Bilingual

The Obligatory AI Post
5 hours ago
- The author has minimal firsthand experience with AI and eschews its use, but feels compelled to write a blog post to defend against false accusations of writing 'AI slop' based on stylistic choices like em dashes.
- AI is criticized for promoting ignorance and laziness by replacing the need to learn or struggle, with the goal of making humans 'super-dumb' rather than AI becoming super-intelligent.
- AI's resource consumption exacerbates climate change, causes hardware shortages, and raises electricity costs, while vendors engage in intellectual property theft and regulatory capture.
- The author views AI as an investment bubble or 'black hole' that incentivizes job automation, worker burnout, and productivity theater—such as generating countless apps that nobody buys.
- Speed in software development is seen as a vice; AI-generated code may be fast but software engineering is an art, and mature software should be changed cautiously.
- Using the example of Google fixing bugs with AI, the author argues that corporations could have hired humans but prioritized feature release over bug fixes, making AI a cheap substitute.
- Drawing on '2001: A Space Odyssey,' the author warns that AI like HAL 9000 is an 'idiot savant' that makes catastrophic errors due to human overtrust, and AI should not be dismissed as 'just a tool' since it fundamentally changes society.
Who manages the agents?
20 days ago
- The article presents two AI futures: one where AI is a deity controlled by a small clergy-like group, and another where AI amplifies human capabilities with humans at the center.
- A new technical clergy is emerging—those who build frontier AI systems, have privileged access, and decide who can use them, warning of mass unemployment but seeing it as liberating.
- Concerns exist about a two-tier system: powerful AI for a select few (e.g., partners, researchers) vs. constrained AI for the public, due to safety risks like bioweapons and disinformation.
- In fields like software development, a small fraction of superusers achieve 100x productivity gains with AI agents, while the median sees little improvement, widening the gap.
- The future should be built by billions, not just for them, emphasizing universal participation over mere access, with humans managing agents to enhance creativity and productivity.
- Companies should focus on making the median worker 2x more productive through AI agents managed by employees, not on headcount reduction, to avoid disenfranchisement.
- Agents should be sovereign—owned and controlled by companies, with humans at the center managing identities, memory, and skills—to ensure accountability and prevent dependency.
- The article calls for rejecting prophecies of AI concentration, advocating instead for distributing AI to increase human agency and building a future where everyone becomes an agent manager.
Enough
a month ago
- The world was progressing in replicating complex systems, building layers of abstraction that were individually simple but collectively complex, though experts could be trained to understand them.
- Learning and understanding were highly valued in a capitalist society that prized speed and efficiency, yet the foundational step always involved internalizing the structure of things, often through explanation from knowledgeable individuals.
- Companies hired explainers and understanders for high pay, recognizing their essential role in operating machines and systems, despite potential moral conflicts, with workers believing in their collective moral stance and power to resist unethical demands.

Hasty Briefsbeta

technology ethics

The Obligatory AI Post

Who manages the agents?

Enough