🏆 How AI makes you better

🔼 Luckin Coffee. Their revenue has overtaken Starbucks’ in China »»

🔽 Backyard eggs in France. Millions have been warned not to eat eggs from home farms, thanks to “forever” chemicals in soil »»

💬 “It’s a cynical use of nature to promote something incredibly nature-damaging.” Britain banned 2 Toyota truck ads »»

🛫 Mexico City's hotel scene has never been better »»

👗 From investment-worthy coats to The Row bags, here’s what to score in 2023’s Black Friday fashion sales »»

💎 An exceedingly rare 1960s Rolex Space-Dweller is coming to auction »»

Why do all chefs cross their arms in photos? »»

Be warned: Flights are getting bumpier »»

Somehow, Hoka's best shoes are getting better »»

What to know about OpenAI’s failed coup »»

And: How to talk about OpenAI’s drama at Thanksgiving »»

Say goodbye to those absurd side-effects readouts in American pharma commercials »»

London’s Restaurant Story is getting ready to unveil its £2.5m makeover »»

Binance’s Dubai-based CEO stepped down as part of a US$4b settlement with the United States Department of Justice »»

The 19 best nail polish gift sets »»

United Airlines is adding a record number of flights to Asia and the Pacific because of surging demand »»

How to dance without looking awkward »»

If your IG is public, anyone can now download your Instagram Reels »»

A new beachfront InterContinental resort is coming to Kota Kinabalu, on the Malaysian side of Borneo »»

The UK has vowed to pump more cash into AI, as it looks to build the ‘next Silicon Valley’ »»

AppFlowy is an open source Notion clone with some big-name backers »»

Attention Londoners: here’s your first look inside the Piccadilly Line’s new air-conditioned, walk-through trains »»

It’s not Black Friday yet, but online shoppers have already spent more than last year »»

AI’s gonna raise your game

ChatGPT’s competitor claude.ai just got a huge upgrade. The Verge reports that, according to parent company Anthropic, the latest version of Claude can digest and work with 500 pages of material for you in one go.

That’s a lot. But there’s more.

Meanwhile the release of Google's forthcoming Gemini, a multimodal AI system, has been pushed to the first quarter of 2024, presumably to make it even stronger and better before launching. (Multimodal is just fancy AI speak for a system that can work with words, pictures and/or sounds all at the same time.)

I’ve previously written about how generative AI systems like ChatGPT are turning not just knowledge but also many forms of creative and analytical output into “uneconomic goods,” or, something that is so plentiful and free, that people literally cannot charge for it.

As I put it in a piece that kinda blew my own mind back in April of this year, we’re on the precipice of a “post scarcity economy”:

I find the idea that Large Language Models (LLMs) are slowly (or quickly) turning knowledge, analysis and creative output into an "uneconomic goods” totally fascinating. I’ve even coined a name for this phenomenon: “the reverse industrial revolution,” because this wild new era is shifting value away from knowledge and (back) towards labor.

AND YET

As LLMs get more and more competent and impressive, I remain convinced that the implications are even more nuanced than they first appear.

Consider the following:

Even now, in our post privacy, post everything online era, not all info is accessible— even to ChatGPT.

Sidebar: it’s not like they’re not trying. I’ve written about some of the efforts that the firms behind LLMs, like Anthropic and OpenAI are undertaking in order to build their systems. In addition to scraping every last corner of the open Internet, and —in Google’s case— every YouTube video ever uploaded, these AI training efforts includes things like paying actual human beings to have random, idle chats about anything in the world with their systems (this actually happens), and (probably) also includes efforts like signing up to every online course ever and feeding transcripts of all of that in to their algos.

All this work on the part of generative AI firms are making a vast expanse of info readily accessible to the system, and therefore all of us, basically for free. (Most basic or entry-level versions of LLMs are free to use.)

This “democratization of knowledge” could be seen as devaluing the information in a traditional economic sense, where value is often driven by scarcity.

But hang on.

While LLMs do provide broad access to information, there's still a massive distinction between surface-level information and deep, specialized, first hand knowledge. It makes logical sense that human expertise in areas requiring years of study or unique experiences in extremely niche fields may still hold significant economic value. At least for a little while.

The shift towards knowledge as an uneconomic good, is already encouraging new business models: there’s a whole cottage industry of side hustlers using ChatGPT and Canva to generate countless e-books on any topic you can think of and selling them to the public, via short-and-sweet TikToks and Instagram Reels.(Pro tip: health, wealth, and relationships are 3 that tend to sell well.) But these e-books are fast becoming uneconomic goods themselves as more and more people push into this side hustle.

Look for the opportunity to shift from simply selling information, to the idea of individuals offering personalized experiences, insights, or consulting services that use human charisma and charm to leverage all this now-freely available knowledge.

OK, WHAT’S THE POINT HERE?

LLMs can provide practically endless information, impressive creative output and some level of sophisticated analysis, and are only getting better and better at doing all three by the day.

I believe that this “rising tide” will force our own human ingenuity to come up with ever more novel and valuable ways in which we can interpret, apply, and synthesize information into new ideas or solutions —at least if we want to continue getting paid for what’s in between our ears.

Think of all this as a value chain, with raw data at the low end of it, and organized information slightly higher up, and valuable insights, higher still.

I’ve written a lot about this value chain in the context of data, using the metaphor of oil to help it make sense. The crude that’s deep under ground is valuable, but it’s not all that practical or usable in that format. Oil really becomes valuable (on a per barrel basis) when it is refined into high octane fuel, and more valuable still, when it is further refined and shaped into high ticket products like official baseball jerseys, or Phillipe Starck chairs.

Continuing this train of thought, it’s easy to imagine economic value shifting from the raw information and entry level analysis and image making (provided by ChatGPT and the like) and towards human experts offer greater depth, more nuanced understanding, delivered with charisma.

Bottom line: as LLMs get more and more impressive and capable, rather than rendering human knowledge “unsellable,” systems like Claude 2.1, Gemini, and the eventual GPT 5 are going to force us to use our newly freed up cognitive resources to perform ever more complex, creative, or empathetic tasks (that machines can't replicate).

In short, AI is going to force us all to raise our game.

How are you going to raise yours?

More:

OpenAI rival Anthropic makes its Claude chatbot even more useful »»

Google’s Bard AI chatbot can now answer questions about YouTube videos »»

Written by Jon Kallus. Thanks for reading.