Il est fort probable que l'avenir des LLM (que je pense hélas destinées à rester) soit des modèles ouverts utilisables sur le pc de tout un chacun + qq gros modèles spécialisés sur datacenter, et je pense que ça terrifie encore plus OpenAI et compagnie qu'une disparition pure et simple.

Show Post

arXiv did something similar last year - restricted CS review-article submissions after an LLM-generated survey paper flood. Every lightly-reviewed venue seems to be hitting this same wall now.

Show Post

Study: AI isn't killing art but it poses a significant threat to human creativity. The question isn't "can AI make convincing art?" It's whether the creative process adds something that the finished product alone lacks. www.jpost.com/science/ai-n...

Show Post

Donald Norman was definitely a huge influence, particularly his Cognitive Artifacts paper (www.researchgate.net/publication/...) So many of those core UX design concepts are not only remarkably durable, we can enact them in new ways with responsible & purposeful use of AI in our HCIs.

Show Post

I mean, the big frontier AI model firms currently get the bulk of their revenue from a small handful of firms, and need demand to grow exponentially at a rate much higher than it has done to date to get to profit. 'Increasingly worrying' means 'resting on multiple fantasies'.

Show Post