksgiving, full of good food and better company.

We will be off tomorrow, but we’ll be back on Friday with a special cybersecurity breakdown of digital footprints.

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

In today’s newsletter:

AI for Good: Plant-based meats

Moving away from meat-based diets could have a significant impact on greenhouse gas emissions, reducing carbon and helping spare some of the resources that are currently expended on maintaining flocks of animals. 

But the plant-based alternatives tend to not taste great. 

What happened: So, some researchers at Stanford are applying machine learning and generative AI to a series of tests and experiments designed to find the closest (plant-based) textural alternative. 

The team then affirmed those equations with a series of human taste tests, which largely supported the conclusions derived by the algorithm. 

Why it matters: Though the team has plenty more work to do, here, the goal now is empirically-tested, AI-generated recipe development:

“Instead of using a trial-and-error approach to improve the texture of plant-based meat, we could envision using generative artificial intelligence to scientifically generate recipes for plant-based meat products with precisely desired properties,” the researchers wrote in a paper that was published in Nature.

The #1 Rated AI Notetaker

Why are 26,000+ companies using Fathom to boost productivity?

Start using Fathom today (at zero cost).

Elon Musk’s robotic arm 

One of the big promises of brain-computer interfaces (BCIs) centers around Star Wars-esque prosthetics. Though we’re not there yet, these systems today use sensors and artificial intelligence to decode and process brain activity, which is then read by computers; some, like Elon Musk’s Neuralink, involve highly invasive surgeries. Other BCI systems simply sit, either on top of the brain, or on the exterior of a patient’s head. 

What happened: Neuralink, which is still in the midst of its initial human trial, says it has received approval to launch a second human trial, this one focused on testing the link between its BCI system and robotic arms. 

The idea is that, if the chip can read and process brain activity enough to control a computer cursor, it could also be applied to controlling prosthetic arm and finger movements. 

But, as Dr. Nicholas Vachicouras, the CEO of brain implant firm Neurosoft, has told me: “the reason why this is not widespread today is a hardware problem.” The wet environment of the brain makes for an incredibly complex engineering challenge, even if the software component is perfect. 

Some context: Neuralink has also faced plenty of scrutiny, due both to its approach and its alleged practices. The company has been accused — and, later, cleared — of animal welfare violations. 

The Physicians Committee for Responsible Medicine said last year: “It seems obvious to everyone but Elon Musk that Neuralink’s device is unsafe and dangerous.” 

🚀SmallCon: Free Virtual GenAI Conference featuring Meta, Mistral, Salesforce, Harvey AI and More🚀

Learn what it takes to build big with small models from AI trailblazers like Meta, Mistral AI, Salesforce, Harvey AI, Upstage, Convirza, Nubank, Nvidia, and more. Dive into cutting-edge tech talks, live demos, and interactive panels on topics like:

Save your spot for SmallCon to learn how to build the GenAI stack of the future!

Full agenda and registration here

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.

Anthropic launches new method to connect LLMs with data sources 

Anthropic this week unveiled and open-sourced a new standard for connecting Large Language Models (LLMs) with data sources: the Model Context Protocol (MCP).

The problem: As Alex Albert, Anthropic’s head of Claude relations, wrote, “getting LLMs to interact with external systems isn't usually that easy.”

The solution: MCP acts as a universal, open standard for this connection, allowing developers to ”build secure, two-way connections between their data sources and AI-powered tools.” It follows a client-server architecture, according to Anthropic, enabling developers to “expose their data through MCP servers or build AI applications (MCP clients) that connect to these servers.”

Dhanji R. Prasanna, CTO at Block, an early adopter of the MCP, likened the protocol to the “bridges that connect AI to real-world applications, ensuring innovation is accessible, transparent, and rooted in collaboration.”

Zoom is becoming an ‘AI-first’ company 

Zoom Video Communications Inc. — that pandemic-era stalwart of digital connection — announced this week that it is dropping the “video” from its legal name as an aesthetic part of its effort to become an “AI-first company.” 

The details: In a blog post, Zoom CEO and founder Eric Yuan said that, “from day one, our mission has been to deliver happiness to our customers.”

The AI companion sits at the center of the Zoom Workplace, a new product offering that starts at $13 per user per month. 

The vision: In the blog post, Yuan said that, by summarizing meetings, drafting emails and helping with meeting prep, Zoom’s AI assistant will decrease workloads to the point that a four-day workweek will become possible. 

The longer-term vision remains the same as it was in June, when Yuan fantasized about a near-term future full of AI-generated digital clones that users can send out into the ether to take the bulk of their meetings and handle the bulk of their work. 

Neither the training data nor the energy costs of Zoom’s generative model are known. 

Reporting third-quarter earnings recently, Zoom detailed a 3.6% year-over-year increase in revenue to $1.18 billion, something a JP Morgan analyst called a “muted” recovery. Shares of the company, up 15% for the year, fell around 7% on Tuesday. 

I’d love to see generative AI leveraged to usher in a 4-day workweek, I just don’t really see that happening. What seems more likely — and what, in certain cases, is already happening — is that employees will become required to use generative AI to handle larger workloads as employers reduce staff. Unless this all was paired with specific regulations, which seems unlikely, I don’t really see the bridge to a shorter workweek here.  

And on the digital clones front, while we’re starting to see some progress in that general vicinity with the ever-increasing push toward “agents,” the tech isn’t ready for something like that, yet. And, if it does get there, there are a lot of questions that ought to be addressed first, spanning everything from security and privacy guarantees to guardrails around AI actions and decisions and a lack of clarity around the liability and accountability for the people using these systems if things were to go wrong. 

Then, you’ve got a kind of central issue with this idea of digital clones — if the bulk of work is being done by AI clones, why would any company employ any people at all? Why not just reduce staff down to a handful of executives, and hundreds of their digital clones? And if that’s the case, the impact on the economy — with or without some version of a universal basic income, which is dubious in and of itself — would be enormous, simply by removing purchasing power from the bulk of consumers. 

This wouldn’t be good. 

All that to say, there’s been a trend — and I think Zoom seems anxious to participate in it — of tech CEOs painting futuristic portraits filled with seemingly high-impact buzzwords. “Agents,” “digital clones” and “superintelligence” all sound, at the very least, intriguing, but they’re unserious and uncritical terms and phrases — the tech isn’t there, the science isn’t there … and if it was, there are too many unanswered questions and unresolved obstacles to call these visions legitimate. 

Instead, what we have is a healthy dose of hype.

🤔 Your thought process:

Selected Image 2 (Left):

💭 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-generated music:

40% of you said: ‘pay the human artists, then it’s fine.’ 25% of you said it’s never fine; the rest were pretty evenly split between loving it and not really caring too much.

Something else: