Moats and the AI Arms Race

It's a free for all.

Every day Refind picks 5 links from around the web that make you smarter and tailored to your interests. Loved by 100k+ curious minds. Sign up here for their free newsletter

Friends,

An internal memo leaked from a Google researcher suggests that open-source artificial intelligence (AI) could soon outpace both Google and OpenAI. The confidential document highlights the tremendous progress made in open-source AI, including the ability to run fundamental models on mobile devices and fine-tune personalized AI on a laptop in just a few hours. While Google and OpenAI may have a slight edge in terms of model quality, open-source models are reportedly faster, more adaptable, and more capable, pound-for-pound.

The memo warns that Google does not have any secret sauce and should instead collaborate with external players to integrate third-party solutions. With free and unrestricted alternatives available, it cautions that people may not be willing to pay for restricted models. The piece concludes by advising Google to prioritize making small variations of models more easily iterated upon and focus on enabling third-party integrations now that the possibilities in the <20B parameter regime have been demonstrated.

This internal document shines a spotlight on the rapid progress of open-source AI and the need for Google to keep up with external players. It is a clear call for the tech giant to adapt and collaborate in order to stay ahead of the curve in the competitive world of AI. As the race for AI innovation continues, it remains to be seen how Google will respond to the challenge posed by the open-source community.

The first thing that Google is doing is to stop its AI researchers from sharing their work with the outside world. OpenAI's ChatGPT kept up with Google by reading their scientific papers. Google will take advantage of its own AI discoveries and only share papers after the lab work has been turned into products. The policy change is part of a larger shift inside Google, which is focusing on launching products quickly and has overhauled its AI operations to protect its core search business, stock price, and future.

And this is the theory of moats in technology and business:

The term "economic moat," popularized by Warren Buffett, refers to a business's ability to maintain competitive advantages over its competitors in order to protect its long-term profits and market share. Just like a medieval castle, the moat serves to protect those inside the fortress and their riches from outsiders.

At the moment we are in a free for all to win the AI Arms race. And while we may think that OpenAI is winning the race, they and Google will be competing with these flexible open-source LLMs which will allow many others to play in the space. It is not guaranteed that Google will win this race - it’s not guaranteed that OpenAI will win this race.

I’ll go so far as to say that we will look back on this simple ChatGPT interface that got us excited about what was possible. Once imaginations build upon LLaMA (Meta’s open-source LLM) as well others, we will see an astonishing change in the world.

According to the Oxford English Dictionary,

"A moat is a deep wide channel that was dug around a castle, etc. and filled with water to make it more difficult for enemies to attack."

Unfortunately, no one quite has a castle just yet.

Stay Curious - and don’t forget to be amazing,

Rahim

Here are my recommendations for this week:

Now

1. The Great Progression 2025-2050 - For far too long we've been looking at our current situation and the coming decades through the lens of the past. Most people are stuck in the familiar default frame that sees many of our old systems breaking down in the face of myriad challenges like climate change, polarized politics, economic and social inequities, the paralysis of liberal democracies, and the rise of authoritarian states. This piece has been recommended to me a couple of times now, and it’s very much on the money of what I believe. Highly recommended.

2. Longevity researchers share 9 of the best antiaging hacks — and a few that are total scams - The obvious ones worth reminding yourself and NMN, Meformin which may be game-changers, if you subscribe to the work of David Sinclair (as I do)

3. 5 Ways to Future-Proof Your Career in the Age of AI: an excellent overview of being human in the face of tech. Be human. Be unique.

4. You Have a New Memory: What do we mean when we say the internet is reading our minds?

5. The Future of Jobs Report 2023: The Future of Jobs Report 2023 explores how jobs and skills will evolve over the next five years. This fourth edition of the series continues the analysis of employer expectations to provide new insights on how socio-economic and technology trends will shape the workplace of the future. 300 page report for the nerds!

Next

1. Scary 'Emergent' AI Abilities Are Just a 'Mirage' Produced by Researchers, Stanford Study Says: In a new paper, Stanford researchers say they have shown that so-called "emergent abilities" in AI models—when a large model suddenly displays an ability it ostensibly was not designed to possess—are actually a "mirage" produced by researchers.

2. ‘The Godfather of A.I.’ Leaves Google and Warns of Danger Ahead: For half a century, Geoffrey Hinton nurtured the technology at the heart of chatbots like ChatGPT. Now he worries it will cause serious harm.

3. AI Singers Are Unnervingly Good and Already Ubiquitous: The software that cloned Drake and the Weeknd is easy to use—and impossible to shut down.

4. I Cloned Myself With AI. She Fooled My Bank and My Family.: This WSJ columnist replaced herself with AI voice and video to see how humanlike the tech can be. The results were eerie.

5. My Weekend With an Emotional Support A.I. Companion: Pi, an A.I. tool that debuted this week, is a twist on the new wave of chatbots: It assists people with their wellness and emotions.

Every day Refind picks 5 links from around the web that make you smarter and tailored to your interests. Loved by 100k+ curious minds. Sign up here for their free newsletter

If you enjoyed this edition of Box of Amazing, please share and help me grow this group. If you share it on Twitter or LinkedIn, please tag me so that I can thank you.