• The Signal
  • Posts
  • Sunday Signal: Zuck building AGI, career planning and 10,000 hours

Sunday Signal: Zuck building AGI, career planning and 10,000 hours

Author

Alex Banks
January 21, 2024

Hey friends 👋 Happy Sunday.

Here’s your weekly dose of AI and introspection.

AI Highlights

Meta is currently training their next-gen model Llama-3. Their long-term vision is to build general intelligence, open-source it responsibly, and make it widely available so everyone can benefit.

Alex’s take: I was surprised at how much compute power Meta plans to assemble. 350,000 H100s by the end of the year and an overall 600,000 H100 equivalents of compute. Nvidia is licking their lips right now.

YC's W24 cohort kicked off last week. Sam Altman, the CEO of OpenAI, gave an introductory talk stating that GPT-5 and AGI will be achieved 'relatively soon'.

Alex’s take: It’s more important than ever before to create something that makes sense alongside the ranks of GPT-5 and AGI. Don't build companies that address GPT-4 limitations when OpenAI is already halfway through training GPT-5.

The title of the piece at hand: “DeepMind co-founder Mustafa Suleyman warns AI is a ‘fundamentally labour replacing’ tool over the long term.” Yet so was the wheel, the printing press, the spinning jenny, the diesel engine, the personal computer and the smartphone. The list goes on.

Alex’s take: I really enjoyed Daniel Jeffries’ analysis of this sensational headline. Big tech likes to stoke the AGI doom narrative to build hype around their products. This then drip feeds into the mainstream media. They love pushing AI panic as it fits their agenda also—to generate clicks.

These doomsday predictions are, as Daniel puts it, ‘one dimensional’, akin to the thinking of a child. It has no consideration for the iterative approach of learning from new technologies that make them smarter, safer and better for all of humanity.

AGI will arrive and it will absolutely change the world. But not as much as we expect, just like every technology before it. The risk lies not in AI replacing you, but in a small group of individuals controlling the AI to control the narrative it feeds you.

Technophobia and moral panic aren’t something new. It’s very common when a new technology comes around. What’s important is we reach a place of balance and position humanity through an open lens.

“…fear people who think open is dangerous, when open is the foundation for all greatness.”

1 Article I Enjoyed

Marc Andreessen is the Co-Founder and General Partner of the venture capital firm Andreessen Horowitz.

He also co-created the Mosaic internet browser and founded Netscape, which paved the digital path for the modern internet.

‘Pmarchive’ is an archive of the best articles from his 2007 blog. I particularly enjoyed part 1 of his guide to career planning: “Opportunity”.

My favourite takeaways:

  • Do not plan your career. Career planning = career limiting.

  • Instead of planning your career, focus on developing skills and pursuing opportunities.

  • Be continuously alert to opportunities that present themselves to you spontaneously, when you happen to be in the right place at the right time.

  • If you don’t jump all over an opportunity, someone else generally will and it will vanish.

  • The world is a very malleable place. If you know what you want, and you go for it with maximum energy and drive and passion, the world will often reconfigure itself around you much more quickly and easily than you would think.

1 Idea I Learned

In episode 333 of the Lex Fridman podcast, Andrej Karpathy stated:

“Beginners are often focused on what to do. The focus should be on how much you do.”

You have to pick the thing you care about and put in 10,000 hours of work.

It doesn’t matter as much where you put it.

It could be:

  • Learning Python

  • Understanding machine learning

  • Educating yourself on large language models

You’ll make mistakes and waste time doing something wrong.

And that’s okay.

What’s important is you’ll learn and improve. You’ll develop the intuition of how to figure it out. You’ll accumulate scar tissue so that the next time you come to a similar situation, you’ll have learned from it and know how to solve the problem.

The beauty of this concept is you can apply this rule to anything. It doesn’t have to be AI.

Apply 10,000 hours of deliberate effort, and you’ll become an expert at it. That’s an incredibly freeing thought.

So the question we should be asking ourselves is not “Am I picking the right thing to do”, but instead “Am I spending 10,000 hours”.

Focus on the work.

1 Quote to Share

Steve Jobs on asking for something:

“Most people never pick up the phone and call. Most people never ask.

And that's what separates sometimes the people who do things from people who just dream about them.

You’ve got to be willing to act and fail. You’ve got to be willing to crash and burn.”

If you’re afraid of failing, you won’t get very far.

If you don't ask, you don't get. Worst case, it's a no. Best case, it changes your life.

1 Question to Ponder

What did I get done this week?

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

See you next week,

Alex Banks