Last Week in AI #102 - Nonsense Sentience, Condemning GPT-4chan, DeepFake bans, CVPR Plagiarism
Jun 26, 2022
Dive into the absurdity of LaMDA's supposed sentience and the widespread condemnation of GPT-4chan. Discover how AI is reshaping data center efficiencies at tech giants like Microsoft and Meta. Explore the revolutionary use of AI in addressing long COVID and the ethical concerns surrounding deepfakes, with the EU stepping in to regulate. Witness the intersection of creativity and AI as it crafts the first AI-generated magazine cover while also stirring controversies in academia with cases of plagiarism.
AI Snips
Chapters
Transcript
Episode notes
LaMDA's Sentience
- LaMDA's sentience claims are overblown and lack scientific basis.
- Its behavior stems from sophisticated pattern matching, not genuine understanding.
GPT-4chan Controversy
- Yannic Kilcher trained a GPT model on 4chan data and deployed it, causing controversy.
- Stanford professors circulated a condemnation statement, signed by many, including Yoshua Bengio.
AI-Powered Data Centers
- Microsoft and Meta, like Google, leverage AI for data center optimization.
- AI helps manage energy consumption, safety, and construction schedules, showcasing practical applications.