Y el problema de Marcus en concreto es que el mismo que hace uno meses firmaba lo de "hay que parar la IA porque es muy peligrosa", es el mismo que jalea a las empresas para que publiquen más rápido, y si no es que "hemos chocado con un muro".
SAFETY
-
OpenAI implements opt-out toggle for voice model training data
By
–
Btw, Voice data training will have a separate toggle and will be turned off by default. +1 to @OpenAI for this
-
Voice data training settings and privacy defaults
By
–
Not really, voice data training will be a separate option and off by default
-
TTS Systems Show Improved Error Recovery Despite Occasional Gibberish
By
–
TTS' still spews gibberish sometimes, but the realization and error recovery in the first example is astounding 😯 https://t.co/krMFGypACE
— Dr. Angelica Lim @petitegeek.bsky.social (@petitegeek) 14 mai 2024TTS' still spews gibberish sometimes, but the realization and error recovery in the first example is astounding
-
AI Interface Design: Balancing Adaptation and Human Connection
By
–
Overall this jives with my dream of having an interface that adapts to the way humans communicate. But the elephant in the room is how to not go down the dystopian path of people getting too attached and neglecting real human relationships
-
Expressive Robot Voices: Designing Clear AI Identity
By
–
What if a voice AI could be super expressive, but still *clearly* be a robot voice? Just like I want my robots to look like robots, not humans.
-
Concerns regarding reliability in voice and vision AI models
By
–
Hope it won’t be the case with a voice & vision model
-
OpenAI Must Address Safety Issues Before Scaling Models
By
–
Sólo le pido a OpenAI que antes de abrir sus modelos al mundo y hacerlos más potentes, solucionen esto…
-
OpenAI’s Reveal and Latest AI Tools Announcements Today
By
–
Top stories in AI today: -Rumor mill: OpenAI’s Monday reveal
-Anthropic’s new tool automates prompting
-How to create realistic AI avatar videos
-MIT studies AI deception capabilities
-6 new AI tools & 4 new AI jobs Read more: http://
therundown.ai/p/openais-big-
reveal
… -
LLM-Generated Papers Risk Slowing Scientific Progress
By
–
What's holding back research isn't a lack of verbose, low-signal, high-noise papers. Using LLMs to automatically generate 100x more of those will not accelerate science, it will slow it down.