8,773
edits
Talk:Restless Souls/Technology: Difference between revisions
Talk:Restless Souls/Technology (view source)
Revision as of 12:52, 12 March 2026
, Thursday at 12:52no edit summary
Paradox-01 (talk | contribs) mNo edit summary |
Paradox-01 (talk | contribs) mNo edit summary |
||
| Line 564: | Line 564: | ||
*** Unique training data can also sometimes be returned. Even if a piece of data appeared only once, if it is highly distinctive or contextually reinforced by similar patterns, the model may reproduce it with surprising fidelity. This occurs because the model has overfitted locally to these rare signals, making them more "retrievable" than typical generalized content. | *** Unique training data can also sometimes be returned. Even if a piece of data appeared only once, if it is highly distinctive or contextually reinforced by similar patterns, the model may reproduce it with surprising fidelity. This occurs because the model has overfitted locally to these rare signals, making them more "retrievable" than typical generalized content. | ||
** Reinforcement learning from human feedback (RLHF) and (its successor RLAIF) can be named as another important feature that added a reward model for higher quality and alignment. | ** Reinforcement learning from human feedback (RLHF) and (its successor RLAIF) can be named as another important feature that added a reward model for higher quality and alignment. | ||
** Other features | ** Other features like chain of thought (reasoning), mixture of experts (MoE), context expansion and the use of external tools via model context protocol (MCP) to compensate own shortcomings are better described as incremental improvements in the evolution of GenAI.<!-- | ||
* GPT = Generative pre-trained transformers (Large Language Model with the actual "learning" part de facto outsourced to humans: Reinforcement learning from human feedback (RLHF), in best case GPTs have a ''transplanted base intelligence'' but they lack the important feature to really learn for themselves. Low quality "synthetic data" can even worsen the models. | * GPT = Generative pre-trained transformers (Large Language Model with the actual "learning" part de facto outsourced to humans: Reinforcement learning from human feedback (RLHF), in best case GPTs have a ''transplanted base intelligence'' but they lack the important feature to really learn for themselves. Low quality "synthetic data" can even worsen the models. | ||
:: After the obvious slowdown in advancement through scaling, chain of thought (COT) was introduced. It is also known under the marketing term "reasoning [AI]". (See: Gemini 2.0 Flash Thinking and ChatGPT o3, the later one especially [https://techxplore.com/news/2024-12-ai-human-general-intelligence.html trained to cheat … erm ... score high in the ARC-AGI test].) | :: After the obvious slowdown in advancement through scaling, chain of thought (COT) was introduced. It is also known under the marketing term "reasoning [AI]". (See: Gemini 2.0 Flash Thinking and ChatGPT o3, the later one especially [https://techxplore.com/news/2024-12-ai-human-general-intelligence.html trained to cheat … erm ... score high in the ARC-AGI test].) | ||
| Line 587: | Line 587: | ||
:: In discussions AGI is often equated with super intelligence. The argument is that as soon as AGI is achieved ASI is just around the corner. This understates the '''wide scope of human intelligence''' and that AGI is achieved first by hyperscalers, making further improvement difficult through further scaling. | :: In discussions AGI is often equated with super intelligence. The argument is that as soon as AGI is achieved ASI is just around the corner. This understates the '''wide scope of human intelligence''' and that AGI is achieved first by hyperscalers, making further improvement difficult through further scaling. | ||
:: Major roadblocks to AGI include: | :: Major roadblocks to AGI include: | ||
::: An architecture that supports continuous learning which avoids "[https://www.ibm.com/think/topics/catastrophic-forgetting catastrophic ] [https://www.fz-juelich.de/en/news/archive/press-release/2025/novel-memristors-to-overcome-ai2019s-catastrophic-forgetting forgetting]". | ::: An architecture that supports continuous learning which avoids "[https://www.ibm.com/think/topics/catastrophic-forgetting catastrophic] [https://www.fz-juelich.de/en/news/archive/press-release/2025/novel-memristors-to-overcome-ai2019s-catastrophic-forgetting forgetting]". | ||
::: Environment that allows | ::: Environment that allows self-improvement. -- Alignment concerns require any AGI's self-improvement to be carefully sandboxed. (This is not exactly a roadblock but a limiting factor in terms of how fast an AGI can be built / will emerge.) | ||
::: The ability to acquire a form of machine wisdom analogous to human wisdom. | ::: The ability to acquire a form of machine wisdom analogous to human wisdom. | ||
::: Abstract reasoning and generation of completely new thought patterns (beyond pattern remixing and transmission). | ::: Abstract reasoning and generation of completely new thought patterns (beyond pattern remixing and transmission). | ||