Agents
Abstract decision-making units in economic models; Goldin contrasts this framing with 'humans' to explain economics' recruitment problem.
As defined in cwt_claudia-goldin
Other definitions
Historical term for autonomous AI entities used in 1990s research, perceived as more ominous than modern 'chatbots'—refers to AI systems designed to act independently on behalf of users.
Jaron Lanier Looks into AI's Future | AI IRL
Where it comes up · 18
- Skill Issue: Andrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AI14 claims
- AI 2027: month-by-month model of intelligence explosion — Scott Alexander & Daniel Kokotajlo5 claims
- The coming AI revolution and the question no one can ignore | Max Tegmark | Strange Loop3 claims
- Nvidia v. Google, Scaling Laws, and the Economics of AI3 claims
- Andrej Karpathy — AGI is still a decade away2 claims
- Mark Zuckerberg — Llama 3, $10B models, Caesar Augustus, & 1 GW datacenters2 claims
- Anthropic CEO Dario Amodei Talks Scaling Laws, AI Arms Races, and Radical Abundance1 claim
- DeepSeek's Invisible Founder Just Talked for 4 Hours. It Leaked.1 claim
and 10 more
Most important claims · 10 of 41Search all claims for “Agents” →
- Treating AI as analogous to past economic transitions (agriculture to manufacturing to service) is clearly delusional because, unlike prior revolutions, AI in the limit cancels the need for human cognition almost entirely—the 'almost' surviving only where we retain a sentimental attachment to the human provenance of products.457 - More From Sam The Epstein Files The Newsom Factor Don Lemon S Arrest Ai Market Disruption
- Even the purely optimistic AI scenario—massive productivity gains where one person can spin up a thousand agents and run multiple companies at once—is itself an emergency, because a world where human labor becomes optional is untenable unless society anticipates the transition and builds mechanisms to share the wealth rather than concentrating it in a handful of founders.459 - More From Sam Corruption Immigration The End Of White-collar Work And More
- AI models pose two distinct safety risk categories: (1) autonomous behavior risks (models acting autonomously without human oversight), which are not yet significant but will become concerning as models become more capable and autonomous; and (2) misuse risks from distributed actors using models for harmful purposes.Anthropic CEO Dario Amodei Talks Scaling Laws, AI Arms Races, and Radical Abundance
- Coding agents built by OpenAI are already extremely good at their task, and coding is a sufficiently general-purpose capability that AI systems able to write code can accomplish many other things.Sam Altman talks the NYT lawsuit, Meta's talent poaching, and Trump on AI | Interview
- Coding agents are good at boilerplate and code common on the internet, but they fail on unique, intellectually intense code: they over-rely on memorized common patterns, bloat code with defensive try-catch and production scaffolding, use deprecated APIs, and can't internalize custom non-standard implementations—so for novel work autocomplete is the sweet spot rather than full agents.Andrej Karpathy — AGI is still a decade away
- Current LLM agents have severe jaggedness: they excel at verifiable, RL-trained tasks (code generation, tool use) but fail catastrophically at soft/unverifiable tasks (like telling jokes), because they're only optimized along dimensions where rewards can be computed, leaving trained-off domains untouched.Skill Issue: Andrej Karpathy on Code Agents, AutoResearch, and the Loopy Era of AI
- There won't be one singular AI users interact with; every business will want an AI representing its own interests rather than one that might sell competitors' products, and the ~200 million creators on Meta's platforms will want AIs they own and train to engage their communities beyond the limits of their own time.Mark Zuckerberg — Llama 3, $10B models, Caesar Augustus, & 1 GW datacenters
- The alignment problem may be partially solved 'by default' through properties of large language models that naturally develop common sense understanding of goals from next-token prediction training, making them less prone to the Malign Omniscience failure modes that earlier alignment researchers feared.AI 2027: month-by-month model of intelligence explosion — Scott Alexander & Daniel Kokotajlo
- If we escape misalignment and malicious-use risks and spread the wealth around, the human future could be one of pure leisure, art appreciation, and AI-enabled creativity—where one creative person can do the work of a thousand by spinning up agents, and where AI delivers cures for every disease, even if it takes 400 different cures for cancer's various forms.455 - More From Sam Guns Propaganda Ai And Power Unbound
- Large language models can be extended from outputting tokens to outputting commands that control robots, humanoid robots, or self-driving cars, making them not just language systems but controllers of physical agents.The coming AI revolution and the question no one can ignore | Max Tegmark | Strange Loop