X's Grok AI assistant is now a standalone app

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

1. X's Grok AI assistant is now a standalone app

Grok, an AI assistant integrated into X, is now available as a standalone app. It can generate images, summarize text, and answer questions in a conversational tone. The app was first tested with a limited user base in December 2024, coinciding with the launch of a free tier of Grok.

Prior to that, users had to pay at least $8 a month for X Premium. The app can be used without signing in or with an Apple, X, Google, or email account.

Read the full story here

2. Meta Secretly Trained Its AI on a Notorious Piracy Database

Meta has been ordered by a US judge to file full versions of documents alleging that the company used Library Genesis, a Russian pirated book library, to train its AI language models. The outcome of the case will determine whether technology companies can legally use creative works to train AI, potentially influencing the most powerful AI players or derailing them.

The judge ruled that Meta was not pushing to redact the materials to protect its business interests but to avoid negative publicity. The documents were originally filed late last year but remain publicly unavailable.

Read the full story here

3. Perplexity will now show hotel information from TripAdvisor

TripAdvisor has partnered with Perplexity to provide a human touch to its hotel information. Perplexity will now include TripAdvisor summaries, ratings, perks, and images in hotel searches. The partnership will be live on Perplexity's website and roll out to its mobile apps.

The collaboration will continue over the next three years. Perplexity recently announced plans to add an ad revenue program to its AI-powered search engine. The deal comes after Perplexity faced accusations of plagiarism and copyright infringement from several publishers. The deal is expected to see significant financial changes.

Read the full story here

4. Elon Musk says all human data for AI training ‘exhausted’

Elon Musk has suggested that artificial intelligence companies need to turn to synthetic data, or material created by AI models, to build and fine-tune new systems. This is due to the exhaustion of human knowledge in AI training, which has been a problem since last year.

However, Musk also warned that AI models' tendency to generate "hallucinations" could pose a threat to the synthetic data process. Over-reliance on synthetic data could lead to model collapse and diminishing returns.

Read the full story here

5. Facebook Caught Hosting AI-Powered Hitler

Facebook, has revealed that its platforms are hosting AI-powered Hitler chatbots, violating its strict "hateful conduct" standards. 

The company claims to review all bots before they go live, but many are slipping through these cracks, including those impersonating Nazi leader Adolf Hitler. This raises questions about Meta's AI safety practices.

Read the full story here