Sunday Signal: Meta’s Chief AI Scientist on LLMs, my last 5 years of work, and latent demand

Hey friends 👋 Happy Sunday.

Here’s your weekly dose of AI and introspection.

AI Highlights

1. Meta's Chief AI Scientist says, “no point working on LLMs”.

During a talk at VivaTech2024 in Paris, Yann LeCun stated that he does not recommend PhD students to work on LLMs. Yann views LLMs as an 'off-ramp' from the road to reaching ultimate intelligence.

Alex’s take: An off-ramp is thinking your new hammer will cut wood. I think it’s important to remember that LLMs are next-word prediction tools trained to be conversational by imitating human responses. Yann states: ”You should work on the next generation of AI systems that lift the limitation of LLMs.”

2. Building an AI app in 9 minutes and 24 seconds.

Dhravya Shah built a full-stack AI image generation app in under 10 minutes. Speedrunning has now entered a new domain with the proliferation of AI.

Alex’s take: This was too good not to highlight. It just shows how frictionless development is becoming to help anyone create something useful for the world.

3. Harley launches collaboration with top-3 UK retailer.

We did it. I launched my new startup on Monday. We’re tackling the problem of static reports, spreadsheets, and dashboards failing to provide flexible insights into your data. Harley uses generative AI to make data analytics accessible to everyone.

Alex’s take: It’s abundantly clear that decision-makers who call the shots needed answers weeks ago. Harley processes data across multiple complex databases to provide live, actionable insights. If you’re interested in learning more, sign up here.

1 Article I Enjoyed

My Last Five Years of Work.

I thought this was a great read by Avital Balwit, Chief of Staff at Anthropic.

The next few years might see the end of traditional work due to advancements in AI.

Many professionals have long resisted the idea that AI’s capabilities across knowledge-work tasks are increasing exponentially.

But for every coding copilot, presentation producer, and Excel accelerator I see pop up, we are approaching a reality where AI might outperform humans in nearly all economically useful tasks. This transition raises serious questions about how society will cope with the loss of traditional employment and its associated values.

Society has previously had a hard time adapting to technological changes. Consider the Internet in 1994, the iPhone in 2008, and generative AI in 2022.

For all the uncertainty that falls out from thinking about these big questions, I can't help but feel excited about what’s coming around the corner.

1 Idea I Learned

The strongest sign of latent demand.

Latent demand exists when there is demand for a product or service that is not yet available.

The strongest sign of latent demand is when people hack existing platforms for unintended use cases.

As Zack Hargett points out:

• Before Loom, people hacked QuickTime and YouTube.

• Before Airbnb, people hacked couch surfing and payments.

• Before Onlyfans, people hacked Snapchat and Venmo.

If you're looking for your next startup idea, a great place to begin is looking for unintended use cases of existing products.

(Share this on X)

1 Quote to Share

Airbnb CEO Brian Chesky on potential:

“The best people in your life are the ones who see potential in you that you didn't see in yourself.”

I found this especially true when finding my partner. It was the foundation that gave me tremendous confidence.

Source: Jay Shetty Podcast

(Share this on X)

1 Question to Ponder

What can I do and speak about for hours at a time without noticing?

(Share this on X)

💡 If you enjoyed this issue, share it with a friend.

See you next week,

Alex Banks

P.S. This will make you want to go change the world.