Useful Confusion
The 'AI' Black Box Myth ā§ From Firms to Networks ā§ Betting Against AI Agents (in 2025)
š On time for your weekend: a round-up of this weekās remarkable stories at the intersection of technology, business, design, and culture. Three to read and three to listen toāno fluff, just stuff ā”
Our technologies are ultimately not contrary to life, but are in fact an extension of life, enabling it to develop yet more options and possibilities at a faster rate. Increasing options and possibilities is also known as progress, so in the end, what the technium brings us humans is progressāKevin Kelly
š Reading
The Black Box Myth: What the Industry Pretends Not to Know About AI:
The shared myth-making around the black box of AI cultivates useful confusion about what can and cannot be known. It helps to cultivate the systems as more mysterious, even sublime, than they really are. It leads to a cottage industry of thinkersā¦This helps sell them, but also drives fears of doom that can redirect concerns from real people to hypothetical systems and abstract future scenarios.
Eryk SalvaggioāTech Policy Press | 10 minutes
From firms to networks: How coordination technology will reorganise everything:
Economic actors are increasingly machines and this machine economy will be larger over timeā¦The internet isnāt dead - it is alive with machinesā¦Human-based coordination mechanisms arenāt relevant as machines don't use language and trust relationships the way humans do. Our current coordination layer is designed for human-to-human interaction, but now we need one that works at machine scale and speed.
Matt Law | 12 minutes
Why I'm Betting Against AI Agents in 2025 (Despite Building Them):
The dirty secret of every production agent system is that the AI is doing maybe 30% of the work. The other 70% is tool engineering: designing feedback interfaces, managing context efficiently, handling partial failures, and building recovery mechanisms that the AI can actually understand and useā¦ Integration is where AI agents go to die.
Utkarsh Kanwat | 9 minutes
š§ Listening
Meta vs. OpenAI: Who Wins the Next Attention War?
The way that most people are using these LLMs is in a very deeply personal way. I don't think you'd have ChatGPTā¦now nearing Instagram's daily time spent. And I don't think that that would be the case if people weren't using it for productivity, for personal use cases. You just have to go on Twitter, go on Reddit and look at people debating the different ways and the different use cases that they're using these things for to come across people using it for deeply personal reasons.
Dave MorināMore or Less | 60 minutes
AI's Enterprise Adoption:
AI and AI agents are the perfect consumers of an API; they basically become these super users within your system on your APIsā¦To go from pre-cloud to post-cloud, that ripped through the entire stack all the way down to the infrastructure, for example, like tenancy, you have to rewrite everything. AI is more of a consumption layer thing, which is you just treat the existing systems as they are, and then the AI becomes the consumption layer.
Aaron Levieāa16z | 53 minutes
Decades of Disruption:
Traditional software is deterministic and does things that are easy to explain to machinesā¦things that are very hard for people to do, but they're easy to explainā¦Machine learning is stuff that's hard to explain to a computer. It's hard to explain why that credit card transaction is weird; it's hard to explain why that's a picture of a dog and not a catā¦And then it's like you tried to make a mechanical horse. It always falls over until robotics comes along. So that was machine learning
Benedict EvansāThe Network State Podcast | 124 minutes
š Timeless
1ļøā£ year agoāThe End of Software
2ļøā£ years agoāIn Defense of Strategy
3ļøā£ years agoāThe Great Fiction of AI