3 ms·
AI snake-oil usually degrades and wears off quicker over time when it is approaching its expiry date as more people realize that it hallucinates often without t
by rvz 3y ago
AI snake-oil usually degrades and wears off quicker over time when it is approaching its expiry date as more people realize that it hallucinates often without transparent explanation; study finds.
Unexplainable black-box stochastic parrots are no different to this no matter how cleverly packaged they are by AI bros.
- ftxbro 3y agodon't forget to say that it's a blurry jpeg of the web, and therefore has fake artefacts and is a degraded form of the original
- GaggiX 3y agoI'm surprised he didn't mention that it's a glorified Markov chain.
- ftxbro 3y agoi'm as hype for ai as anyone but mathematically it is a markov chain; a trained llm model is a function that gives a probability distribution over token n+1 conditional on the preceding n tokens.
- GaggiX 3y agoWell, on Wikipedia you can read "Random process independent of past history" and "a stochastic model describing a sequence of possible events in which the probability of each event depends only on the state attained in the previous event." honestly I would not consider the entire context window to be part of a single event, since the model generates tokens one at a time, separately.
- ftxbro 3y ago> honestly I would not consider the entire context window to be part of a single event, since the model generates tokens one at a time, separately. OK but traditional small language model Markov chains have context windows too, and they also generate tokens one at a time, separately. Maybe you are arguing that there is a qualitative difference between an eight token context window and an eight thousand token context window. I guess that would be fair enough.
- Zambyte 3y agoIt's interesting to me that people use the word "hallucinate" to describe when an LLM provides incorrect information, when it is doing the exact same process as when it happens to get things right: predicting a reasonably likely next token. It's "hallucinating" in the same way y=x^2 will predict the next point in a curve.
- willsmith72 3y agoDo you honestly believe it's snake-oil? I can't believe people still think this, nearly everyone I've talked about it with has been able to find an amazing use for AI now.
- ftxbro 3y agothe idea of AI skepticism or AI snake-oil has shifted, from meaning that it's ineffective to meaning that it's so effective that it's bad for society