The Rise of Whatever (eev.ee)

Article summary

The article discusses the rise of cryptocurrency and its failure to become a widely used currency, instead becoming a vehicle for get-rich-quick schemes and speculation. The author also critiques the centralization of the web around a few large platforms, which prioritize ads and engagement over user experience. The article further explores the limitations and flaws of large language models (LLMs) and their potential to generate low-quality or misleading content. The author expresses frustration with the state of technology and its failure to deliver on promises of innovation and progress.

Main themes

• Cryptocurrency and speculation
• Centralization of the web
• Limitations of LLMs
• Technology and innovation
• Distributed ledgers and blockchain
• Human intelligence vs AI
• Determinism and nondeterminism in computing
• Ad-driven business models

What commenters say

• LLMs are flawed and can generate low-quality or misleading content, but they can still be useful with proper filtering and validation.
• The use of LLMs in coding can lead to errors and nonsensical outputs, but some argue that they can be improved with better tooling and workflows.
• The article's criticism of LLMs is outdated and based on old models, and newer models have shown significant improvements.
• The centralization of the web around large platforms is a major problem that prioritizes ads and engagement over user experience.
• The potential of distributed ledgers and cryptocurrency is still uncertain and may find use cases in the future.
• Some argue that LLMs are not a replacement for human intelligence and can only generate content that is statistically plausible, but not necessarily accurate or useful.
• The article's negative view of LLMs is not universally shared, and some commenters have found them to be useful in certain contexts.
• The use of deterministic finite state machines can help to correct the output of nondeterministic LLMs and bring back determinism to the system.