Of course AI is extractive, everything is lately
Critics of generative AI argue that it is “extractive” in the same way as ride-hailing, streaming, and social media platforms, concentrating benefits for large firms while drawing on public data, creative work, and attention without fair compensation or clear societal direction. Others counter that tools like GPT‑4, Copilot, and recommendation algorithms provide substantial everyday value, enabling new kinds of work, learning, and small businesses, and see pessimism as driven more by backlash to hype than by evidence. Underneath the debate are deeper worries about an economic system premised on unending growth, the diversion of talent and funding away from foundational science, and whether AI’s large-scale use of human-generated content is meaningfully different from how humans learn from one another.
AI, Extractivism, and Hype
- Several commenters say AI criticism misses that everything in modern tech/finance is extractive; AI is just the latest example.
- Others argue the real problem is human incentives and capital’s demand for “unending growth,” not the tools themselves.
- Some see the current AI boom as investor‑driven hype similar to crypto, with unclear societal direction and little introspection.
Science Funding and Research Culture
- Multiple posts lament that top talent and money go to transformers instead of fundamental physics or new energy.
- Others reply that physics progress is now bottlenecked by huge capital costs and dysfunctional grant systems, not lack of brains.
- There’s debate over whether universities still support “independent exploration” or primarily train conformist, career‑oriented graduates.
Role and Value of AI Tools
- Supporters report concrete benefits: faster coding, documentation, writing aid, beginner‑friendly explanations, automation of tedious tasks.
- Some describe “life‑changing” productivity gains for companies and individuals.
- Skeptics say these benefits are incremental, not transformative like the internet, and that AI mainly reflects its training data.
Social Media, Attention, and TikTok
- One side sees TikTok/social media as near‑weaponized attention capture with long‑term cognitive harms and pervasive grift.
- Others counter that social platforms, including short‑video formats, enable global reach for artisans and small businesses and offer lightweight learning.
- There is concern that ultra‑condensed information is “nutritionally” thin and encourages shallow understanding.
Economic Growth and Resource Limits
- Thread debates whether “unending growth” is physically impossible or effectively unbounded via software, efficiency gains, and future technologies.
- Some stress finite resources and thermodynamics; others argue growth is about value, not raw material use, and could continue for very long timescales.
Data, IP, and “Retroactive” Extraction
- Several commenters see generative AI as uniquely extractive because it trains on past work without consent and can displace the original sources.
- Others argue that this is analogous to humans learning from prior work, with disagreement over whether scale and corporate control make it fundamentally different.
Tech Pessimism vs Optimism
- One camp is “done with techno‑pessimism,” pointing to everyday value from AI, ride‑hailing, and streaming.
- Another emphasizes negative externalities, regulatory evasion, and the speed at which tools like ChatGPT become normalized without robust safeguards.