Systems

The changing nature of Change ⓧ Systems as fiction ⓧ Scaling LLMs

👋 On time for your weekend: a round-up of this week’s remarkable stories at the intersection of technology, business, design, and culture. Three reads and three listens; no fluff, just stuff ⚡

There is absolutely no inevitability as long as there is a willingness to contemplate what is happening—Marshall McLuhan

📚 Reading

Chegg, ChatGPT, and the changing nature of change:

Structural uncertainty reshapes the entire playing field itself, and with that, you are often left playing the wrong game. This type of uncertainty doesn’t just change how companies compete—it changes what they are competing for or against […] Structural uncertainty often leaves you fixing problems with the wrong tools. The key mistake leaders make when confronted with uncertainty is they try to solve structural problems using operational responses.

Sangeet Paul Choudary | 9 minutes

Places to intervene in a system:

All systems are fictions that we project onto our environment for our own instrumental purposes. Those purposes are determined by our mindset, by the way we see the world around us. If the pursuit of our instrumental goals results in catastrophe, perhaps it is possible to change our mindset, the mindset or paradigm out of which the system — its goals, structure, rules, delays, parameters — arises?

Grodon Brander—Squishy | 2 minutes

Thoughts on LLM scaling slowing down:

We don't appreciate the jump because we've been frog boiled. The OG GPT-4 would feel dumb now, especially for coding. GPT-4-0613 is 65th on LM Arena, behind 8B param open source models. Many evals are saturated so it's hard to see how far we've come. The models have also saturated just like...being amazing chatbots, so most chat users won't tell the difference until there's a more fundamental shift in how we use these models.

Toby—X | 1 minute

🎧 Listening

Leading with AI:

Humans and AI are basically going to be collaborating. You see, interpersonal skills and collaboration are already important […] this is how we're going to evaluate leaders today, still show those leadership behaviors, but now leveraging the value of AI […] 70% of their tech budget goes to the adoption of AI itself. Business leaders don't understand it. They think it's all about technology.

David de Cremer—Lancefield on the Line | 42 minutes

How an Anonymous Researcher Predicted AI's Trajectory:

The AIs will be too myopic to execute any kind of novel long-term strategy and seize new opportunities. So that would presumably give you this paradigm where you have like a human CEO who does the vision thing and then the AI corporation kind of like scurries around underneath them doing the CEO's bidding. And they don't have the taste that the CEO has.

Gwern Branwen—Dwarkesh Patel | 97 minutes

Science is in trouble and it worries me:

Academic research is becoming increasingly risk averse. This is because there's so much competition for funding and jobs that researchers have very strong incentives to pick a topic that are reliably produced results. This encourages them to pursue low-risk mainstream research and plausibly explains the lack of novelty and disruptive ideas […] why does it persist?

Sabine Hossenfelder | 24 minutes

💎 Timeless

1️⃣ year ago—What Happened to the New Internet?

2️⃣ years ago—Picture Limitless Creativity at Your Fingertips

3️⃣ years ago—Science Fiction Is a Luddite Literature