6 ms·
> The hazard we are experiencing with LLM right now is not how freely accessible and powerfully truthy it's content is, but it is precisely the controls upon it
by nonethewiser 3y ago
> The hazard we are experiencing with LLM right now is not how freely accessible and powerfully truthy it's content is, but it is precisely the controls upon it which are trying to be injected by the large model operators which are generating mistrust and a poor understanding of what these models are useful for.
This slices through a lot of double speak about AI safety. At the same time, people use “safety” to mean not letting AI control electrical grids and to ensure AIs adhere to partisan moral guidelines.
Virtually all of the current “safety” issues fall into the latter category. Which many don’t consider a safety issue at all. But they get snuck in with real concerns about integrating an AI too deeply into critical systems.
Just wait until google integrates it deeply into search. Might finally kill search.
- sroussey 3y agoWhat are you talking bout? It’s been deeply running google search for many years. And AI for electrical grids and factories has also been a thing for a couple years.
- Jensson 3y agoLLMs hasn't been deeply integrated into google search for many years. The snippets you see predates LLMs, it is based on other techniques.
- consp 3y agoWhat people call AI might be an algorithm but algorithms are not AI. And it's definitely algorithms which do what you describe. There is very little magic in algorithms.
- User23 3y agoMy read of "safety" is that the proponents of "safety" consider "safe" to be their having a monopoly on control and keeping control out of the hands of those they disapprove of. I don't think whatever ideology happens to be fashionable at the moment, be it ahistorical portraits or whatever else, is remotely relevant compared to who has the power and whom it is exercised on. The "safety" proponents very clearly get that.