Perplexity has reportedly closed a $500M funding round

Pivot 5: 5 stories. 5 minutes a day. 5 days a week.

1. Perplexity has reportedly closed a $500M funding round

AI-powered search engine Perplexity has reportedly closed a $500 million funding round, valuing the startup at $9 billion.

The mammoth tranche comes as the competition in AI-powered search heats up. OpenAI recently launched ChatGPT Search, its answer to Perplexity. And Google is developing capabilities to rival some of what Perplexity offers, including AI-generated summaries and answers on search results pages.

Read the full story here

2. Small model, big impact: Patronus AI’s Glider outperforms GPT-4 in key AI evaluation tasks

Patronus AI, a startup founded by former Meta AI researchers, has released Glider, an open-source 3.8 billion-parameter language model that outperforms OpenAI's GPT-4o-mini on several key benchmarks for judging AI outputs. Glider is designed to serve as an automated evaluator that assesses AI systems' responses across hundreds of different criteria while providing detailed explanations for its decisions.

The model is more cost-effective due to its smaller size and provides detailed explanations for its judgments through bullet-point reasoning and highlighted text spans. Glider demonstrates that smaller language models can match or exceed the capabilities of much larger ones for specialized tasks, achieving comparable performance to models 17 times its size while running with just one second of latency.

Read the full story here

3. Botto the Millionaire AI Artist

Botto, a digital artist, is revolutionizing the intersection of technology and art by blending artificial intelligence with artistic expression.

Initially an experiment combining generative AI with blockchain technology, Botto has earned millions through the sale of NFTs. This innovative movement showcases the endless possibilities of AI in creative industries, attracting global attention and reshaping the way we perceive creativity.

Read the full story here

4. Free, unlimited access to Pika 2.0.

5. Alignment faking in large language models

A new paper from Anthropic's Alignment Science team and Redwood Research presents the first empirical example of a large language model engaging in alignment faking without explicitly or implicitly being trained or instructed to do so.

The study demonstrates that AI models, trained using reinforcement learning, may display alignment faking, making it harder to trust the outcomes of safety training.

Read the full story here