If you’re new here, this is the latest edition of the Build Mode newsletter, where we deliver insights and strategies from the leaders shaping the future of AI, designing the future of work, and building the most important companies of the next decade. Subscribe here to get the top insights in your inbox every week.
THE BIG IDEA
Why so many fintech entrepreneurs are cautious about GenAI
Imagine for a moment that you call up your financial advisor to check on your portfolio, and instead of the usual human on the other end—who is, to be fair, probably younger than you and may not have personally lived through a recession—you end up connected to a chatbot who has the economic history of the world in its training data.
Let’s say, for the sake of argument, that this is a personal financial advisor powered by a large language model who not only knows you and your spending patterns intimately, but has also internalized the latest market data. Compelling!
The AI suggests a moderately conservative investment strategy while setting aside several thousand dollars per year so you can take your daughter to see Taylor Swift every time there’s a new tour.
Are you taking the chatbot’s advice?
If you’d be hesitant, you’re with the majority of Americans. Only 37% of U.S. adults would take investment advice from an AI tool, according to a CNBC study from December; 51% said they have little or no trust in these tools.
"People have this kind of dream of a PFM — the personal financial manager," said neobank Mercury co-founder and CEO Immad Akhund. "It just hasn't worked out so far.”
As an industry, financial services is regulated like crazy. It handles extremely sensitive data. It requires vast customer support resources. And it involves layers and layers of manual, human-led processes. Banks have remained stubbornly labor-intensive over the years. For all of these reasons, finserv is ripe for generative AI disruption. And, for the same reasons, fintech entrepreneurs and CIOs at banks are wary.
"You can't afford in financial services for [AI] to give you a random tip," said Sachin Dev Duggal, founder of Builder.ai. "The bank would be running an indefinite amount of liability."
Which means that for all the opportunities here—and McKinsey analysts say there are many—risk management is going to be key.
As Paolo Sironi, an IBM researcher specializing in banking recently put it on LinkedIn: “Every banker must be an AI risk manager.”
CHART OF THE WEEK
McKinsey: Generative AI is a $340B opportunity for banks
Applying generative AI in financial services will be different than traditional AI/ML initiatives. Banks have been using regular old machine learning and algorithmic decision-making for decades. Wisely, they are cautious about the risks involved with generative AI.
That said, a recent McKinsey analysis suggests that there is an annual potential of $200 billion to $340 billion for banks to capture, largely from increased productivity. The biggest economic impact will be in the corporate and retail banking sectors.
But the challenges are many. First up: finding the right talent. As McKinsey has it, “scaling up gen AI has unique talent-related challenges, whose magnitude will depend greatly on a bank’s talent base.”
CLIENT SPOTLIGHT
Building the world’s first augmented reality laptop with Spacetop
Two former Magic Leap executives wanted to build the world’s first augmented reality laptop. The vision was right out of Sci-Fi: Blur the boundaries between virtual and physical, offering a computing experience far beyond conventional screens. Sounds easy, right?
Here’s the big question with spatial computing: How do you combine the portability of smartphones and the productivity of multi-monitor setups? Traditional laptops, while versatile, fall short of being truly mobile.
Enter Sightful, and their signature product: Spacetop, the world’s first AR laptop.
It’s like using a normal laptop — except instead of a screen, an enormous 100” curved display appears in front of you. You can move and pin windows, zoom in with simple gestures, and pin windows side-by-side. That means you get the ease and mobility of a laptop with the ergonomic relief and productivity boost of a home monitor setup.
Creating such a game-changing piece of hardware presented many challenges. Chief among them was building an operating system that could integrate a seamless user experience. That’s where A.Team came in.
WATERCOOLER
Do we even want this?
Last week, Reid Hoffman posted a video on X of him talking with an AI version of himself. He fed it all of his books and podcasts into GPT-4, and then used some kind of AI video avatar to animate the responses, although he wasn’t explicit about which tool he used.
According to Hoffman, the premise is that deepfaked himself in order to inspire “self-reflection, new insights into my thought patterns, and deep truths.”
Also last week: It turned out that a racist rant by a Baltimore school principal, for which he was placed on administrative leave, was a deepfake. The disgruntled former employee who made the video was arrested. It’s like a bad episode of Community.
This seems like a good time to ask: Do we even want AI with these kind of deepfake capabilities?
Yes, there are some useful applications if you stretch your imagination. Although Hoffman’s idea that it will help us “be more human” feels like wishful thinking.
Seriously though: How does this not make the internet worse? When you have no idea whether a video is real, whether a person actually said what they said, or even if the person you’re zooming with is real. Forget Community — we’re already living in an episode of Black Mirror.
EVENTS
Gen AI Salon: The Future of Finance
The giants of American banking are betting big on AI to revolutionize their operations. But, despite their enthusiasm, most are still grappling with the basic implementation of AI products internally and are struggling to define what success looks like in this new era.
On May 29th, join an invite-only group of fintech founders, VCs, researchers, and decision-makers in NYC for a Generative AI Salon dedicated to the future of finance. You’ll hear from leaders like Lisa Carmen Wang, founder & GP of Bad Bitch Empire, Nicole Casperson of Fintech is Femme, and a panel moderated by M13’s Morgan Blumberg.
Apply to attend in-person at our NYC Clubhouse or join virtually.
DISCOVERY ZONE
Just in time for Mother’s Day, you can now officially buy your mom her very own flame-throwing robot dog for less than 10K USD. It even connects to Bluetooth!
If you have an idea for an AI tool we should feature here, put a comment on Joe’s post calling for nominations on LinkedIn.
MEME