Most discussions I see about generative AI ask what’s possible in order to figure out what the future might be. I think that’s the wrong question.

Instead we should be asking what’s the opportunity to get a sense of where we’re heading. Every new technology brings the promise of something better. Generative AI arrived with that promise in a macroeconomic climate where the cost to invest is high.

There are two commonly touted opportunities being promised.

AI as an automation tool

The most obvious opportunity is one that is already happening: generative AI automates some knowledge-based tasks for productivity gains. Microsoft is leading the way with its $10b investment in OpenAI and multiple Copilot launches. The natural extension is AI agents, where multiple tasks are connected and performed autonomously. For a company the productivity opportunity here is obvious.

However, what’s quickly becoming an appealing is generating low quality content, because it doesn’t matter what the quality is if it still makes money. That impact is feeling the most immediate:

AI is killing the old web, and the new web struggles to be born (The Verge / James Vincent)

It’s the dynamics of AI — producing cheap content based on others’ work — that is underwriting this change, and if Google goes ahead with its current AI search experience, the effects would be difficult to predict. Potentially, it would damage whole swathes of the web that most of us find useful — from product reviews to recipe blogs, hobbyist homepages, news outlets, and wikis. Sites could protect themselves by locking down entry and charging for access, but this would also be a huge reordering of the web’s economy. In the end, Google might kill the ecosystem that created its value, or change it so irrevocably that its own existence is threatened.

AI as just an automation tool can sound disappointing because it’s evolutionary and not revolutionary.

AI as a new computer paradigm

The other big promised opportunity is that AI will bring a whole new way we use computers. The pitch is we’ll use natural language as the interface. That’s silly. ChatGPT sure looks like natural language, but it isn’t. To use it we need to learn how to write effective prompts in much the same way we learnt how to use Google search operators.

The vision of a new computer paradigm comes from the success of ChatGPT and the assumption that it could become the next platform to build apps and businesses on top of.

As AI Gets Better, Could It Swallow The Companies Building With It? (Big Technology / Alex Kantrowitz)

This type of eager and widespread developer adoption was always the goal for the AI platforms, but something funny happened along the way. The consumer tools they built to show their capabilities — like ChatGPT and Bing Chat — garnered so much interest they’ve become destinations themselves. And as their AI gets better, it may well end up subsuming some of the products and features developers are building with the technology today.

This opportunity is hard to imagine in the short term, and somewhat vague in the longer term. For the short term the investment cost is high with a number of barriers still unsolved: mass adoption, teaching users a new interaction model, still unsolved technical problems, and user retention.

In contrast, as a long term opportunity the future feels too uncertain to make a compelling big bet on how real this will be. Instead it feels more opportunistic to make smaller bets based on what will be most valuable (to users or businesses) and discover how the way we use tech shifts with AI. Again, an evolution not a revolution.

Seeing how much generative AI automation is already changing the web is a strong sign that there’s a lot more change coming, but how that will look will keep changing in ways we can’t see yet, over a much longer time horizon.

💩 Cool shit

Moderator Mayhem - Play this game and see what it’s like to be a content moderator for a social media platform.

Counter Forms - I honestly struggled to describe this link. It caught my interest for two reasons. One; it uses a unique free-form grid design. Two; it explores Australian and New Zealand typefaces and texts.

Hidden Door - I’m not sure how real this is, but I games built with AI are an intriguing look at what the tech can do. This one is based on a generative narrative.

ChatGPT Games - And this Github is a list of games you can try yourself entirely in ChatGPT.

CopyPastaDB - If you’ve spent any amount of time on the internet you’ve likely come across copypastas. I’m surprised how many there are.

Ooak Finder - A search engine to find out if a “handmade” product you’ve found online is actually unique.

Jelly Gummies - Gifs of jelly gummies. There’s not much more to it.

Share this email with a friend because citizens are already raking in millions via this loophole.